xemacs-beta: man/texinfo.tex comparison

comparison man/texinfo.tex @ 5492:e82f5b7010fe

Merge some stuff in man, fix up Makefile -------------------- ChangeLog entries follow: -------------------- man/ChangeLog addition: 2010-02-19 Ben Wing <ben@xemacs.org> * widget.texi: * widget.texi (Top): * widget.texi (Introduction): * widget.texi (User Interface): * widget.texi (Programming Example): * widget.texi (Setting Up the Buffer): * widget.texi (Basic Types): * widget.texi (link): * widget.texi (url-link): * widget.texi (info-link): * widget.texi (push-button): * widget.texi (editable-field): * widget.texi (text): * widget.texi (menu-choice): * widget.texi (radio-button-choice): * widget.texi (item): * widget.texi (choice-item): * widget.texi (toggle): * widget.texi (checkbox): * widget.texi (checklist): * widget.texi (editable-list): * widget.texi (group): * widget.texi (Sexp Types): * widget.texi (constants): * widget.texi (generic): * widget.texi (atoms): * widget.texi (composite): * widget.texi (Widget Properties): * widget.texi (Defining New Widgets): * widget.texi (Widget Browser): * widget.texi (Widget Minor Mode): * widget.texi (Utilities): * widget.texi (Widget Wishlist): * widget.texi (Widget Internals): * widget.texi (GNU Free Documentation License): * widget.texi (Index): Sync with FSF 23.1.92. 2010-02-19 Ben Wing <ben@xemacs.org> * texinfo/fdl.texi: New file. * texinfo/texinfo.texi: * texinfo/texinfo.texi (Top): * texinfo/texinfo.texi (Copying Conditions): * texinfo/texinfo.texi (Overview): * texinfo/texinfo.texi (Reporting Bugs): * texinfo/texinfo.texi (Using Texinfo): * texinfo/texinfo.texi (Output Formats): * texinfo/texinfo.texi (Info Files): * texinfo/texinfo.texi (Printed Books): * texinfo/texinfo.texi (Formatting Commands): * texinfo/texinfo.texi (Conventions): * texinfo/texinfo.texi (Comments): * texinfo/texinfo.texi (Minimum): * texinfo/texinfo.texi (Six Parts): * texinfo/texinfo.texi (Short Sample): * texinfo/texinfo.texi (History): * texinfo/texinfo.texi (Texinfo Mode): * texinfo/texinfo.texi (Texinfo Mode Overview): * texinfo/texinfo.texi (XEmacs Editing): * texinfo/texinfo.texi (Inserting): * texinfo/texinfo.texi (Showing the Structure): * texinfo/texinfo.texi (Updating Nodes and Menus): * texinfo/texinfo.texi (Updating Commands): * texinfo/texinfo.texi (Updating Requirements): * texinfo/texinfo.texi (Other Updating Commands): * texinfo/texinfo.texi (Info Formatting): * texinfo/texinfo.texi (Printing): * texinfo/texinfo.texi (Texinfo Mode Summary): * texinfo/texinfo.texi (Beginning a File): * texinfo/texinfo.texi (Sample Beginning): * texinfo/texinfo.texi (Texinfo File Header): * texinfo/texinfo.texi (First Line): * texinfo/texinfo.texi (Start of Header): * texinfo/texinfo.texi (setfilename): * texinfo/texinfo.texi (settitle): * texinfo/texinfo.texi (End of Header): * texinfo/texinfo.texi (Document Permissions): * texinfo/texinfo.texi (copying): * texinfo/texinfo.texi (insertcopying): * texinfo/texinfo.texi (Titlepage & Copyright Page): * texinfo/texinfo.texi (titlepage): * texinfo/texinfo.texi (titlefont center sp): * texinfo/texinfo.texi (title subtitle author): * texinfo/texinfo.texi (Copyright): * texinfo/texinfo.texi (end titlepage): * texinfo/texinfo.texi (headings on off): * texinfo/texinfo.texi (Contents): * texinfo/texinfo.texi (The Top Node): * texinfo/texinfo.texi (Top Node Example): * texinfo/texinfo.texi (Master Menu Parts): * texinfo/texinfo.texi (Global Document Commands): * texinfo/texinfo.texi (documentdescription): * texinfo/texinfo.texi (setchapternewpage): * texinfo/texinfo.texi (paragraphindent): * texinfo/texinfo.texi (firstparagraphindent): * texinfo/texinfo.texi (exampleindent): * texinfo/texinfo.texi (Software Copying Permissions): * texinfo/texinfo.texi (Ending a File): * texinfo/texinfo.texi (Printing Indices & Menus): * texinfo/texinfo.texi (File End): * texinfo/texinfo.texi (Structuring): * texinfo/texinfo.texi (Tree Structuring): * texinfo/texinfo.texi (Structuring Command Types): * texinfo/texinfo.texi (makeinfo top): * texinfo/texinfo.texi (chapter): * texinfo/texinfo.texi (unnumbered & appendix): * texinfo/texinfo.texi (majorheading & chapheading): * texinfo/texinfo.texi (section): * texinfo/texinfo.texi (unnumberedsec appendixsec heading): * texinfo/texinfo.texi (subsection): * texinfo/texinfo.texi (unnumberedsubsec appendixsubsec subheading): * texinfo/texinfo.texi (subsubsection): * texinfo/texinfo.texi (Raise/lower sections): * texinfo/texinfo.texi (Nodes): * texinfo/texinfo.texi (Two Paths): * texinfo/texinfo.texi (Node Menu Illustration): * texinfo/texinfo.texi (node): * texinfo/texinfo.texi (Node Names): * texinfo/texinfo.texi (Writing a Node): * texinfo/texinfo.texi (Node Line Tips): * texinfo/texinfo.texi (Node Line Requirements): * texinfo/texinfo.texi (First Node): * texinfo/texinfo.texi (makeinfo top command): * texinfo/texinfo.texi (makeinfo Pointer Creation): * texinfo/texinfo.texi (anchor): * texinfo/texinfo.texi (Menus): * texinfo/texinfo.texi (Menu Location): * texinfo/texinfo.texi (Writing a Menu): * texinfo/texinfo.texi (Menu Parts): * texinfo/texinfo.texi (Less Cluttered Menu Entry): * texinfo/texinfo.texi (Menu Example): * texinfo/texinfo.texi (Other Info Files): * texinfo/texinfo.texi (Cross References): * texinfo/texinfo.texi (References): * texinfo/texinfo.texi (Cross Reference Commands): * texinfo/texinfo.texi (Cross Reference Parts): * texinfo/texinfo.texi (xref): * texinfo/texinfo.texi (Reference Syntax): * texinfo/texinfo.texi (One Argument): * texinfo/texinfo.texi (Two Arguments): * texinfo/texinfo.texi (Three Arguments): * texinfo/texinfo.texi (Four and Five Arguments): * texinfo/texinfo.texi (Top Node Naming): * texinfo/texinfo.texi (ref): * texinfo/texinfo.texi (pxref): * texinfo/texinfo.texi (inforef): * texinfo/texinfo.texi (uref): * texinfo/texinfo.texi (cite): * texinfo/texinfo.texi (Marking Text): * texinfo/texinfo.texi (Indicating): * texinfo/texinfo.texi (Useful Highlighting): * texinfo/texinfo.texi (code): * texinfo/texinfo.texi (kbd): * texinfo/texinfo.texi (key): * texinfo/texinfo.texi (samp): * texinfo/texinfo.texi (verb): * texinfo/texinfo.texi (var): * texinfo/texinfo.texi (env): * texinfo/texinfo.texi (file): * texinfo/texinfo.texi (command): * texinfo/texinfo.texi (option): * texinfo/texinfo.texi (dfn): * texinfo/texinfo.texi (abbr): * texinfo/texinfo.texi (acronym): * texinfo/texinfo.texi (indicateurl): * texinfo/texinfo.texi (email): * texinfo/texinfo.texi (Emphasis): * texinfo/texinfo.texi (emph & strong): * texinfo/texinfo.texi (Smallcaps): * texinfo/texinfo.texi (Fonts): * texinfo/texinfo.texi (Quotations and Examples): * texinfo/texinfo.texi (Block Enclosing Commands): * texinfo/texinfo.texi (quotation): * texinfo/texinfo.texi (example): * texinfo/texinfo.texi (verbatim): * texinfo/texinfo.texi (verbatiminclude): * texinfo/texinfo.texi (lisp): * texinfo/texinfo.texi (small): * texinfo/texinfo.texi (display): * texinfo/texinfo.texi (format): * texinfo/texinfo.texi (exdent): * texinfo/texinfo.texi (flushleft & flushright): * texinfo/texinfo.texi (noindent): * texinfo/texinfo.texi (indent): * texinfo/texinfo.texi (cartouche): * texinfo/texinfo.texi (Lists and Tables): * texinfo/texinfo.texi (Introducing Lists): * texinfo/texinfo.texi (itemize): * texinfo/texinfo.texi (enumerate): * texinfo/texinfo.texi (Two-column Tables): * texinfo/texinfo.texi (table): * texinfo/texinfo.texi (ftable vtable): * texinfo/texinfo.texi (itemx): * texinfo/texinfo.texi (Multi-column Tables): * texinfo/texinfo.texi (Multitable Column Widths): * texinfo/texinfo.texi (Multitable Rows): * texinfo/texinfo.texi (Special Displays): * texinfo/texinfo.texi (Floats): * texinfo/texinfo.texi (float): * texinfo/texinfo.texi (caption shortcaption): * texinfo/texinfo.texi (listoffloats): * texinfo/texinfo.texi (Images): * texinfo/texinfo.texi (Image Syntax): * texinfo/texinfo.texi (Image Scaling): * texinfo/texinfo.texi (Footnotes): * texinfo/texinfo.texi (Footnote Commands): * texinfo/texinfo.texi (Footnote Styles): * texinfo/texinfo.texi (Indices): * texinfo/texinfo.texi (Index Entries): * texinfo/texinfo.texi (Predefined Indices): * texinfo/texinfo.texi (Indexing Commands): * texinfo/texinfo.texi (Combining Indices): * texinfo/texinfo.texi (syncodeindex): * texinfo/texinfo.texi (synindex): * texinfo/texinfo.texi (New Indices): * texinfo/texinfo.texi (Insertions): * texinfo/texinfo.texi (Atsign Braces Comma): * texinfo/texinfo.texi (Inserting an Atsign): * texinfo/texinfo.texi (Inserting Braces): * texinfo/texinfo.texi (Inserting a Comma): * texinfo/texinfo.texi (Inserting Quote Characters): * texinfo/texinfo.texi (Inserting Space): * texinfo/texinfo.texi (Not Ending a Sentence): * texinfo/texinfo.texi (Ending a Sentence): * texinfo/texinfo.texi (Multiple Spaces): * texinfo/texinfo.texi (frenchspacing): * texinfo/texinfo.texi (dmn): * texinfo/texinfo.texi (Inserting Accents): * texinfo/texinfo.texi (Inserting Quotation Marks): * texinfo/texinfo.texi (Dots Bullets): * texinfo/texinfo.texi (dots): * texinfo/texinfo.texi (bullet): * texinfo/texinfo.texi (TeX and copyright): * texinfo/texinfo.texi (tex): * texinfo/texinfo.texi (copyright symbol): * texinfo/texinfo.texi (registered symbol): * texinfo/texinfo.texi (euro): * texinfo/texinfo.texi (pounds): * texinfo/texinfo.texi (textdegree): * texinfo/texinfo.texi (minus): * texinfo/texinfo.texi (geq leq): * texinfo/texinfo.texi (math): * texinfo/texinfo.texi (Click Sequences): * texinfo/texinfo.texi (Glyphs): * texinfo/texinfo.texi (Glyphs Summary): * texinfo/texinfo.texi (result): * texinfo/texinfo.texi (expansion): * texinfo/texinfo.texi (Print Glyph): * texinfo/texinfo.texi (Error Glyph): * texinfo/texinfo.texi (Equivalence): * texinfo/texinfo.texi (Point Glyph): * texinfo/texinfo.texi (Breaks): * texinfo/texinfo.texi (Break Commands): * texinfo/texinfo.texi (Line Breaks): * texinfo/texinfo.texi (- and hyphenation): * texinfo/texinfo.texi (allowcodebreaks): * texinfo/texinfo.texi (w): * texinfo/texinfo.texi (tie): * texinfo/texinfo.texi (sp): * texinfo/texinfo.texi (page): * texinfo/texinfo.texi (group): * texinfo/texinfo.texi (need): * texinfo/texinfo.texi (Definition Commands): * texinfo/texinfo.texi (Def Cmd Template): * texinfo/texinfo.texi (Def Cmd Continuation Lines): * texinfo/texinfo.texi (Optional Arguments): * texinfo/texinfo.texi (deffnx): * texinfo/texinfo.texi (Def Cmds in Detail): * texinfo/texinfo.texi (Functions Commands): * texinfo/texinfo.texi (Variables Commands): * texinfo/texinfo.texi (Typed Functions): * texinfo/texinfo.texi (Typed Variables): * texinfo/texinfo.texi (Data Types): * texinfo/texinfo.texi (Abstract Objects): * texinfo/texinfo.texi (Object-Oriented Variables): * texinfo/texinfo.texi (Object-Oriented Methods): * texinfo/texinfo.texi (Defining Macros): * texinfo/texinfo.texi (Invoking Macros): * texinfo/texinfo.texi (Macro Details): * texinfo/texinfo.texi (alias): * texinfo/texinfo.texi (definfoenclose): * texinfo/texinfo.texi (Hardcopy): * texinfo/texinfo.texi (Use TeX): * texinfo/texinfo.texi (Format with tex/texindex): * texinfo/texinfo.texi (Format with texi2dvi): * texinfo/texinfo.texi (Print with lpr): * texinfo/texinfo.texi (Within XEmacs): * texinfo/texinfo.texi (Texinfo Mode Printing): * texinfo/texinfo.texi (Compile-Command): * texinfo/texinfo.texi (Requirements Summary): * texinfo/texinfo.texi (Preparing for TeX): * texinfo/texinfo.texi (Overfull hboxes): * texinfo/texinfo.texi (smallbook): * texinfo/texinfo.texi (A4 Paper): * texinfo/texinfo.texi (pagesizes): * texinfo/texinfo.texi (Cropmarks and Magnification): * texinfo/texinfo.texi (PDF Output): * texinfo/texinfo.texi (Obtaining TeX): * texinfo/texinfo.texi (Creating and Installing Info Files): * texinfo/texinfo.texi (Creating an Info File): * texinfo/texinfo.texi (makeinfo advantages): * texinfo/texinfo.texi (Invoking makeinfo): * texinfo/texinfo.texi (makeinfo options): * texinfo/texinfo.texi (Pointer Validation): * texinfo/texinfo.texi (makeinfo in XEmacs): * texinfo/texinfo.texi (texinfo-format commands): * texinfo/texinfo.texi (Batch Formatting): * texinfo/texinfo.texi (Tag and Split Files): * texinfo/texinfo.texi (Installing an Info File): * texinfo/texinfo.texi (Directory File): * texinfo/texinfo.texi (New Info File): * texinfo/texinfo.texi (Other Info Directories): * texinfo/texinfo.texi (Installing Dir Entries): * texinfo/texinfo.texi (Invoking install-info): * texinfo/texinfo.texi (Generating HTML): * texinfo/texinfo.texi (HTML Translation): * texinfo/texinfo.texi (HTML Splitting): * texinfo/texinfo.texi (HTML CSS): * texinfo/texinfo.texi (HTML Xref): * texinfo/texinfo.texi (HTML Xref Link Basics): * texinfo/texinfo.texi (HTML Xref Node Name Expansion): * texinfo/texinfo.texi (HTML Xref Command Expansion): * texinfo/texinfo.texi (HTML Xref 8-bit Character Expansion): * texinfo/texinfo.texi (HTML Xref Mismatch): * texinfo/texinfo.texi (Command List): * texinfo/texinfo.texi (Command Syntax): * texinfo/texinfo.texi (Tips): * texinfo/texinfo.texi (Sample Texinfo Files): * texinfo/texinfo.texi (Short Sample Texinfo File): * texinfo/texinfo.texi (GNU Sample Texts): * texinfo/texinfo.texi (Invoking sample): * texinfo/texinfo.texi (GNU Free Documentation License): * texinfo/texinfo.texi (Index): * texinfo/texinfo.texi (Verbatim Copying License): * texinfo/texinfo.texi (All-permissive Copying License): * texinfo/texinfo.texi (Include Files): * texinfo/texinfo.texi (Using Include Files): * texinfo/texinfo.texi (texinfo-multiple-files-update): * texinfo/texinfo.texi (Include Files Requirements): * texinfo/texinfo.texi (Sample Include File): * texinfo/texinfo.texi (Include Files Evolution): * texinfo/texinfo.texi (Headings): * texinfo/texinfo.texi (Headings Introduced): * texinfo/texinfo.texi (Heading Format): * texinfo/texinfo.texi (Heading Choice): * texinfo/texinfo.texi (Custom Headings): * texinfo/texinfo.texi (Catching Mistakes): * texinfo/texinfo.texi (makeinfo Preferred): * texinfo/texinfo.texi (Debugging with Info): * texinfo/texinfo.texi (Debugging with TeX): * texinfo/texinfo.texi (Using texinfo-show-structure): * texinfo/texinfo.texi (Using occur): * texinfo/texinfo.texi (Running Info-Validate): * texinfo/texinfo.texi (Using Info-validate): * texinfo/texinfo.texi (Unsplit): * texinfo/texinfo.texi (Tagifying): * texinfo/texinfo.texi (Splitting): * texinfo/texinfo.texi (Refilling Paragraphs): * texinfo/texinfo.texi (Command and Variable Index): * texinfo/texinfo.texi (General Index): * texinfo/version.texi: New file. Sync with FSF 23.1.92. Make new directory to hold the files needed to generate texinfo.info, since there are three such files now. 2010-02-19 Ben Wing <ben@xemacs.org> * Makefile: * Makefile (src_files1): * Makefile (DIR): * Makefile (texinfo-srcs): * Makefile ($(INFODIR)/widget.info): * Makefile ($(INFODIR)/texinfo.info): * Makefile (.PHONY): * Makefile (texinfo.dvi): * Makefile (texinfo.pdf): * Makefile ($(HTMLDIR)/widget.html): * Makefile ($(HTMLDIR)/texinfo.html): Incorporate texinfo.texi moving to a subdirectory texinfo/. Do some tricks to reduce the amount of duplication while still maintaining compatible with non-GNU make (at least, with Solaris make). * doclicense.texi: New file. * info.texi: * info.texi (Top): * info.texi (Getting Started): * info.texi (Help-Small-Screen): * info.texi (Help): * info.texi (Help-P): * info.texi (Help-^L): * info.texi (Help-Inv): * info.texi (Help-]): * info.texi (Help-M): * info.texi (Help-FOO): * info.texi (Help-Xref): * info.texi (Help-Int): * info.texi (Help-Q): * info.texi (Advanced): * info.texi (Search Text): * info.texi (Search Index): * info.texi (Go to node): * info.texi (Choose menu subtopic): * info.texi (Create Info buffer): * info.texi (XEmacs Info Variables): * info.texi (Expert Info): * info.texi (Add): * info.texi (Menus): * info.texi (Cross-refs): * info.texi (Help-Cross): * info.texi (Tags): * info.texi (Checking): * info.texi (Index): * texinfo.tex: * texinfo.tex (paragraphindent{%): * texinfo.tex (sectionheading will have): * texinfo.tex (chapterzzz{#3}%): * texinfo.tex (subsubsection = \numberedsubsubsec): * texinfo.tex (subsubsection = \appendixsubsubsec): * texinfo.tex (subsubsection = \unnumberedsubsubsec): * texinfo.tex (sectionheading{#1}{sec}{Ynumbered}{\the\chapno.\the\secno}%): * texinfo.tex (sectionheading{#1}{sec}{Yappendix}{\appendixletter.\the\secno}%): * texinfo.tex (sectionheading{#1}{sec}{Ynothing}{\the\unnumberedno.\the\secno}%): * texinfo.tex (sectionheading{#1}{subsec}{Ynumbered}{\the\chapno.\the\secno.\the\subsecno}%): * texinfo.tex (sectionheading{#1}{subsec}{Yappendix}%): * texinfo.tex (sectionheading{#1}{subsec}{Ynothing}%): * texinfo.tex (sectionheading{#1}{subsubsec}{Ynumbered}%): * texinfo.tex (sectionheading{#1}{subsubsec}{Yappendix}%): * texinfo.tex (sectionheading{#1}{subsubsec}{Ynothing}%): * texinfo.tex (sectionheading{#1}{subsubsec}{Yomitfromtoc}{}): * texinfo.tex (sectionheading to do the printing.): * texinfo.tex (sectionlevel}{#1}{#4}%): * texinfo.tex (sectionheading, q.v.): Sync with FSF 23.1.92.

author	Ben Wing <ben@xemacs.org>
date	Fri, 19 Feb 2010 22:39:19 -0600
parents	3ecd8885ac67
children	1f0b15040456 a7da359711ce

comparison

equal deleted inserted replaced

-:06dd936cde16
+:e82f5b7010fe
 % texinfo.tex -- TeX macros to handle Texinfo files.
 %
 % Load plain if necessary, i.e., if running under initex.
 \expandafter\ifx\csname fmtname\endcsname\relax\input plain\fi
 %
-\def\texinfoversion{1999-09-25.10}
+\def\texinfoversion{2008-04-18.10}
 %
-% Copyright (C) 1985, 86, 88, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99
+% Copyright (C) 1985, 1986, 1988, 1990, 1991, 1992, 1993, 1994, 1995,
-% Free Software Foundation, Inc.
+% 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006,
-%
+% 2007, 2008 Free Software Foundation, Inc.
-% This texinfo.tex file is free software; you can redistribute it and/or
+%
+% This texinfo.tex file is free software: you can redistribute it and/or
 % modify it under the terms of the GNU General Public License as
-% published by the Free Software Foundation; either version 2, or (at
+% published by the Free Software Foundation, either version 3 of the
-% your option) any later version.
+% License, or (at your option) any later version.
 %
 % This texinfo.tex file is distributed in the hope that it will be
 % useful, but WITHOUT ANY WARRANTY; without even the implied warranty
 % of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 % General Public License for more details.
 %
 % You should have received a copy of the GNU General Public License
-% along with this texinfo.tex file; see the file COPYING.  If not, write
+% along with this program.  If not, see <http://www.gnu.org/licenses/>.
-% to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+%
-% Boston, MA 02111-1307, USA.
+% As a special exception, when this file is read by TeX when processing
-%
+% a Texinfo source document, you may use the result without
-% In other words, you are welcome to use, share and improve this program.
+% restriction.  (This has been our intent since Texinfo was invented.)
-% You are forbidden to forbid anyone else to use, share and improve
-% what you give them.   Help stamp out software-hoarding!
 %
 % Please try the latest version of texinfo.tex before submitting bug
 % reports; you can get the latest version from:
-%   ftp://ftp.gnu.org/gnu/texinfo.tex
+%   http://www.gnu.org/software/texinfo/ (the Texinfo home page), or
-%   (and all GNU mirrors, see http://www.gnu.org/order/ftp.html)
+%   ftp://tug.org/tex/texinfo.tex
-%   ftp://texinfo.org/tex/texinfo.tex
+%     (and all CTAN mirrors, see http://www.ctan.org).
-%   ftp://us.ctan.org/macros/texinfo/texinfo.tex
+% The texinfo.tex in any given distribution could well be out
-%   (and all CTAN mirrors, finger ctan@us.ctan.org for a list).
-%   /home/gd/gnu/doc/texinfo.tex on the GNU machines.
-% The texinfo.tex in any given Texinfo distribution could well be out
 % of date, so if that's what you're using, please check.
-% Texinfo has a small home page at http://texinfo.org/.
 %
 % Send bug reports to bug-texinfo@gnu.org.  Please include including a
 % complete document in each bug report with which we can reproduce the
 % problem.  Patches are, of course, greatly appreciated.
 %
 % manual foo.texi, however, you can get away with this:
 %   tex foo.texi
 %   texindex foo.??
 %   tex foo.texi
 %   tex foo.texi
-%   dvips foo.dvi -o # or whatever, to process the dvi file; this makes foo.ps.
+%   dvips foo.dvi -o  # or whatever; this makes foo.ps.
-% The extra runs of TeX get the cross-reference information correct.
+% The extra TeX runs get the cross-reference information correct.
 % Sometimes one run after texindex suffices, and sometimes you need more
 % than two; texi2dvi does it as many times as necessary.
 %
-% It is possible to adapt texinfo.tex for other languages.  You can get
+% It is possible to adapt texinfo.tex for other languages, to some
-% the existing language-specific files from ftp://ftp.gnu.org/gnu/texinfo/.
+% extent.  You can get the existing language-specific files from the
+% full Texinfo distribution.
+%
+% The GNU Texinfo home page is http://www.gnu.org/software/texinfo.
 \message{Loading texinfo [version \texinfoversion]:}
 % If in a .fmt file, print the version number
 % and turn on active characters that we couldn't do earlier because
 % they might have appeared in the input file name.
 \everyjob{\message{[Texinfo version \texinfoversion]}%
 \catcode`+=\active \catcode`\_=\active}
-% Save some parts of plain tex whose names we will redefine.
+\chardef\other=12
+% We never want plain's \outer definition of \+ in Texinfo.
+% For @tex, we can use \tabalign.
+\let\+ = \relax
+% Save some plain tex macros whose names we will redefine.
 \let\ptexb=\b
 \let\ptexbullet=\bullet
 \let\ptexc=\c
 \let\ptexcomma=\,
 \let\ptexdot=\.
 \let\ptexdots=\dots
 \let\ptexend=\end
 \let\ptexequiv=\equiv
 \let\ptexexclam=\!
+\let\ptexfootnote=\footnote
+\let\ptexgtr=>
+\let\ptexhat=^
 \let\ptexi=\i
+\let\ptexindent=\indent
+\let\ptexinsert=\insert
 \let\ptexlbrace=\{
+\let\ptexless=<
+\let\ptexnewwrite\newwrite
+\let\ptexnoindent=\noindent
+\let\ptexplus=+
 \let\ptexrbrace=\}
+\let\ptexslash=\/
 \let\ptexstar=\*
 \let\ptext=\t
+\let\ptextop=\top
-% We never want plain's outer \+ definition in Texinfo.
-% For @tex, we can use \tabalign.
-\let\+ = \relax
-\message{Basics,}
-\chardef\other=12
 % If this character appears in an error message or help string, it
 % starts a new line in the output.
 \newlinechar = `^^J
+% Use TeX 3.0's \inputlineno to get the line number, for better error
+% messages, but if we're using an old version of TeX, don't do anything.
+%
+\ifx\inputlineno\thisisundefined
+\let\linenumber = \empty % Pre-3.0.
+\else
+\def\linenumber{l.\the\inputlineno:\space}
+\fi
 % Set up fixed words for English if not already set.
 \ifx\putwordAppendix\undefined  \gdef\putwordAppendix{Appendix}\fi
 \ifx\putwordChapter\undefined   \gdef\putwordChapter{Chapter}\fi
 \ifx\putwordfile\undefined      \gdef\putwordfile{file}\fi
 %
 \ifx\putwordDefmac\undefined    \gdef\putwordDefmac{Macro}\fi
 \ifx\putwordDefspec\undefined   \gdef\putwordDefspec{Special Form}\fi
 \ifx\putwordDefvar\undefined    \gdef\putwordDefvar{Variable}\fi
 \ifx\putwordDefopt\undefined    \gdef\putwordDefopt{User Option}\fi
-\ifx\putwordDeftypevar\undefined\gdef\putwordDeftypevar{Variable}\fi
 \ifx\putwordDeffunc\undefined   \gdef\putwordDeffunc{Function}\fi
-\ifx\putwordDeftypefun\undefined\gdef\putwordDeftypefun{Function}\fi
+% Since the category of space is not known, we have to be careful.
+\chardef\spacecat = 10
+\def\spaceisspace{\catcode`\ =\spacecat}
+% sometimes characters are active, so we need control sequences.
+\chardef\colonChar = `\:
+\chardef\commaChar = `\,
+\chardef\dashChar  = `\-
+\chardef\dotChar   = `\.
+\chardef\exclamChar= `\!
+\chardef\lquoteChar= `\`
+\chardef\questChar = `\?
+\chardef\rquoteChar= `\'
+\chardef\semiChar  = `\;
+\chardef\underChar = `\_
 % Ignore a token.
 %
 \def\gobble#1{}
-\hyphenation{ap-pen-dix}
+% The following is used inside several \edef's.
-\hyphenation{mini-buf-fer mini-buf-fers}
+\def\makecsname#1{\expandafter\noexpand\csname#1\endcsname}
-\hyphenation{eshell}
-\hyphenation{white-space}
+% Hyphenation fixes.
+\hyphenation{
+Flor-i-da Ghost-script Ghost-view Mac-OS Post-Script
+ap-pen-dix bit-map bit-maps
+data-base data-bases eshell fall-ing half-way long-est man-u-script
+man-u-scripts mini-buf-fer mini-buf-fers over-view par-a-digm
+par-a-digms rath-er rec-tan-gu-lar ro-bot-ics se-vere-ly set-up spa-ces
+spell-ing spell-ings
+stand-alone strong-est time-stamp time-stamps which-ever white-space
+wide-spread wrap-around
+}
 % Margin to add to right of even pages, to left of odd pages.
-\newdimen \bindingoffset
+\newdimen\bindingoffset
-\newdimen \normaloffset
+\newdimen\normaloffset
 \newdimen\pagewidth \newdimen\pageheight
+% For a final copy, take out the rectangles
+% that mark overfull boxes (in case you have decided
+% that the text looks ok even though it passes the margin).
+%
+\def\finalout{\overfullrule=0pt}
+% @| inserts a changebar to the left of the current line.  It should
+% surround any changed text.  This approach does *not* work if the
+% change spans more than two lines of output.  To handle that, we would
+% have adopt a much more difficult approach (putting marks into the main
+% vertical list for the beginning and end of each change).
+%
+\def\|{%
+% \vadjust can only be used in horizontal mode.
+\leavevmode
+%
+% Append this vertical mode material after the current line in the output.
+\vadjust{%
+% We want to insert a rule with the height and depth of the current
+% leading; that is exactly what \strutbox is supposed to record.
+\vskip-\baselineskip
+%
+% \vadjust-items are inserted at the left edge of the type.  So
+% the \llap here moves out into the left-hand margin.
+\llap{%
+%
+% For a thicker or thinner bar, change the `1pt'.
+\vrule height\baselineskip width1pt
+%
+% This is the space between the bar and the text.
+\hskip 12pt
+}%
+}%
+}
 % Sometimes it is convenient to have everything in the transcript file
 % and nothing on the terminal.  We don't just call \tracingall here,
-% since that produces some useless output on the terminal.
+% since that produces some useless output on the terminal.  We also make
+% some effort to order the tracing commands to reduce output in the log
+% file; cf. trace.sty in LaTeX.
 %
 \def\gloggingall{\begingroup \globaldefs = 1 \loggingall \endgroup}%
-\ifx\eTeXversion\undefined
+\def\loggingall{%
-\def\loggingall{\tracingcommands2 \tracingstats2
+\tracingstats2
-\tracingpages1 \tracingoutput1 \tracinglostchars1
+\tracingpages1
-\tracingmacros2 \tracingparagraphs1 \tracingrestores1
+\tracinglostchars2  % 2 gives us more in etex
-\showboxbreadth\maxdimen\showboxdepth\maxdimen
+\tracingparagraphs1
+\tracingoutput1
+\tracingmacros2
+\tracingrestores1
+\showboxbreadth\maxdimen \showboxdepth\maxdimen
+\ifx\eTeXversion\undefined\else % etex gives us more logging
+\tracingscantokens1
+\tracingifs1
+\tracinggroups1
+\tracingnesting2
+\tracingassigns1
+\fi
+\tracingcommands3  % 3 gives us more in etex
+\errorcontextlines16
 }%
-\else
-\def\loggingall{\tracingcommands3 \tracingstats2
+% add check for \lastpenalty to plain's definitions.  If the last thing
-\tracingpages1 \tracingoutput1 \tracinglostchars1
+% we did was a \nobreak, we don't want to insert more space.
-\tracingmacros2 \tracingparagraphs1 \tracingrestores1
+%
-\tracingscantokens1 \tracingassigns1 \tracingifs1
+\def\smallbreak{\ifnum\lastpenalty<10000\par\ifdim\lastskip<\smallskipamount
-\tracinggroups1 \tracingnesting2
+\removelastskip\penalty-50\smallskip\fi\fi}
-\showboxbreadth\maxdimen\showboxdepth\maxdimen
+\def\medbreak{\ifnum\lastpenalty<10000\par\ifdim\lastskip<\medskipamount
-}%
+\removelastskip\penalty-100\medskip\fi\fi}
-\fi
+\def\bigbreak{\ifnum\lastpenalty<10000\par\ifdim\lastskip<\bigskipamount
+\removelastskip\penalty-200\bigskip\fi\fi}
 % For @cropmarks command.
 % Do @cropmarks to get crop marks.
 %
 \newif\ifcropmarks
 \newdimen\outerhsize \newdimen\outervsize % set by the paper size routines
 \newdimen\cornerlong  \cornerlong=1pc
 \newdimen\cornerthick \cornerthick=.3pt
 \newdimen\topandbottommargin \topandbottommargin=.75in
+% Output a mark which sets \thischapter, \thissection and \thiscolor.
+% We dump everything together because we only have one kind of mark.
+% This works because we only use \botmark / \topmark, not \firstmark.
+%
+% A mark contains a subexpression of the \ifcase ... \fi construct.
+% \get*marks macros below extract the needed part using \ifcase.
+%
+% Another complication is to let the user choose whether \thischapter
+% (\thissection) refers to the chapter (section) in effect at the top
+% of a page, or that at the bottom of a page.  The solution is
+% described on page 260 of The TeXbook.  It involves outputting two
+% marks for the sectioning macros, one before the section break, and
+% one after.  I won't pretend I can describe this better than DEK...
+\def\domark{%
+\toks0=\expandafter{\lastchapterdefs}%
+\toks2=\expandafter{\lastsectiondefs}%
+\toks4=\expandafter{\prevchapterdefs}%
+\toks6=\expandafter{\prevsectiondefs}%
+\toks8=\expandafter{\lastcolordefs}%
+\mark{%
+\the\toks0 \the\toks2
+\noexpand\or \the\toks4 \the\toks6
+\noexpand\else \the\toks8
+}%
+}
+% \topmark doesn't work for the very first chapter (after the title
+% page or the contents), so we use \firstmark there -- this gets us
+% the mark with the chapter defs, unless the user sneaks in, e.g.,
+% @setcolor (or @url, or @link, etc.) between @contents and the very
+% first @chapter.
+\def\gettopheadingmarks{%
+\ifcase0\topmark\fi
+\ifx\thischapter\empty \ifcase0\firstmark\fi \fi
+}
+\def\getbottomheadingmarks{\ifcase1\botmark\fi}
+\def\getcolormarks{\ifcase2\topmark\fi}
+% Avoid "undefined control sequence" errors.
+\def\lastchapterdefs{}
+\def\lastsectiondefs{}
+\def\prevchapterdefs{}
+\def\prevsectiondefs{}
+\def\lastcolordefs{}
 % Main output routine.
 \chardef\PAGE = 255
 \output = {\onepageout{\pagecontents\PAGE}}
 \newbox\headlinebox
 \ifodd\pageno  \advance\hoffset by \bindingoffset
 \else \advance\hoffset by -\bindingoffset\fi
 %
 % Do this outside of the \shipout so @code etc. will be expanded in
 % the headline as they should be, not taken literally (outputting ''code).
+\ifodd\pageno \getoddheadingmarks \else \getevenheadingmarks \fi
 \setbox\headlinebox = \vbox{\let\hsize=\pagewidth \makeheadline}%
+\ifodd\pageno \getoddfootingmarks \else \getevenfootingmarks \fi
 \setbox\footlinebox = \vbox{\let\hsize=\pagewidth \makefootline}%
 %
 {%
 % Have to do this stuff outside the \shipout because we want it to
 % take effect in \write's, yet the group defined by the \vbox ends
 % before the \shipout runs.
 %
-\escapechar = `\\     % use backslash in output files.
 \indexdummies         % don't expand commands in the output.
 \normalturnoffactive  % \ in index entries must not stay \, e.g., if
 % the page break happens to be in the middle of an example.
+% We don't want .vr (or whatever) entries like this:
+% \entry{{\tt \indexbackslash }acronym}{32}{\code {\acronym}}
+% "\acronym" won't work when it's read back in;
+% it needs to be
+% {\code {{\tt \backslashcurfont }acronym}
 \shipout\vbox{%
+% Do this early so pdf references go to the beginning of the page.
+\ifpdfmakepagedest \pdfdest name{\the\pageno} xyz\fi
+%
 \ifcropmarks \vbox to \outervsize\bgroup
 \hsize = \outerhsize
 \vskip-\topandbottommargin
 \vtop to0pt{%
 \line{\ewtop\hfil\ewtop}%
 %
 \unvbox\headlinebox
 \pagebody{#1}%
 \ifdim\ht\footlinebox > 0pt
 % Only leave this space if the footline is nonempty.
-% (We lessened \vsize for it in \oddfootingxxx.)
+% (We lessened \vsize for it in \oddfootingyyy.)
 % The \baselineskip=24pt in plain's \makefootline has no effect.
-\vskip 2\baselineskip
+\vskip 24pt
 \unvbox\footlinebox
 \fi
-%
-\ifpdfmakepagedest \pdfmkdest{\the\pageno} \fi
 %
 \ifcropmarks
 \egroup % end of \vbox\bgroup
 \hfil\egroup % end of (centering) \line\bgroup
 \vskip\topandbottommargin plus1fill minus1fill
 \line{\ewbot\hfil\ewbot}%
 }%
 \egroup % \vbox from first cropmarks clause
 \fi
 }% end of \shipout\vbox
-}% end of group with \turnoffactive
+}% end of group with \indexdummies
 \advancepageno
 \ifnum\outputpenalty>-20000 \else\dosupereject\fi
 }
 \newinsert\margin \dimen\margin=\maxdimen
 {\catcode`\@ =11
 \gdef\pagecontents#1{\ifvoid\topins\else\unvbox\topins\fi
 % marginal hacks, juha@viisa.uucp (Juha Takala)
 \ifvoid\margin\else % marginal info is present
 \rlap{\kern\hsize\vbox to\z@{\kern1pt\box\margin \vss}}\fi
-\dimen@=\dp#1 \unvbox#1
+\dimen@=\dp#1\relax \unvbox#1\relax
 \ifvoid\footins\else\vskip\skip\footins\footnoterule \unvbox\footins\fi
 \ifr@ggedbottom \kern-\dimen@ \vfil \fi}
 }
 % Here are the rules for the cropmarks.  Note that they are
 % Parse an argument, then pass it to #1.  The argument is the rest of
 % the input line (except we remove a trailing comment).  #1 should be a
 % macro which expects an ordinary undelimited TeX argument.
 %
-\def\parsearg#1{%
+\def\parsearg{\parseargusing{}}
-\let\next = #1%
+\def\parseargusing#1#2{%
+\def\argtorun{#2}%
 \begingroup
 \obeylines
-\futurelet\temp\parseargx
+\spaceisspace
-}
+#1%
+\parseargline\empty% Insert the \empty token, see \finishparsearg below.
-% If the next token is an obeyed space (from an @example environment or
+}
-% the like), remove it and recurse.  Otherwise, we're done.
-\def\parseargx{%
-% \obeyedspace is defined far below, after the definition of \sepspaces.
-\ifx\obeyedspace\temp
-\expandafter\parseargdiscardspace
-\else
-\expandafter\parseargline
-\fi
-}
-% Remove a single space (as the delimiter token to the macro call).
-{\obeyspaces %
-\gdef\parseargdiscardspace {\futurelet\temp\parseargx}}
 {\obeylines %
 \gdef\parseargline#1^^M{%
 \endgroup % End of the group started in \parsearg.
-%
+\argremovecomment #1\comment\ArgTerm%
-% First remove any @c comment, then any @comment.
-% Result of each macro is put in \toks0.
-\argremovec #1\c\relax %
-\expandafter\argremovecomment \the\toks0 \comment\relax %
-%
-% Call the caller's macro, saved as \next in \parsearg.
-\expandafter\next\expandafter{\the\toks0}%
 }%
 }
-% Since all \c{,omment} does is throw away the argument, we can let TeX
+% First remove any @comment, then any @c comment.
-% do that for us.  The \relax here is matched by the \relax in the call
+\def\argremovecomment#1\comment#2\ArgTerm{\argremovec #1\c\ArgTerm}
-% in \parseargline; it could be more or less anything, its purpose is
+\def\argremovec#1\c#2\ArgTerm{\argcheckspaces#1\^^M\ArgTerm}
-% just to delimit the argument to the \c.
-\def\argremovec#1\c#2\relax{\toks0 = {#1}}
+% Each occurrence of `\^^M' or `<space>\^^M' is replaced by a single space.
-\def\argremovecomment#1\comment#2\relax{\toks0 = {#1}}
+%
+% \argremovec might leave us with trailing space, e.g.,
-% \argremovec{,omment} might leave us with trailing spaces, though; e.g.,
 %    @end itemize  @c foo
-% will have two active spaces as part of the argument with the
+% This space token undergoes the same procedure and is eventually removed
-% `itemize'.  Here we remove all active spaces from #1, and assign the
+% by \finishparsearg.
-% result to \toks0.
+%
-%
+\def\argcheckspaces#1\^^M{\argcheckspacesX#1\^^M \^^M}
-% This loses if there are any *other* active characters besides spaces
+\def\argcheckspacesX#1 \^^M{\argcheckspacesY#1\^^M}
-% in the argument -- _ ^ +, for example -- since they get expanded.
+\def\argcheckspacesY#1\^^M#2\^^M#3\ArgTerm{%
-% Fortunately, Texinfo does not define any such commands.  (If it ever
+\def\temp{#3}%
-% does, the catcode of the characters in questionwill have to be changed
+\ifx\temp\empty
-% here.)  But this means we cannot call \removeactivespaces as part of
+% Do not use \next, perhaps the caller of \parsearg uses it; reuse \temp:
-% \argremovec{,omment}, since @c uses \parsearg, and thus the argument
+\let\temp\finishparsearg
-% that \parsearg gets might well have any character at all in it.
+\else
-%
+\let\temp\argcheckspaces
-\def\removeactivespaces#1{%
+\fi
-\begingroup
+% Put the space token in:
-\ignoreactivespaces
+\temp#1 #3\ArgTerm
-\edef\temp{#1}%
+}
-\global\toks0 = \expandafter{\temp}%
-\endgroup
+% If a _delimited_ argument is enclosed in braces, they get stripped; so
-}
+% to get _exactly_ the rest of the line, we had to prevent such situation.
+% We prepended an \empty token at the very beginning and we expand it now,
-% Change the active space to expand to nothing.
+% just before passing the control to \argtorun.
-%
+% (Similarly, we have to think about #3 of \argcheckspacesY above: it is
-\begingroup
+% either the null string, or it ends with \^^M---thus there is no danger
+% that a pair of braces would be stripped.
+%
+% But first, we have to remove the trailing space token.
+%
+\def\finishparsearg#1 \ArgTerm{\expandafter\argtorun\expandafter{#1}}
+% \parseargdef\foo{...}
+%	is roughly equivalent to
+% \def\foo{\parsearg\Xfoo}
+% \def\Xfoo#1{...}
+%
+% Actually, I use \csname\string\foo\endcsname, ie. \\foo, as it is my
+% favourite TeX trick.  --kasal, 16nov03
+\def\parseargdef#1{%
+\expandafter \doparseargdef \csname\string#1\endcsname #1%
+}
+\def\doparseargdef#1#2{%
+\def#2{\parsearg#1}%
+\def#1##1%
+}
+% Several utility definitions with active space:
+{
 \obeyspaces
-\gdef\ignoreactivespaces{\obeyspaces\let =\empty}
+\gdef\obeyedspace{ }
-\endgroup
+% Make each space character in the input produce a normal interword
+% space in the output.  Don't allow a line break at this space, as this
+% is used only in environments like @example, where each line of input
+% should produce a line of output anyway.
+%
+\gdef\sepspaces{\obeyspaces\let =\tie}
+% If an index command is used in an @example environment, any spaces
+% therein should become regular spaces in the raw index file, not the
+% expansion of \tie (\leavevmode \penalty \@M \ ).
+\gdef\unsepspaces{\let =\space}
+}
 \def\flushcr{\ifx\par\lisppar \def\next##1{}\else \let\next=\relax \fi \next}
-%% These are used to keep @begin/@end levels from running away
+% Define the framework for environments in texinfo.tex.  It's used like this:
-%% Call \inENV within environments (after a \begingroup)
+%
-\newif\ifENV \ENVfalse \def\inENV{\ifENV\relax\else\ENVtrue\fi}
+%   \envdef\foo{...}
-\def\ENVcheck{%
+%   \def\Efoo{...}
-\ifENV\errmessage{Still within an environment; press RETURN to continue}
+%
-\endgroup\fi} % This is not perfect, but it should reduce lossage
+% It's the responsibility of \envdef to insert \begingroup before the
+% actual body; @end closes the group after calling \Efoo.  \envdef also
-% @begin foo  is the same as @foo, for now.
+% defines \thisenv, so the current environment is known; @end checks
+% whether the environment name matches.  The \checkenv macro can also be
+% used to check whether the current environment is the one expected.
+%
+% Non-false conditionals (@iftex, @ifset) don't fit into this, so they
+% are not treated as environments; they don't open a group.  (The
+% implementation of @end takes care not to call \endgroup in this
+% special case.)
+% At run-time, environments start with this:
+\def\startenvironment#1{\begingroup\def\thisenv{#1}}
+% initialize
+\let\thisenv\empty
+% ... but they get defined via ``\envdef\foo{...}'':
+\long\def\envdef#1#2{\def#1{\startenvironment#1#2}}
+\def\envparseargdef#1#2{\parseargdef#1{\startenvironment#1#2}}
+% Check whether we're in the right environment:
+\def\checkenv#1{%
+\def\temp{#1}%
+\ifx\thisenv\temp
+\else
+\badenverr
+\fi
+}
+% Environment mismatch, #1 expected:
+\def\badenverr{%
+\errhelp = \EMsimple
+\errmessage{This command can appear only \inenvironment\temp,
+not \inenvironment\thisenv}%
+}
+\def\inenvironment#1{%
+\ifx#1\empty
+out of any environment%
+\else
+in environment \expandafter\string#1%
+\fi
+}
+% @end foo executes the definition of \Efoo.
+% But first, it executes a specialized version of \checkenv
+%
+\parseargdef\end{%
+\if 1\csname iscond.#1\endcsname
+\else
+% The general wording of \badenverr may not be ideal, but... --kasal, 06nov03
+\expandafter\checkenv\csname#1\endcsname
+\csname E#1\endcsname
+\endgroup
+\fi
+}
 \newhelp\EMsimple{Press RETURN to continue.}
-\outer\def\begin{\parsearg\beginxxx}
-\def\beginxxx #1{%
-\expandafter\ifx\csname #1\endcsname\relax
-{\errhelp=\EMsimple \errmessage{Undefined command @begin #1}}\else
-\csname #1\endcsname\fi}
-% @end foo executes the definition of \Efoo.
-%
-\def\end{\parsearg\endxxx}
-\def\endxxx #1{%
-\removeactivespaces{#1}%
-\edef\endthing{\the\toks0}%
-%
-\expandafter\ifx\csname E\endthing\endcsname\relax
-\expandafter\ifx\csname \endthing\endcsname\relax
-% There's no \foo, i.e., no ``environment'' foo.
-\errhelp = \EMsimple
-\errmessage{Undefined command `@end \endthing'}%
-\else
-\unmatchedenderror\endthing
-\fi
-\else
-% Everything's ok; the right environment has been started.
-\csname E\endthing\endcsname
-\fi
-}
-% There is an environment #1, but it hasn't been started.  Give an error.
-%
-\def\unmatchedenderror#1{%
-\errhelp = \EMsimple
-\errmessage{This `@end #1' doesn't have a matching `@#1'}%
-}
-% Define the control sequence \E#1 to give an unmatched @end error.
-%
-\def\defineunmatchedend#1{%
-\expandafter\def\csname E#1\endcsname{\unmatchedenderror{#1}}%
-}
-% Single-spacing is done by various environments (specifically, in
-% \nonfillstart and \quotations).
-\newskip\singlespaceskip \singlespaceskip = 12.5pt
-\def\singlespace{%
-% Why was this kern here?  It messes up equalizing space above and below
-% environments.  --karl, 6may93
-%{\advance \baselineskip by -\singlespaceskip
-%\kern \baselineskip}%
-\setleading \singlespaceskip
-}
 %% Simple single-character @ commands
 % @@ prints an @
 % Kludge this until the fonts are right (grr).
 \def\mylbrace {{\tt\char123}}
 \def\myrbrace {{\tt\char125}}
 \let\{=\mylbrace
 \let\}=\myrbrace
 \begingroup
-% Definitions to produce actual \{ & \} command in an index.
+% Definitions to produce \{ and \} commands for indices,
-\catcode`\{ = 12 \catcode`\} = 12
+% and @{ and @} for the aux/toc files.
+\catcode`\{ = \other \catcode`\} = \other
 \catcode`\[ = 1 \catcode`\] = 2
-\catcode`\@ = 0 \catcode`\\ = 12
+\catcode`\! = 0 \catcode`\\ = \other
-@gdef@lbracecmd[\{]%
+!gdef!lbracecmd[\{]%
-@gdef@rbracecmd[\}]%
+!gdef!rbracecmd[\}]%
-@endgroup
+!gdef!lbraceatcmd[@{]%
+!gdef!rbraceatcmd[@}]%
+!endgroup
+% @comma{} to avoid , parsing problems.
+\let\comma = ,
 % Accents: @, @dotaccent @ringaccent @ubaraccent @udotaccent
-% Others are defined by plain TeX: @` @' @" @^ @~ @= @v @H.
+% Others are defined by plain TeX: @` @' @" @^ @~ @= @u @v @H.
 \let\, = \c
 \let\dotaccent = \.
 \def\ringaccent#1{{\accent23 #1}}
 \let\tieaccent = \t
 \let\ubaraccent = \b
 \let\udotaccent = \d
-% Other special characters: @questiondown @exclamdown
+% Other special characters: @questiondown @exclamdown @ordf @ordm
-% Plain TeX defines: @AA @AE @O @OE @L (and lowercase versions) @ss.
+% Plain TeX defines: @AA @AE @O @OE @L (plus lowercase versions) @ss.
 \def\questiondown{?`}
 \def\exclamdown{!`}
+\def\ordf{\leavevmode\raise1ex\hbox{\selectfonts\lllsize \underbar{a}}}
+\def\ordm{\leavevmode\raise1ex\hbox{\selectfonts\lllsize \underbar{o}}}
 % Dotless i and dotless j, used for accents.
 \def\imacro{i}
 \def\jmacro{j}
 \def\dotless#1{%
 \def\temp{#1}%
-\ifx\temp\imacro \ptexi
+\ifx\temp\imacro \ifmmode\imath \else\ptexi \fi
-\else\ifx\temp\jmacro \j
+\else\ifx\temp\jmacro \ifmmode\jmath \else\j \fi
 \else \errmessage{@dotless can be used only with i or j}%
 \fi\fi
+}
+% The \TeX{} logo, as in plain, but resetting the spacing so that a
+% period following counts as ending a sentence.  (Idea found in latex.)
+%
+\edef\TeX{\TeX \spacefactor=1000 }
+% @LaTeX{} logo.  Not quite the same results as the definition in
+% latex.ltx, since we use a different font for the raised A; it's most
+% convenient for us to use an explicitly smaller font, rather than using
+% the \scriptstyle font (since we don't reset \scriptstyle and
+% \scriptscriptstyle).
+%
+\def\LaTeX{%
+L\kern-.36em
+{\setbox0=\hbox{T}%
+\vbox to \ht0{\hbox{\selectfonts\lllsize A}\vss}}%
+\kern-.15em
+\TeX
 }
 % Be sure we're in horizontal mode when doing a tie, since we make space
 % equivalent to this in @example-like environments. Otherwise, a space
 % at the beginning of a line will start with \penalty -- and
 \def\:{\spacefactor=1000 }
 % @* forces a line break.
 \def\*{\hfil\break\hbox{}\ignorespaces}
+% @/ allows a line break.
+\let\/=\allowbreak
 % @. is an end-of-sentence period.
-\def\.{.\spacefactor=3000 }
+\def\.{.\spacefactor=\endofsentencespacefactor\space}
 % @! is an end-of-sentence bang.
-\def\!{!\spacefactor=3000 }
+\def\!{!\spacefactor=\endofsentencespacefactor\space}
 % @? is an end-of-sentence query.
-\def\?{?\spacefactor=3000 }
+\def\?{?\spacefactor=\endofsentencespacefactor\space}
+% @frenchspacing on|off  says whether to put extra space after punctuation.
+%
+\def\onword{on}
+\def\offword{off}
+%
+\parseargdef\frenchspacing{%
+\def\temp{#1}%
+\ifx\temp\onword \plainfrenchspacing
+\else\ifx\temp\offword \plainnonfrenchspacing
+\else
+\errhelp = \EMsimple
+\errmessage{Unknown @frenchspacing option `\temp', must be on/off}%
+\fi\fi
+}
 % @w prevents a word break.  Without the \leavevmode, @w at the
 % beginning of a paragraph, when TeX is still in vertical mode, would
 % produce a whole line of output instead of starting the paragraph.
 \def\w#1{\leavevmode\hbox{#1}}
 % \topskip (p.114 of the TeXbook), the glue inserted is
 % max (\topskip - \ht (first item), 0).  If that height is large,
 % therefore, no glue is inserted, and the space between the headline and
 % the text is small, which looks bad.
 %
-\def\group{\begingroup
+% Another complication is that the group might be very large.  This can
-\ifnum\catcode13=\active \else
+% cause the glue on the previous page to be unduly stretched, because it
+% does not have much material.  In this case, it's better to add an
+% explicit \vfill so that the extra space is at the bottom.  The
+% threshold for doing this is if the group is more than \vfilllimit
+% percent of a page (\vfilllimit can be changed inside of @tex).
+%
+\newbox\groupbox
+\def\vfilllimit{0.7}
+%
+\envdef\group{%
+\ifnum\catcode`\^^M=\active \else
 \errhelp = \groupinvalidhelp
 \errmessage{@group invalid in context where filling is enabled}%
 \fi
-%
+\startsavinginserts
-% The \vtop we start below produces a box with normal height and large
+%
-% depth; thus, TeX puts \baselineskip glue before it, and (when the
+\setbox\groupbox = \vtop\bgroup
-% next line of text is done) \lineskip glue after it.  (See p.82 of
-% the TeXbook.)  Thus, space below is not quite equal to space
-% above.  But it's pretty close.
-\def\Egroup{%
-\egroup           % End the \vtop.
-\endgroup         % End the \group.
-}%
-%
-\vtop\bgroup
-% We have to put a strut on the last line in case the @group is in
-% the midst of an example, rather than completely enclosing it.
-% Otherwise, the interline space between the last line of the group
-% and the first line afterwards is too small.  But we can't put the
-% strut in \Egroup, since there it would be on a line by itself.
-% Hence this just inserts a strut at the beginning of each line.
-\everypar = {\strut}%
-%
-% Since we have a strut on every line, we don't need any of TeX's
-% normal interline spacing.
-\offinterlineskip
-%
-% OK, but now we have to do something about blank
-% lines in the input in @example-like environments, which normally
-% just turn into \lisppar, which will insert no space now that we've
-% turned off the interline space.  Simplest is to make them be an
-% empty paragraph.
-\ifx\par\lisppar
-\edef\par{\leavevmode \par}%
-%
-% Reset ^^M's definition to new definition of \par.
-\obeylines
-\fi
-%
 % Do @comment since we are called inside an environment such as
 % @example, where each end-of-line in the input causes an
 % end-of-line in the output.  We don't want the end-of-line after
 % the `@group' to put extra space in the output.  Since @group
 % should appear on a line by itself (according to the Texinfo
 % manual), we don't worry about eating any user text.
 \comment
 }
 %
+% The \vtop produces a box with normal height and large depth; thus, TeX puts
+% \baselineskip glue before it, and (when the next line of text is done)
+% \lineskip glue after it.  Thus, space below is not quite equal to space
+% above.  But it's pretty close.
+\def\Egroup{%
+% To get correct interline space between the last line of the group
+% and the first line afterwards, we have to propagate \prevdepth.
+\endgraf % Not \par, as it may have been set to \lisppar.
+\global\dimen1 = \prevdepth
+\egroup           % End the \vtop.
+% \dimen0 is the vertical size of the group's box.
+\dimen0 = \ht\groupbox  \advance\dimen0 by \dp\groupbox
+% \dimen2 is how much space is left on the page (more or less).
+\dimen2 = \pageheight   \advance\dimen2 by -\pagetotal
+% if the group doesn't fit on the current page, and it's a big big
+% group, force a page break.
+\ifdim \dimen0 > \dimen2
+\ifdim \pagetotal < \vfilllimit\pageheight
+\page
+\fi
+\fi
+\box\groupbox
+\prevdepth = \dimen1
+\checkinserts
+}
+%
 % TeX puts in an \escapechar (i.e., `@') at the beginning of the help
 % message, so this ends up printing `@group can only ...'.
 %
 \newhelp\groupinvalidhelp{%
 group can only be used in environments such as @example,^^J%
 % @need space-in-mils
 % forces a page break if there is not space-in-mils remaining.
 \newdimen\mil  \mil=0.001in
-\def\need{\parsearg\needx}
 % Old definition--didn't work.
-%\def\needx #1{\par %
+%\parseargdef\need{\par %
 %% This method tries to make TeX break the page naturally
 %% if the depth of the box does not fit.
 %{\baselineskip=0pt%
 %\vtop to #1\mil{\vfil}\kern -#1\mil\nobreak
 %\prevdepth=-1000pt
 %}}
-\def\needx#1{%
+\parseargdef\need{%
 % Ensure vertical mode, so we don't make a big box in the middle of a
 % paragraph.
 \par
 %
 % If the @need value is less than one line space, it's useless.
 % Do not allow a page break right after this kern.
 \nobreak
 \fi
 }
-% @br   forces paragraph break
+% @br   forces paragraph break (and is undocumented).
 \let\br = \par
-% @dots{} output an ellipsis using the current font.
+% @page forces the start of a new page.
-% We do .5em per period so that it has the same spacing in a typewriter
-% font as three actual period characters.
-%
-\def\dots{%
-\leavevmode
-\hbox to 1.5em{%
-\hskip 0pt plus 0.25fil minus 0.25fil
-.\hss.\hss.%
-\hskip 0pt plus 0.5fil minus 0.5fil
-}%
-}
-% @enddots{} is an end-of-sentence ellipsis.
-%
-\def\enddots{%
-\leavevmode
-\hbox to 2em{%
-\hskip 0pt plus 0.25fil minus 0.25fil
-.\hss.\hss.\hss.%
-\hskip 0pt plus 0.5fil minus 0.5fil
-}%
-\spacefactor=3000
-}
-% @page    forces the start of a new page
 %
 \def\page{\par\vfill\supereject}
 % @exdent text....
 % outputs text on separate line in roman font, starting at standard page margin
 % This records the amount of indent in the innermost environment.
 % That's how much \exdent should take out.
 \newskip\exdentamount
 % This defn is used inside fill environments such as @defun.
-\def\exdent{\parsearg\exdentyyy}
+\parseargdef\exdent{\hfil\break\hbox{\kern -\exdentamount{\rm#1}}\hfil\break}
-\def\exdentyyy #1{{\hfil\break\hbox{\kern -\exdentamount{\rm#1}}\hfil\break}}
 % This defn is used inside nofill environments such as @example.
-\def\nofillexdent{\parsearg\nofillexdentyyy}
+\parseargdef\nofillexdent{{\advance \leftskip by -\exdentamount
-\def\nofillexdentyyy #1{{\advance \leftskip by -\exdentamount
+\leftline{\hskip\leftskip{\rm#1}}}}
-\leftline{\hskip\leftskip{\rm#1}}}}
+% @inmargin{WHICH}{TEXT} puts TEXT in the WHICH margin next to the current
-% @inmargin{TEXT} puts TEXT in the margin next to the current paragraph.
+% paragraph.  For more general purposes, use the \margin insertion
+% class.  WHICH is `l' or `r'.
-\def\inmargin#1{%
+%
-\strut\vadjust{\nobreak\kern-\strutdepth
-\vtop to \strutdepth{\baselineskip\strutdepth\vss
-\llap{\rightskip=\inmarginspacing \vbox{\noindent #1}}\null}}}
 \newskip\inmarginspacing \inmarginspacing=1cm
 \def\strutdepth{\dp\strutbox}
+%
-%\hbox{{\rm#1}}\hfil\break}}
+\def\doinmargin#1#2{\strut\vadjust{%
+\nobreak
-% @include file    insert text of that file as input.
+\kern-\strutdepth
-% Allow normal characters that  we make active in the argument (a file name).
+\vtop to \strutdepth{%
-\def\include{\begingroup
+\baselineskip=\strutdepth
-\catcode`\\=12
+\vss
-\catcode`~=12
+% if you have multiple lines of stuff to put here, you'll need to
-\catcode`^=12
+% make the vbox yourself of the appropriate size.
-\catcode`_=12
+\ifx#1l%
-\catcode`|=12
+\llap{\ignorespaces #2\hskip\inmarginspacing}%
-\catcode`<=12
+\else
-\catcode`>=12
+\rlap{\hskip\hsize \hskip\inmarginspacing \ignorespaces #2}%
-\catcode`+=12
+\fi
-\parsearg\includezzz}
+\null
-% Restore active chars for included file.
+}%
-\def\includezzz#1{\endgroup\begingroup
+}}
-% Read the included file in a group so nested @include's work.
+\def\inleftmargin{\doinmargin l}
+\def\inrightmargin{\doinmargin r}
+%
+% @inmargin{TEXT [, RIGHT-TEXT]}
+% (if RIGHT-TEXT is given, use TEXT for left page, RIGHT-TEXT for right;
+% else use TEXT for both).
+%
+\def\inmargin#1{\parseinmargin #1,,\finish}
+\def\parseinmargin#1,#2,#3\finish{% not perfect, but better than nothing.
+\setbox0 = \hbox{\ignorespaces #2}%
+\ifdim\wd0 > 0pt
+\def\lefttext{#1}%  have both texts
+\def\righttext{#2}%
+\else
+\def\lefttext{#1}%  have only one text
+\def\righttext{#1}%
+\fi
+%
+\ifodd\pageno
+\def\temp{\inrightmargin\righttext}% odd page -> outside is right margin
+\else
+\def\temp{\inleftmargin\lefttext}%
+\fi
+\temp
+}
+% @include FILE -- \input text of FILE.
+%
+\def\include{\parseargusing\filenamecatcodes\includezzz}
+\def\includezzz#1{%
+\pushthisfilestack
 \def\thisfile{#1}%
-\input\thisfile
+{%
-\endgroup}
+\makevalueexpandable  % we want to expand any @value in FILE.
+\turnoffactive        % and allow special characters in the expansion
+\edef\temp{\noexpand\input #1 }%
+%
+% This trickery is to read FILE outside of a group, in case it makes
+% definitions, etc.
+\expandafter
+}\temp
+\popthisfilestack
+}
+\def\filenamecatcodes{%
+\catcode`\\=\other
+\catcode`~=\other
+\catcode`^=\other
+\catcode`_=\other
+\catcode`|=\other
+\catcode`<=\other
+\catcode`>=\other
+\catcode`+=\other
+\catcode`-=\other
+}
+\def\pushthisfilestack{%
+\expandafter\pushthisfilestackX\popthisfilestack\StackTerm
+}
+\def\pushthisfilestackX{%
+\expandafter\pushthisfilestackY\thisfile\StackTerm
+}
+\def\pushthisfilestackY #1\StackTerm #2\StackTerm {%
+\gdef\popthisfilestack{\gdef\thisfile{#1}\gdef\popthisfilestack{#2}}%
+}
+\def\popthisfilestack{\errthisfilestackempty}
+\def\errthisfilestackempty{\errmessage{Internal error:
+the stack of filenames is empty.}}
 \def\thisfile{}
-% @center line   outputs that line, centered
+% @center line
+% outputs that line, centered.
-\def\center{\parsearg\centerzzz}
+%
-\def\centerzzz #1{{\advance\hsize by -\leftskip
+\parseargdef\center{%
-\advance\hsize by -\rightskip
+\ifhmode
-\centerline{#1}}}
+\let\next\centerH
+\else
+\let\next\centerV
+\fi
+\next{\hfil \ignorespaces#1\unskip \hfil}%
+}
+\def\centerH#1{%
+{%
+\hfil\break
+\advance\hsize by -\leftskip
+\advance\hsize by -\rightskip
+\line{#1}%
+\break
+}%
+}
+\def\centerV#1{\line{\kern\leftskip #1\kern\rightskip}}
 % @sp n   outputs n lines of vertical space
-\def\sp{\parsearg\spxxx}
+\parseargdef\sp{\vskip #1\baselineskip}
-\def\spxxx #1{\vskip #1\baselineskip}
 % @comment ...line which is ignored...
 % @c is the same as @comment
 % @ignore ... @end ignore  is another way to write a comment
 \let\c=\comment
 % @paragraphindent NCHARS
 % We'll use ems for NCHARS, close enough.
-% We cannot implement @paragraphindent asis, though.
+% NCHARS can also be the word `asis' or `none'.
-%
+% We cannot feasibly implement @paragraphindent asis, though.
+%
 \def\asisword{asis} % no translation, these are keywords
 \def\noneword{none}
 %
-\def\paragraphindent{\parsearg\doparagraphindent}
+\parseargdef\paragraphindent{%
-\def\doparagraphindent#1{%
 \def\temp{#1}%
 \ifx\temp\asisword
 \else
 \ifx\temp\noneword
 \defaultparindent = 0pt
 % @exampleindent NCHARS
 % We'll use ems for NCHARS like @paragraphindent.
 % It seems @exampleindent asis isn't necessary, but
 % I preserve it to make it similar to @paragraphindent.
-\def\exampleindent{\parsearg\doexampleindent}
+\parseargdef\exampleindent{%
-\def\doexampleindent#1{%
 \def\temp{#1}%
 \ifx\temp\asisword
 \else
 \ifx\temp\noneword
 \lispnarrowing = 0pt
 \lispnarrowing = #1em
 \fi
 \fi
 }
+% @firstparagraphindent WORD
+% If WORD is `none', then suppress indentation of the first paragraph
+% after a section heading.  If WORD is `insert', then do indent at such
+% paragraphs.
+%
+% The paragraph indentation is suppressed or not by calling
+% \suppressfirstparagraphindent, which the sectioning commands do.
+% We switch the definition of this back and forth according to WORD.
+% By default, we suppress indentation.
+%
+\def\suppressfirstparagraphindent{\dosuppressfirstparagraphindent}
+\def\insertword{insert}
+%
+\parseargdef\firstparagraphindent{%
+\def\temp{#1}%
+\ifx\temp\noneword
+\let\suppressfirstparagraphindent = \dosuppressfirstparagraphindent
+\else\ifx\temp\insertword
+\let\suppressfirstparagraphindent = \relax
+\else
+\errhelp = \EMsimple
+\errmessage{Unknown @firstparagraphindent option `\temp'}%
+\fi\fi
+}
+% Here is how we actually suppress indentation.  Redefine \everypar to
+% \kern backwards by \parindent, and then reset itself to empty.
+%
+% We also make \indent itself not actually do anything until the next
+% paragraph.
+%
+\gdef\dosuppressfirstparagraphindent{%
+\gdef\indent{%
+\restorefirstparagraphindent
+\indent
+}%
+\gdef\noindent{%
+\restorefirstparagraphindent
+\noindent
+}%
+\global\everypar = {%
+\kern -\parindent
+\restorefirstparagraphindent
+}%
+}
+\gdef\restorefirstparagraphindent{%
+\global \let \indent = \ptexindent
+\global \let \noindent = \ptexnoindent
+\global \everypar = {}%
+}
 % @asis just yields its argument.  Used with @table, for example.
 %
 \def\asis#1{#1}
-% @math means output in math mode.
+% @math outputs its argument in math mode.
-% We don't use $'s directly in the definition of \math because control
+%
-% sequences like \math are expanded when the toc file is written.  Then,
+% One complication: _ usually means subscripts, but it could also mean
-% we read the toc file back, the $'s will be normal characters (as they
+% an actual _ character, as in @math{@var{some_variable} + 1}.  So make
-% should be, according to the definition of Texinfo).  So we must use a
+% _ active, and distinguish by seeing if the current family is \slfam,
-% control sequence to switch into and out of math mode.
+% which is what @var uses.
-%
+{
-% This isn't quite enough for @math to work properly in indices, but it
+\catcode`\_ = \active
-% seems unlikely it will ever be needed there.
+\gdef\mathunderscore{%
-%
+\catcode`\_=\active
-\let\implicitmath = $
+\def_{\ifnum\fam=\slfam \_\else\sb\fi}%
-\def\math#1{\implicitmath #1\implicitmath}
+}
+}
-% @bullet and @minus need the same treatment as @math, just above.
+% Another complication: we want \\ (and @\) to output a \ character.
-\def\bullet{\implicitmath\ptexbullet\implicitmath}
+% FYI, plain.tex uses \\ as a temporary control sequence (why?), but
-\def\minus{\implicitmath-\implicitmath}
+% this is not advertised and we don't care.  Texinfo does not
+% otherwise define @\.
+%
+% The \mathchar is class=0=ordinary, family=7=ttfam, position=5C=\.
+\def\mathbackslash{\ifnum\fam=\ttfam \mathchar"075C \else\backslash \fi}
+%
+\def\math{%
+\tex
+\mathunderscore
+\let\\ = \mathbackslash
+\mathactive
+% make the texinfo accent commands work in math mode
+\let\"=\ddot
+\let\'=\acute
+\let\==\bar
+\let\^=\hat
+\let\`=\grave
+\let\u=\breve
+\let\v=\check
+\let\~=\tilde
+\let\dotaccent=\dot
+$\finishmath
+}
+\def\finishmath#1{#1$\endgroup}  % Close the group opened by \tex.
+% Some active characters (such as <) are spaced differently in math.
+% We have to reset their definitions in case the @math was an argument
+% to a command which sets the catcodes (such as @item or @section).
+%
+{
+\catcode`^ = \active
+\catcode`< = \active
+\catcode`> = \active
+\catcode`+ = \active
+\gdef\mathactive{%
+\let^ = \ptexhat
+\let< = \ptexless
+\let> = \ptexgtr
+\let+ = \ptexplus
+}
+}
+% Some math mode symbols.
+\def\bullet{$\ptexbullet$}
+\def\geq{\ifmmode \ge\else $\ge$\fi}
+\def\leq{\ifmmode \le\else $\le$\fi}
+\def\minus{\ifmmode -\else $-$\fi}
+% @dots{} outputs an ellipsis using the current font.
+% We do .5em per period so that it has the same spacing in the cm
+% typewriter fonts as three actual period characters; on the other hand,
+% in other typewriter fonts three periods are wider than 1.5em.  So do
+% whichever is larger.
+%
+\def\dots{%
+\leavevmode
+\setbox0=\hbox{...}% get width of three periods
+\ifdim\wd0 > 1.5em
+\dimen0 = \wd0
+\else
+\dimen0 = 1.5em
+\fi
+\hbox to \dimen0{%
+\hskip 0pt plus.25fil
+.\hskip 0pt plus1fil
+.\hskip 0pt plus1fil
+.\hskip 0pt plus.5fil
+}%
+}
+% @enddots{} is an end-of-sentence ellipsis.
+%
+\def\enddots{%
+\dots
+\spacefactor=\endofsentencespacefactor
+}
+% @comma{} is so commas can be inserted into text without messing up
+% Texinfo's parsing.
+%
+\let\comma = ,
 % @refill is a no-op.
 \let\refill=\relax
 % If working on a large document in chapters, it is convenient to
 % @setfilename is done at the beginning of every texinfo file.
 % So open here the files we need to have open while reading the input.
 % This makes it possible to make a .fmt file for texinfo.
 \def\setfilename{%
+\fixbackslash  % Turn off hack to swallow `\input texinfo'.
 \iflinks
-\readauxfile
+\tryauxfile
+% Open the new aux file.  TeX will close it automatically at exit.
+\immediate\openout\auxfile=\jobname.aux
 \fi % \openindices needs to do some work in any case.
 \openindices
-\fixbackslash  % Turn off hack to swallow `\input texinfo'.
+\let\setfilename=\comment % Ignore extra @setfilename cmds.
-\global\let\setfilename=\comment % Ignore extra @setfilename cmds.
 %
 % If texinfo.cnf is present on the system, read it.
 % Useful for site-wide @afourpaper, etc.
-% Just to be on the safe side, close the input stream before the \input.
 \openin 1 texinfo.cnf
-\ifeof1 \let\temp=\relax \else \def\temp{\input texinfo.cnf }\fi
+\ifeof 1 \else \input texinfo.cnf \fi
-\closein1
+\closein 1
-\temp
 %
 \comment % Ignore the actual filename.
 }
 % Called from \setfilename.
 \newbox\boxA
 \newcount\countA
 \newif\ifpdf
 \newif\ifpdfmakepagedest
+% when pdftex is run in dvi mode, \pdfoutput is defined (so \pdfoutput=1
+% can be set).  So we test for \relax and 0 as well as \undefined,
+% borrowed from ifpdf.sty.
 \ifx\pdfoutput\undefined
-\pdffalse
-\let\pdfmkdest = \gobble
-\let\pdfurl = \gobble
-\let\endlink = \relax
-\let\linkcolor = \relax
-\let\pdfmakeoutlines = \relax
 \else
-\pdftrue
+\ifx\pdfoutput\relax
-\pdfoutput = 1
+\else
-\input pdfcolor
+\ifcase\pdfoutput
+\else
+\pdftrue
+\fi
+\fi
+\fi
+% PDF uses PostScript string constants for the names of xref targets,
+% for display in the outlines, and in other places.  Thus, we have to
+% double any backslashes.  Otherwise, a name like "\node" will be
+% interpreted as a newline (\n), followed by o, d, e.  Not good.
+% http://www.ntg.nl/pipermail/ntg-pdftex/2004-July/000654.html
+% (and related messages, the final outcome is that it is up to the TeX
+% user to double the backslashes and otherwise make the string valid, so
+% that's what we do).
+% double active backslashes.
+%
+{\catcode`\@=0 \catcode`\\=\active
+@gdef@activebackslashdouble{%
+@catcode`@\=@active
+@let\=@doublebackslash}
+}
+% To handle parens, we must adopt a different approach, since parens are
+% not active characters.  hyperref.dtx (which has the same problem as
+% us) handles it with this amazing macro to replace tokens, with minor
+% changes for Texinfo.  It is included here under the GPL by permission
+% from the author, Heiko Oberdiek.
+%
+% #1 is the tokens to replace.
+% #2 is the replacement.
+% #3 is the control sequence with the string.
+%
+\def\HyPsdSubst#1#2#3{%
+\def\HyPsdReplace##1#1##2\END{%
+##1%
+\ifx\\##2\\%
+\else
+#2%
+\HyReturnAfterFi{%
+\HyPsdReplace##2\END
+}%
+\fi
+}%
+\xdef#3{\expandafter\HyPsdReplace#3#1\END}%
+}
+\long\def\HyReturnAfterFi#1\fi{\fi#1}
+% #1 is a control sequence in which to do the replacements.
+\def\backslashparens#1{%
+\xdef#1{#1}% redefine it as its expansion; the definition is simply
+% \lastnode when called from \setref -> \pdfmkdest.
+\HyPsdSubst{(}{\realbackslash(}{#1}%
+\HyPsdSubst{)}{\realbackslash)}{#1}%
+}
+\newhelp\nopdfimagehelp{Texinfo supports .png, .jpg, .jpeg, and .pdf images
+with PDF output, and none of those formats could be found.  (.eps cannot
+be supported due to the design of the PDF format; use regular TeX (DVI
+output) for that.)}
+\ifpdf
+%
+% Color manipulation macros based on pdfcolor.tex.
+\def\cmykDarkRed{0.28 1 1 0.35}
+\def\cmykBlack{0 0 0 1}
+%
+\def\pdfsetcolor#1{\pdfliteral{#1 k}}
+% Set color, and create a mark which defines \thiscolor accordingly,
+% so that \makeheadline knows which color to restore.
+\def\setcolor#1{%
+\xdef\lastcolordefs{\gdef\noexpand\thiscolor{#1}}%
+\domark
+\pdfsetcolor{#1}%
+}
+%
+\def\maincolor{\cmykBlack}
+\pdfsetcolor{\maincolor}
+\edef\thiscolor{\maincolor}
+\def\lastcolordefs{}
+%
+\def\makefootline{%
+\baselineskip24pt
+\line{\pdfsetcolor{\maincolor}\the\footline}%
+}
+%
+\def\makeheadline{%
+\vbox to 0pt{%
+\vskip-22.5pt
+\line{%
+\vbox to8.5pt{}%
+% Extract \thiscolor definition from the marks.
+\getcolormarks
+% Typeset the headline with \maincolor, then restore the color.
+\pdfsetcolor{\maincolor}\the\headline\pdfsetcolor{\thiscolor}%
+}%
+\vss
+}%
+\nointerlineskip
+}
+%
+%
+\pdfcatalog{/PageMode /UseOutlines}
+%
+% #1 is image name, #2 width (might be empty/whitespace), #3 height (ditto).
 \def\dopdfimage#1#2#3{%
-\def\imagewidth{#2}%
+\def\imagewidth{#2}\setbox0 = \hbox{\ignorespaces #2}%
-\def\imageheight{#3}%
+\def\imageheight{#3}\setbox2 = \hbox{\ignorespaces #3}%
+%
+% pdftex (and the PDF format) support .png, .jpg, .pdf (among
+% others).  Let's try in that order.
+\let\pdfimgext=\empty
+\begingroup
+\openin 1 #1.png \ifeof 1
+\openin 1 #1.jpg \ifeof 1
+\openin 1 #1.jpeg \ifeof 1
+\openin 1 #1.JPG \ifeof 1
+\openin 1 #1.pdf \ifeof 1
+\openin 1 #1.PDF \ifeof 1
+\errhelp = \nopdfimagehelp
+\errmessage{Could not find image file #1 for pdf}%
+\else \gdef\pdfimgext{PDF}%
+\fi
+\else \gdef\pdfimgext{pdf}%
+\fi
+\else \gdef\pdfimgext{JPG}%
+\fi
+\else \gdef\pdfimgext{jpeg}%
+\fi
+\else \gdef\pdfimgext{jpg}%
+\fi
+\else \gdef\pdfimgext{png}%
+\fi
+\closein 1
+\endgroup
+%
+% without \immediate, ancient pdftex seg faults when the same image is
+% included twice.  (Version 3.14159-pre-1.0-unofficial-20010704.)
 \ifnum\pdftexversion < 14
-\pdfimage
+\immediate\pdfimage
 \else
-\pdfximage
+\immediate\pdfximage
 \fi
-\ifx\empty\imagewidth\else width \imagewidth \fi
+\ifdim \wd0 >0pt width \imagewidth \fi
-\ifx\empty\imageheight\else height \imageheight \fi
+\ifdim \wd2 >0pt height \imageheight \fi
-{#1.pdf}%
+\ifnum\pdftexversion<13
+#1.\pdfimgext
+\else
+{#1.\pdfimgext}%
+\fi
 \ifnum\pdftexversion < 14 \else
 \pdfrefximage \pdflastximage
 \fi}
-\def\pdfmkdest#1{\pdfdest name{#1@} xyz}
+%
-\def\pdfmkpgn#1{#1@}
+\def\pdfmkdest#1{{%
-\let\linkcolor = \Cyan
+% We have to set dummies so commands such as @code, and characters
-\def\endlink{\Black\pdfendlink}
+% such as \, aren't expanded when present in a section title.
+\indexnofonts
+\turnoffactive
+\activebackslashdouble
+\makevalueexpandable
+\def\pdfdestname{#1}%
+\backslashparens\pdfdestname
+\safewhatsit{\pdfdest name{\pdfdestname} xyz}%
+}}
+%
+% used to mark target names; must be expandable.
+\def\pdfmkpgn#1{#1}
+%
+% by default, use a color that is dark enough to print on paper as
+% nearly black, but still distinguishable for online viewing.
+\def\urlcolor{\cmykDarkRed}
+\def\linkcolor{\cmykDarkRed}
+\def\endlink{\setcolor{\maincolor}\pdfendlink}
+%
 % Adding outlines to PDF; macros for calculating structure of outlines
 % come from Petr Olsak
 \def\expnumber#1{\expandafter\ifx\csname#1\endcsname\relax 0%
 \else \csname#1\endcsname \fi}
 \def\advancenumber#1{\tempnum=\expnumber{#1}\relax
-\advance\tempnum by1
+\advance\tempnum by 1
 \expandafter\xdef\csname#1\endcsname{\the\tempnum}}
-\def\pdfmakeoutlines{{%
+%
-\openin 1 \jobname.toc
+% #1 is the section text, which is what will be displayed in the
-\ifeof 1\else\bgroup
+% outline by the pdf viewer.  #2 is the pdf expression for the number
-\closein 1
+% of subentries (or empty, for subsubsections).  #3 is the node text,
-\indexnofonts
+% which might be empty if this toc entry had no corresponding node.
-\def\tt{}
+% #4 is the page number
-% thanh's hack / proper braces in bookmarks
+%
+\def\dopdfoutline#1#2#3#4{%
+% Generate a link to the node text if that exists; else, use the
+% page number.  We could generate a destination for the section
+% text in the case where a section has no node, but it doesn't
+% seem worth the trouble, since most documents are normally structured.
+\def\pdfoutlinedest{#3}%
+\ifx\pdfoutlinedest\empty
+\def\pdfoutlinedest{#4}%
+\else
+% Doubled backslashes in the name.
+{\activebackslashdouble \xdef\pdfoutlinedest{#3}%
+\backslashparens\pdfoutlinedest}%
+\fi
+%
+% Also double the backslashes in the display string.
+{\activebackslashdouble \xdef\pdfoutlinetext{#1}%
+\backslashparens\pdfoutlinetext}%
+%
+\pdfoutline goto name{\pdfmkpgn{\pdfoutlinedest}}#2{\pdfoutlinetext}%
+}
+%
+\def\pdfmakeoutlines{%
+\begingroup
+% Thanh's hack / proper braces in bookmarks
 \edef\mylbrace{\iftrue \string{\else}\fi}\let\{=\mylbrace
 \edef\myrbrace{\iffalse{\else\string}\fi}\let\}=\myrbrace
 %
-\def\chapentry ##1##2##3{}
+% Read toc silently, to get counts of subentries for \pdfoutline.
-\def\unnumbchapentry ##1##2{}
+\def\numchapentry##1##2##3##4{%
-\def\secentry ##1##2##3##4{\advancenumber{chap##2}}
+	\def\thischapnum{##2}%
-\def\unnumbsecentry ##1##2{}
+	\def\thissecnum{0}%
-\def\subsecentry ##1##2##3##4##5{\advancenumber{sec##2.##3}}
+	\def\thissubsecnum{0}%
-\def\unnumbsubsecentry ##1##2{}
+}%
-\def\subsubsecentry ##1##2##3##4##5##6{\advancenumber{subsec##2.##3.##4}}
+\def\numsecentry##1##2##3##4{%
-\def\unnumbsubsubsecentry ##1##2{}
+	\advancenumber{chap\thischapnum}%
-\input \jobname.toc
+	\def\thissecnum{##2}%
-\def\chapentry ##1##2##3{%
+	\def\thissubsecnum{0}%
-\pdfoutline goto name{\pdfmkpgn{##3}}count-\expnumber{chap##2}{##1}}
+}%
-\def\unnumbchapentry ##1##2{%
+\def\numsubsecentry##1##2##3##4{%
-\pdfoutline goto name{\pdfmkpgn{##2}}{##1}}
+	\advancenumber{sec\thissecnum}%
-\def\secentry ##1##2##3##4{%
+	\def\thissubsecnum{##2}%
-\pdfoutline goto name{\pdfmkpgn{##4}}count-\expnumber{sec##2.##3}{##1}}
+}%
-\def\unnumbsecentry ##1##2{%
+\def\numsubsubsecentry##1##2##3##4{%
-\pdfoutline goto name{\pdfmkpgn{##2}}{##1}}
+	\advancenumber{subsec\thissubsecnum}%
-\def\subsecentry ##1##2##3##4##5{%
+}%
-\pdfoutline goto name{\pdfmkpgn{##5}}count-\expnumber{subsec##2.##3.##4}{##1}}
+\def\thischapnum{0}%
-\def\unnumbsubsecentry ##1##2{%
+\def\thissecnum{0}%
-\pdfoutline goto name{\pdfmkpgn{##2}}{##1}}
+\def\thissubsecnum{0}%
-\def\subsubsecentry ##1##2##3##4##5##6{%
+%
-\pdfoutline goto name{\pdfmkpgn{##6}}{##1}}
+% use \def rather than \let here because we redefine \chapentry et
-\def\unnumbsubsubsecentry ##1##2{%
+% al. a second time, below.
-\pdfoutline goto name{\pdfmkpgn{##2}}{##1}}
+\def\appentry{\numchapentry}%
-\input \jobname.toc
+\def\appsecentry{\numsecentry}%
-\egroup\fi
+\def\appsubsecentry{\numsubsecentry}%
-}}
+\def\appsubsubsecentry{\numsubsubsecentry}%
-\def\makelinks #1,{%
+\def\unnchapentry{\numchapentry}%
-\def\params{#1}\def\E{END}%
+\def\unnsecentry{\numsecentry}%
-\ifx\params\E
+\def\unnsubsecentry{\numsubsecentry}%
-\let\nextmakelinks=\relax
+\def\unnsubsubsecentry{\numsubsubsecentry}%
-\else
+\readdatafile{toc}%
-\let\nextmakelinks=\makelinks
+%
-\ifnum\lnkcount>0,\fi
+% Read toc second time, this time actually producing the outlines.
-\picknum{#1}%
+% The `-' means take the \expnumber as the absolute number of
-\startlink attr{/Border [0 0 0]}
+% subentries, which we calculated on our first read of the .toc above.
-goto name{\pdfmkpgn{\the\pgn}}%
+%
-\linkcolor #1%
+% We use the node names as the destinations.
-\advance\lnkcount by 1%
+\def\numchapentry##1##2##3##4{%
-\endlink
+\dopdfoutline{##1}{count-\expnumber{chap##2}}{##3}{##4}}%
-\fi
+\def\numsecentry##1##2##3##4{%
-\nextmakelinks
+\dopdfoutline{##1}{count-\expnumber{sec##2}}{##3}{##4}}%
+\def\numsubsecentry##1##2##3##4{%
+\dopdfoutline{##1}{count-\expnumber{subsec##2}}{##3}{##4}}%
+\def\numsubsubsecentry##1##2##3##4{% count is always zero
+\dopdfoutline{##1}{}{##3}{##4}}%
+%
+% PDF outlines are displayed using system fonts, instead of
+% document fonts.  Therefore we cannot use special characters,
+% since the encoding is unknown.  For example, the eogonek from
+% Latin 2 (0xea) gets translated to a | character.  Info from
+% Staszek Wawrykiewicz, 19 Jan 2004 04:09:24 +0100.
+%
+% xx to do this right, we have to translate 8-bit characters to
+% their "best" equivalent, based on the @documentencoding.  Right
+% now, I guess we'll just let the pdf reader have its way.
+\indexnofonts
+\setupdatafile
+\catcode`\\=\active \otherbackslash
+\input \tocreadfilename
+\endgroup
 }
-\def\picknum#1{\expandafter\pn#1}
+%
-\def\pn#1{%
-\def\p{#1}%
-\ifx\p\lbrace
-\let\nextpn=\ppn
-\else
-\let\nextpn=\ppnn
-\def\first{#1}
-\fi
-\nextpn
-}
-\def\ppn#1{\pgn=#1\gobble}
-\def\ppnn{\pgn=\first}
-\def\pdfmklnk#1{\lnkcount=0\makelinks #1,END,}
-\def\addtokens#1#2{\edef\addtoks{\noexpand#1={\the#1#2}}\addtoks}
 \def\skipspaces#1{\def\PP{#1}\def\D{|}%
 \ifx\PP\D\let\nextsp\relax
 \else\let\nextsp\skipspaces
 \ifx\p\space\else\addtokens{\filename}{\PP}%
 \advance\filenamelength by 1
 \ifnum\pdftexversion < 14
 \let \startlink \pdfannotlink
 \else
 \let \startlink \pdfstartlink
 \fi
+% make a live url in pdf output.
 \def\pdfurl#1{%
 \begingroup
-\normalturnoffactive\def\@{@}%
+% it seems we really need yet another set of dummies; have not
-\leavevmode\Red
+% tried to figure out what each command should do in the context
+% of @url.  for now, just make @/ a no-op, that's the only one
+% people have actually reported a problem with.
+%
+\normalturnoffactive
+\def\@{@}%
+\let\/=\empty
+\makevalueexpandable
+\leavevmode\setcolor{\urlcolor}%
 \startlink attr{/Border [0 0 0]}%
 user{/Subtype /Link /A << /S /URI /URI (#1) >>}%
-% #1
 \endgroup}
 \def\pdfgettoks#1.{\setbox\boxA=\hbox{\toksA={#1.}\toksB={}\maketoks}}
 \def\addtokens#1#2{\edef\addtoks{\noexpand#1={\the#1#2}}\addtoks}
 \def\adn#1{\addtokens{\toksC}{#1}\global\countA=1\let\next=\maketoks}
 \def\poptoks#1#2|ENDTOKS|{\let\first=#1\toksD={#1}\toksA={#2}}
 \def\maketoks{%
-\expandafter\poptoks\the\toksA|ENDTOKS|
+\expandafter\poptoks\the\toksA|ENDTOKS|\relax
 \ifx\first0\adn0
 \else\ifx\first1\adn1 \else\ifx\first2\adn2 \else\ifx\first3\adn3
 \else\ifx\first4\adn4 \else\ifx\first5\adn5 \else\ifx\first6\adn6
 \else\ifx\first7\adn7 \else\ifx\first8\adn8 \else\ifx\first9\adn9
 \else
 \ifnum0=\countA\else\makelink\fi
 \ifx\first.\let\next=\done\else
 \let\next=\maketoks
 \addtokens{\toksB}{\the\toksD}
 \fi\fi\fi\fi\fi\fi\fi\fi\fi\fi
 \next}
 \def\makelink{\addtokens{\toksB}%
 {\noexpand\pdflink{\the\toksC}}\toksC={}\global\countA=0}
 \def\pdflink#1{%
-\startlink attr{/Border [0 0 0]} goto name{\mkpgn{#1}}
+\startlink attr{/Border [0 0 0]} goto name{\pdfmkpgn{#1}}
-\linkcolor #1\endlink}
+\setcolor{\linkcolor}#1\endlink}
-\def\mkpgn#1{#1@}
 \def\done{\edef\st{\global\noexpand\toksA={\the\toksB}}\st}
-\fi % \ifx\pdfoutput
+\else
+\let\pdfmkdest = \gobble
+\let\pdfurl = \gobble
+\let\endlink = \relax
+\let\setcolor = \gobble
+\let\pdfsetcolor = \gobble
+\let\pdfmakeoutlines = \relax
+\fi  % \ifx\pdfoutput
 \message{fonts,}
-% Font-change commands.
+% Change the current font style to #1, remembering it in \curfontstyle.
+% For now, we do not accumulate font styles: @b{@i{foo}} prints foo in
+% italics, not bold italics.
+%
+\def\setfontstyle#1{%
+\def\curfontstyle{#1}% not as a control sequence, because we are \edef'd.
+\csname ten#1\endcsname  % change the current font
+}
+% Select #1 fonts with the current style.
+%
+\def\selectfonts#1{\csname #1fonts\endcsname \csname\curfontstyle\endcsname}
+\def\rm{\fam=0 \setfontstyle{rm}}
+\def\it{\fam=\itfam \setfontstyle{it}}
+\def\sl{\fam=\slfam \setfontstyle{sl}}
+\def\bf{\fam=\bffam \setfontstyle{bf}}\def\bfstylename{bf}
+\def\tt{\fam=\ttfam \setfontstyle{tt}}
 % Texinfo sort of supports the sans serif font style, which plain TeX does not.
-% So we set up a \sf analogous to plain's \rm, etc.
+% So we set up a \sf.
 \newfam\sffam
-\def\sf{\fam=\sffam \tensf}
+\def\sf{\fam=\sffam \setfontstyle{sf}}
 \let\li = \sf % Sometimes we call it \li, not \sf.
-% We don't need math for this one.
+% We don't need math for this font style.
-\def\ttsl{\tenttsl}
+\def\ttsl{\setfontstyle{ttsl}}
-% Use Computer Modern fonts at \magstephalf (11pt).
-\newcount\mainmagstep
+% Default leading.
-\mainmagstep=\magstephalf
+\newdimen\textleading  \textleading = 13.2pt
+% Set the baselineskip to #1, and the lineskip and strut size
+% correspondingly.  There is no deep meaning behind these magic numbers
+% used as factors; they just match (closely enough) what Knuth defined.
+%
+\def\lineskipfactor{.08333}
+\def\strutheightpercent{.70833}
+\def\strutdepthpercent {.29167}
+%
+% can get a sort of poor man's double spacing by redefining this.
+\def\baselinefactor{1}
+%
+\def\setleading#1{%
+\dimen0 = #1\relax
+\normalbaselineskip = \baselinefactor\dimen0
+\normallineskip = \lineskipfactor\normalbaselineskip
+\normalbaselines
+\setbox\strutbox =\hbox{%
+\vrule width0pt height\strutheightpercent\baselineskip
+depth \strutdepthpercent \baselineskip
+}%
+}
+% PDF CMaps.  See also LaTeX's t1.cmap.
+%
+% do nothing with this by default.
+\expandafter\let\csname cmapOT1\endcsname\gobble
+\expandafter\let\csname cmapOT1IT\endcsname\gobble
+\expandafter\let\csname cmapOT1TT\endcsname\gobble
+% if we are producing pdf, and we have \pdffontattr, then define cmaps.
+% (\pdffontattr was introduced many years ago, but people still run
+% older pdftex's; it's easy to conditionalize, so we do.)
+\ifpdf \ifx\pdffontattr\undefined \else
+\begingroup
+\catcode`\^^M=\active \def^^M{^^J}% Output line endings as the ^^J char.
+\catcode`\%=12 \immediate\pdfobj stream {%!PS-Adobe-3.0 Resource-CMap
+%%DocumentNeededResources: ProcSet (CIDInit)
+%%IncludeResource: ProcSet (CIDInit)
+%%BeginResource: CMap (TeX-OT1-0)
+%%Title: (TeX-OT1-0 TeX OT1 0)
+%%Version: 1.000
+%%EndComments
+/CIDInit /ProcSet findresource begin
+12 dict begin
+begincmap
+/CIDSystemInfo
+<< /Registry (TeX)
+/Ordering (OT1)
+/Supplement 0
+>> def
+/CMapName /TeX-OT1-0 def
+/CMapType 2 def
+1 begincodespacerange
+<00> <7F>
+endcodespacerange
+8 beginbfrange
+<00> <01> <0393>
+<09> <0A> <03A8>
+<23> <26> <0023>
+<28> <3B> <0028>
+<3F> <5B> <003F>
+<5D> <5E> <005D>
+<61> <7A> <0061>
+<7B> <7C> <2013>
+endbfrange
+40 beginbfchar
+<02> <0398>
+<03> <039B>
+<04> <039E>
+<05> <03A0>
+<06> <03A3>
+<07> <03D2>
+<08> <03A6>
+<0B> <00660066>
+<0C> <00660069>
+<0D> <0066006C>
+<0E> <006600660069>
+<0F> <00660066006C>
+<10> <0131>
+<11> <0237>
+<12> <0060>
+<13> <00B4>
+<14> <02C7>
+<15> <02D8>
+<16> <00AF>
+<17> <02DA>
+<18> <00B8>
+<19> <00DF>
+<1A> <00E6>
+<1B> <0153>
+<1C> <00F8>
+<1D> <00C6>
+<1E> <0152>
+<1F> <00D8>
+<21> <0021>
+<22> <201D>
+<27> <2019>
+<3C> <00A1>
+<3D> <003D>
+<3E> <00BF>
+<5C> <201C>
+<5F> <02D9>
+<60> <2018>
+<7D> <02DD>
+<7E> <007E>
+<7F> <00A8>
+endbfchar
+endcmap
+CMapName currentdict /CMap defineresource pop
+end
+end
+%%EndResource
+%%EOF
+}\endgroup
+\expandafter\edef\csname cmapOT1\endcsname#1{%
+\pdffontattr#1{/ToUnicode \the\pdflastobj\space 0 R}%
+}%
+%
+% \cmapOT1IT
+\begingroup
+\catcode`\^^M=\active \def^^M{^^J}% Output line endings as the ^^J char.
+\catcode`\%=12 \immediate\pdfobj stream {%!PS-Adobe-3.0 Resource-CMap
+%%DocumentNeededResources: ProcSet (CIDInit)
+%%IncludeResource: ProcSet (CIDInit)
+%%BeginResource: CMap (TeX-OT1IT-0)
+%%Title: (TeX-OT1IT-0 TeX OT1IT 0)
+%%Version: 1.000
+%%EndComments
+/CIDInit /ProcSet findresource begin
+12 dict begin
+begincmap
+/CIDSystemInfo
+<< /Registry (TeX)
+/Ordering (OT1IT)
+/Supplement 0
+>> def
+/CMapName /TeX-OT1IT-0 def
+/CMapType 2 def
+1 begincodespacerange
+<00> <7F>
+endcodespacerange
+8 beginbfrange
+<00> <01> <0393>
+<09> <0A> <03A8>
+<25> <26> <0025>
+<28> <3B> <0028>
+<3F> <5B> <003F>
+<5D> <5E> <005D>
+<61> <7A> <0061>
+<7B> <7C> <2013>
+endbfrange
+42 beginbfchar
+<02> <0398>
+<03> <039B>
+<04> <039E>
+<05> <03A0>
+<06> <03A3>
+<07> <03D2>
+<08> <03A6>
+<0B> <00660066>
+<0C> <00660069>
+<0D> <0066006C>
+<0E> <006600660069>
+<0F> <00660066006C>
+<10> <0131>
+<11> <0237>
+<12> <0060>
+<13> <00B4>
+<14> <02C7>
+<15> <02D8>
+<16> <00AF>
+<17> <02DA>
+<18> <00B8>
+<19> <00DF>
+<1A> <00E6>
+<1B> <0153>
+<1C> <00F8>
+<1D> <00C6>
+<1E> <0152>
+<1F> <00D8>
+<21> <0021>
+<22> <201D>
+<23> <0023>
+<24> <00A3>
+<27> <2019>
+<3C> <00A1>
+<3D> <003D>
+<3E> <00BF>
+<5C> <201C>
+<5F> <02D9>
+<60> <2018>
+<7D> <02DD>
+<7E> <007E>
+<7F> <00A8>
+endbfchar
+endcmap
+CMapName currentdict /CMap defineresource pop
+end
+end
+%%EndResource
+%%EOF
+}\endgroup
+\expandafter\edef\csname cmapOT1IT\endcsname#1{%
+\pdffontattr#1{/ToUnicode \the\pdflastobj\space 0 R}%
+}%
+%
+% \cmapOT1TT
+\begingroup
+\catcode`\^^M=\active \def^^M{^^J}% Output line endings as the ^^J char.
+\catcode`\%=12 \immediate\pdfobj stream {%!PS-Adobe-3.0 Resource-CMap
+%%DocumentNeededResources: ProcSet (CIDInit)
+%%IncludeResource: ProcSet (CIDInit)
+%%BeginResource: CMap (TeX-OT1TT-0)
+%%Title: (TeX-OT1TT-0 TeX OT1TT 0)
+%%Version: 1.000
+%%EndComments
+/CIDInit /ProcSet findresource begin
+12 dict begin
+begincmap
+/CIDSystemInfo
+<< /Registry (TeX)
+/Ordering (OT1TT)
+/Supplement 0
+>> def
+/CMapName /TeX-OT1TT-0 def
+/CMapType 2 def
+1 begincodespacerange
+<00> <7F>
+endcodespacerange
+5 beginbfrange
+<00> <01> <0393>
+<09> <0A> <03A8>
+<21> <26> <0021>
+<28> <5F> <0028>
+<61> <7E> <0061>
+endbfrange
+32 beginbfchar
+<02> <0398>
+<03> <039B>
+<04> <039E>
+<05> <03A0>
+<06> <03A3>
+<07> <03D2>
+<08> <03A6>
+<0B> <2191>
+<0C> <2193>
+<0D> <0027>
+<0E> <00A1>
+<0F> <00BF>
+<10> <0131>
+<11> <0237>
+<12> <0060>
+<13> <00B4>
+<14> <02C7>
+<15> <02D8>
+<16> <00AF>
+<17> <02DA>
+<18> <00B8>
+<19> <00DF>
+<1A> <00E6>
+<1B> <0153>
+<1C> <00F8>
+<1D> <00C6>
+<1E> <0152>
+<1F> <00D8>
+<20> <2423>
+<27> <2019>
+<60> <2018>
+<7F> <00A8>
+endbfchar
+endcmap
+CMapName currentdict /CMap defineresource pop
+end
+end
+%%EndResource
+%%EOF
+}\endgroup
+\expandafter\edef\csname cmapOT1TT\endcsname#1{%
+\pdffontattr#1{/ToUnicode \the\pdflastobj\space 0 R}%
+}%
+\fi\fi
 % Set the font macro #1 to the font named #2, adding on the
 % specified font prefix (normally `cm').
-% #3 is the font's design size, #4 is a scale factor
+% #3 is the font's design size, #4 is a scale factor, #5 is the CMap
-\def\setfont#1#2#3#4{\font#1=\fontprefix#2#3 scaled #4}
+% encoding (currently only OT1, OT1IT and OT1TT are allowed, pass
+% empty to omit).
+\def\setfont#1#2#3#4#5{%
+\font#1=\fontprefix#2#3 scaled #4
+\csname cmap#5\endcsname#1%
+}
+% This is what gets called when #5 of \setfont is empty.
+\let\cmap\gobble
+% emacs-page end of cmaps
 % Use cm as the default font prefix.
 % To specify the font prefix, you must define \fontprefix
 % before you read in texinfo.tex.
 \ifx\fontprefix\undefined
 \def\sfshape{ss}
 \def\sfbshape{ss}
 \def\scshape{csc}
 \def\scbshape{csc}
-\ifx\bigger\relax
+% Definitions for a main text size of 11pt.  This is the default in
-\let\mainmagstep=\magstep1
+% Texinfo.
-\setfont\textrm\rmshape{12}{1000}
+%
-\setfont\texttt\ttshape{12}{1000}
+\def\definetextfontsizexi{%
-\else
+% Text fonts (11.2pt, magstep1).
-\setfont\textrm\rmshape{10}{\mainmagstep}
+\def\textnominalsize{11pt}
-\setfont\texttt\ttshape{10}{\mainmagstep}
+\edef\mainmagstep{\magstephalf}
-\fi
+\setfont\textrm\rmshape{10}{\mainmagstep}{OT1}
-% Instead of cmb10, you many want to use cmbx10.
+\setfont\texttt\ttshape{10}{\mainmagstep}{OT1TT}
-% cmbx10 is a prettier font on its own, but cmb10
+\setfont\textbf\bfshape{10}{\mainmagstep}{OT1}
-% looks better when embedded in a line with cmr10.
+\setfont\textit\itshape{10}{\mainmagstep}{OT1IT}
-\setfont\textbf\bfshape{10}{\mainmagstep}
+\setfont\textsl\slshape{10}{\mainmagstep}{OT1}
-\setfont\textit\itshape{10}{\mainmagstep}
+\setfont\textsf\sfshape{10}{\mainmagstep}{OT1}
-\setfont\textsl\slshape{10}{\mainmagstep}
+\setfont\textsc\scshape{10}{\mainmagstep}{OT1}
-\setfont\textsf\sfshape{10}{\mainmagstep}
+\setfont\textttsl\ttslshape{10}{\mainmagstep}{OT1TT}
-\setfont\textsc\scshape{10}{\mainmagstep}
-\setfont\textttsl\ttslshape{10}{\mainmagstep}
 \font\texti=cmmi10 scaled \mainmagstep
 \font\textsy=cmsy10 scaled \mainmagstep
+\def\textecsize{1095}
-% A few fonts for @defun, etc.
-\setfont\defbf\bxshape{10}{\magstep1} %was 1314
+% A few fonts for @defun names and args.
-\setfont\deftt\ttshape{10}{\magstep1}
+\setfont\defbf\bfshape{10}{\magstep1}{OT1}
-\def\df{\let\tentt=\deftt \let\tenbf = \defbf \bf}
+\setfont\deftt\ttshape{10}{\magstep1}{OT1TT}
+\setfont\defttsl\ttslshape{10}{\magstep1}{OT1TT}
+\def\df{\let\tentt=\deftt \let\tenbf = \defbf \let\tenttsl=\defttsl \bf}
 % Fonts for indices, footnotes, small examples (9pt).
-\setfont\smallrm\rmshape{9}{1000}
+\def\smallnominalsize{9pt}
-\setfont\smalltt\ttshape{9}{1000}
+\setfont\smallrm\rmshape{9}{1000}{OT1}
-\setfont\smallbf\bfshape{10}{900}
+\setfont\smalltt\ttshape{9}{1000}{OT1TT}
-\setfont\smallit\itshape{9}{1000}
+\setfont\smallbf\bfshape{10}{900}{OT1}
-\setfont\smallsl\slshape{9}{1000}
+\setfont\smallit\itshape{9}{1000}{OT1IT}
-\setfont\smallsf\sfshape{9}{1000}
+\setfont\smallsl\slshape{9}{1000}{OT1}
-\setfont\smallsc\scshape{10}{900}
+\setfont\smallsf\sfshape{9}{1000}{OT1}
-\setfont\smallttsl\ttslshape{10}{900}
+\setfont\smallsc\scshape{10}{900}{OT1}
+\setfont\smallttsl\ttslshape{10}{900}{OT1TT}
 \font\smalli=cmmi9
 \font\smallsy=cmsy9
+\def\smallecsize{0900}
-% Fonts for title page:
-\setfont\titlerm\rmbshape{12}{\magstep3}
+% Fonts for small examples (8pt).
-\setfont\titleit\itbshape{10}{\magstep4}
+\def\smallernominalsize{8pt}
-\setfont\titlesl\slbshape{10}{\magstep4}
+\setfont\smallerrm\rmshape{8}{1000}{OT1}
-\setfont\titlett\ttbshape{12}{\magstep3}
+\setfont\smallertt\ttshape{8}{1000}{OT1TT}
-\setfont\titlettsl\ttslshape{10}{\magstep4}
+\setfont\smallerbf\bfshape{10}{800}{OT1}
-\setfont\titlesf\sfbshape{17}{\magstep1}
+\setfont\smallerit\itshape{8}{1000}{OT1IT}
+\setfont\smallersl\slshape{8}{1000}{OT1}
+\setfont\smallersf\sfshape{8}{1000}{OT1}
+\setfont\smallersc\scshape{10}{800}{OT1}
+\setfont\smallerttsl\ttslshape{10}{800}{OT1TT}
+\font\smalleri=cmmi8
+\font\smallersy=cmsy8
+\def\smallerecsize{0800}
+% Fonts for title page (20.4pt):
+\def\titlenominalsize{20pt}
+\setfont\titlerm\rmbshape{12}{\magstep3}{OT1}
+\setfont\titleit\itbshape{10}{\magstep4}{OT1IT}
+\setfont\titlesl\slbshape{10}{\magstep4}{OT1}
+\setfont\titlett\ttbshape{12}{\magstep3}{OT1TT}
+\setfont\titlettsl\ttslshape{10}{\magstep4}{OT1TT}
+\setfont\titlesf\sfbshape{17}{\magstep1}{OT1}
 \let\titlebf=\titlerm
-\setfont\titlesc\scbshape{10}{\magstep4}
+\setfont\titlesc\scbshape{10}{\magstep4}{OT1}
 \font\titlei=cmmi12 scaled \magstep3
 \font\titlesy=cmsy10 scaled \magstep4
 \def\authorrm{\secrm}
+\def\authortt{\sectt}
+\def\titleecsize{2074}
 % Chapter (and unnumbered) fonts (17.28pt).
-\setfont\chaprm\rmbshape{12}{\magstep2}
+\def\chapnominalsize{17pt}
-\setfont\chapit\itbshape{10}{\magstep3}
+\setfont\chaprm\rmbshape{12}{\magstep2}{OT1}
-\setfont\chapsl\slbshape{10}{\magstep3}
+\setfont\chapit\itbshape{10}{\magstep3}{OT1IT}
-\setfont\chaptt\ttbshape{12}{\magstep2}
+\setfont\chapsl\slbshape{10}{\magstep3}{OT1}
-\setfont\chapttsl\ttslshape{10}{\magstep3}
+\setfont\chaptt\ttbshape{12}{\magstep2}{OT1TT}
-\setfont\chapsf\sfbshape{17}{1000}
+\setfont\chapttsl\ttslshape{10}{\magstep3}{OT1TT}
+\setfont\chapsf\sfbshape{17}{1000}{OT1}
 \let\chapbf=\chaprm
-\setfont\chapsc\scbshape{10}{\magstep3}
+\setfont\chapsc\scbshape{10}{\magstep3}{OT1}
 \font\chapi=cmmi12 scaled \magstep2
 \font\chapsy=cmsy10 scaled \magstep3
+\def\chapecsize{1728}
 % Section fonts (14.4pt).
-\setfont\secrm\rmbshape{12}{\magstep1}
+\def\secnominalsize{14pt}
-\setfont\secit\itbshape{10}{\magstep2}
+\setfont\secrm\rmbshape{12}{\magstep1}{OT1}
-\setfont\secsl\slbshape{10}{\magstep2}
+\setfont\secit\itbshape{10}{\magstep2}{OT1IT}
-\setfont\sectt\ttbshape{12}{\magstep1}
+\setfont\secsl\slbshape{10}{\magstep2}{OT1}
-\setfont\secttsl\ttslshape{10}{\magstep2}
+\setfont\sectt\ttbshape{12}{\magstep1}{OT1TT}
-\setfont\secsf\sfbshape{12}{\magstep1}
+\setfont\secttsl\ttslshape{10}{\magstep2}{OT1TT}
+\setfont\secsf\sfbshape{12}{\magstep1}{OT1}
 \let\secbf\secrm
-\setfont\secsc\scbshape{10}{\magstep2}
+\setfont\secsc\scbshape{10}{\magstep2}{OT1}
 \font\seci=cmmi12 scaled \magstep1
 \font\secsy=cmsy10 scaled \magstep2
+\def\sececsize{1440}
-% \setfont\ssecrm\bxshape{10}{\magstep1}    % This size an font looked bad.
-% \setfont\ssecit\itshape{10}{\magstep1}    % The letters were too crowded.
-% \setfont\ssecsl\slshape{10}{\magstep1}
-% \setfont\ssectt\ttshape{10}{\magstep1}
-% \setfont\ssecsf\sfshape{10}{\magstep1}
-%\setfont\ssecrm\bfshape{10}{1315}      % Note the use of cmb rather than cmbx.
-%\setfont\ssecit\itshape{10}{1315}      % Also, the size is a little larger than
-%\setfont\ssecsl\slshape{10}{1315}      % being scaled magstep1.
-%\setfont\ssectt\ttshape{10}{1315}
-%\setfont\ssecsf\sfshape{10}{1315}
-%\let\ssecbf=\ssecrm
 % Subsection fonts (13.15pt).
-\setfont\ssecrm\rmbshape{12}{\magstephalf}
+\def\ssecnominalsize{13pt}
-\setfont\ssecit\itbshape{10}{1315}
+\setfont\ssecrm\rmbshape{12}{\magstephalf}{OT1}
-\setfont\ssecsl\slbshape{10}{1315}
+\setfont\ssecit\itbshape{10}{1315}{OT1IT}
-\setfont\ssectt\ttbshape{12}{\magstephalf}
+\setfont\ssecsl\slbshape{10}{1315}{OT1}
-\setfont\ssecttsl\ttslshape{10}{1315}
+\setfont\ssectt\ttbshape{12}{\magstephalf}{OT1TT}
-\setfont\ssecsf\sfbshape{12}{\magstephalf}
+\setfont\ssecttsl\ttslshape{10}{1315}{OT1TT}
+\setfont\ssecsf\sfbshape{12}{\magstephalf}{OT1}
 \let\ssecbf\ssecrm
-\setfont\ssecsc\scbshape{10}{\magstep1}
+\setfont\ssecsc\scbshape{10}{1315}{OT1}
 \font\sseci=cmmi12 scaled \magstephalf
 \font\ssecsy=cmsy10 scaled 1315
-% The smallcaps and symbol fonts should actually be scaled \magstep1.5,
+\def\ssececsize{1200}
-% but that is not a standard magnification.
+% Reduced fonts for @acro in text (10pt).
+\def\reducednominalsize{10pt}
+\setfont\reducedrm\rmshape{10}{1000}{OT1}
+\setfont\reducedtt\ttshape{10}{1000}{OT1TT}
+\setfont\reducedbf\bfshape{10}{1000}{OT1}
+\setfont\reducedit\itshape{10}{1000}{OT1IT}
+\setfont\reducedsl\slshape{10}{1000}{OT1}
+\setfont\reducedsf\sfshape{10}{1000}{OT1}
+\setfont\reducedsc\scshape{10}{1000}{OT1}
+\setfont\reducedttsl\ttslshape{10}{1000}{OT1TT}
+\font\reducedi=cmmi10
+\font\reducedsy=cmsy10
+\def\reducedecsize{1000}
+% reset the current fonts
+\textfonts
+\rm
+} % end of 11pt text font size definitions
+% Definitions to make the main text be 10pt Computer Modern, with
+% section, chapter, etc., sizes following suit.  This is for the GNU
+% Press printing of the Emacs 22 manual.  Maybe other manuals in the
+% future.  Used with @smallbook, which sets the leading to 12pt.
+%
+\def\definetextfontsizex{%
+% Text fonts (10pt).
+\def\textnominalsize{10pt}
+\edef\mainmagstep{1000}
+\setfont\textrm\rmshape{10}{\mainmagstep}{OT1}
+\setfont\texttt\ttshape{10}{\mainmagstep}{OT1TT}
+\setfont\textbf\bfshape{10}{\mainmagstep}{OT1}
+\setfont\textit\itshape{10}{\mainmagstep}{OT1IT}
+\setfont\textsl\slshape{10}{\mainmagstep}{OT1}
+\setfont\textsf\sfshape{10}{\mainmagstep}{OT1}
+\setfont\textsc\scshape{10}{\mainmagstep}{OT1}
+\setfont\textttsl\ttslshape{10}{\mainmagstep}{OT1TT}
+\font\texti=cmmi10 scaled \mainmagstep
+\font\textsy=cmsy10 scaled \mainmagstep
+\def\textecsize{1000}
+% A few fonts for @defun names and args.
+\setfont\defbf\bfshape{10}{\magstephalf}{OT1}
+\setfont\deftt\ttshape{10}{\magstephalf}{OT1TT}
+\setfont\defttsl\ttslshape{10}{\magstephalf}{OT1TT}
+\def\df{\let\tentt=\deftt \let\tenbf = \defbf \let\tenttsl=\defttsl \bf}
+% Fonts for indices, footnotes, small examples (9pt).
+\def\smallnominalsize{9pt}
+\setfont\smallrm\rmshape{9}{1000}{OT1}
+\setfont\smalltt\ttshape{9}{1000}{OT1TT}
+\setfont\smallbf\bfshape{10}{900}{OT1}
+\setfont\smallit\itshape{9}{1000}{OT1IT}
+\setfont\smallsl\slshape{9}{1000}{OT1}
+\setfont\smallsf\sfshape{9}{1000}{OT1}
+\setfont\smallsc\scshape{10}{900}{OT1}
+\setfont\smallttsl\ttslshape{10}{900}{OT1TT}
+\font\smalli=cmmi9
+\font\smallsy=cmsy9
+\def\smallecsize{0900}
+% Fonts for small examples (8pt).
+\def\smallernominalsize{8pt}
+\setfont\smallerrm\rmshape{8}{1000}{OT1}
+\setfont\smallertt\ttshape{8}{1000}{OT1TT}
+\setfont\smallerbf\bfshape{10}{800}{OT1}
+\setfont\smallerit\itshape{8}{1000}{OT1IT}
+\setfont\smallersl\slshape{8}{1000}{OT1}
+\setfont\smallersf\sfshape{8}{1000}{OT1}
+\setfont\smallersc\scshape{10}{800}{OT1}
+\setfont\smallerttsl\ttslshape{10}{800}{OT1TT}
+\font\smalleri=cmmi8
+\font\smallersy=cmsy8
+\def\smallerecsize{0800}
+% Fonts for title page (20.4pt):
+\def\titlenominalsize{20pt}
+\setfont\titlerm\rmbshape{12}{\magstep3}{OT1}
+\setfont\titleit\itbshape{10}{\magstep4}{OT1IT}
+\setfont\titlesl\slbshape{10}{\magstep4}{OT1}
+\setfont\titlett\ttbshape{12}{\magstep3}{OT1TT}
+\setfont\titlettsl\ttslshape{10}{\magstep4}{OT1TT}
+\setfont\titlesf\sfbshape{17}{\magstep1}{OT1}
+\let\titlebf=\titlerm
+\setfont\titlesc\scbshape{10}{\magstep4}{OT1}
+\font\titlei=cmmi12 scaled \magstep3
+\font\titlesy=cmsy10 scaled \magstep4
+\def\authorrm{\secrm}
+\def\authortt{\sectt}
+\def\titleecsize{2074}
+% Chapter fonts (14.4pt).
+\def\chapnominalsize{14pt}
+\setfont\chaprm\rmbshape{12}{\magstep1}{OT1}
+\setfont\chapit\itbshape{10}{\magstep2}{OT1IT}
+\setfont\chapsl\slbshape{10}{\magstep2}{OT1}
+\setfont\chaptt\ttbshape{12}{\magstep1}{OT1TT}
+\setfont\chapttsl\ttslshape{10}{\magstep2}{OT1TT}
+\setfont\chapsf\sfbshape{12}{\magstep1}{OT1}
+\let\chapbf\chaprm
+\setfont\chapsc\scbshape{10}{\magstep2}{OT1}
+\font\chapi=cmmi12 scaled \magstep1
+\font\chapsy=cmsy10 scaled \magstep2
+\def\chapecsize{1440}
+% Section fonts (12pt).
+\def\secnominalsize{12pt}
+\setfont\secrm\rmbshape{12}{1000}{OT1}
+\setfont\secit\itbshape{10}{\magstep1}{OT1IT}
+\setfont\secsl\slbshape{10}{\magstep1}{OT1}
+\setfont\sectt\ttbshape{12}{1000}{OT1TT}
+\setfont\secttsl\ttslshape{10}{\magstep1}{OT1TT}
+\setfont\secsf\sfbshape{12}{1000}{OT1}
+\let\secbf\secrm
+\setfont\secsc\scbshape{10}{\magstep1}{OT1}
+\font\seci=cmmi12
+\font\secsy=cmsy10 scaled \magstep1
+\def\sececsize{1200}
+% Subsection fonts (10pt).
+\def\ssecnominalsize{10pt}
+\setfont\ssecrm\rmbshape{10}{1000}{OT1}
+\setfont\ssecit\itbshape{10}{1000}{OT1IT}
+\setfont\ssecsl\slbshape{10}{1000}{OT1}
+\setfont\ssectt\ttbshape{10}{1000}{OT1TT}
+\setfont\ssecttsl\ttslshape{10}{1000}{OT1TT}
+\setfont\ssecsf\sfbshape{10}{1000}{OT1}
+\let\ssecbf\ssecrm
+\setfont\ssecsc\scbshape{10}{1000}{OT1}
+\font\sseci=cmmi10
+\font\ssecsy=cmsy10
+\def\ssececsize{1000}
+% Reduced fonts for @acro in text (9pt).
+\def\reducednominalsize{9pt}
+\setfont\reducedrm\rmshape{9}{1000}{OT1}
+\setfont\reducedtt\ttshape{9}{1000}{OT1TT}
+\setfont\reducedbf\bfshape{10}{900}{OT1}
+\setfont\reducedit\itshape{9}{1000}{OT1IT}
+\setfont\reducedsl\slshape{9}{1000}{OT1}
+\setfont\reducedsf\sfshape{9}{1000}{OT1}
+\setfont\reducedsc\scshape{10}{900}{OT1}
+\setfont\reducedttsl\ttslshape{10}{900}{OT1TT}
+\font\reducedi=cmmi9
+\font\reducedsy=cmsy9
+\def\reducedecsize{0900}
+% reduce space between paragraphs
+\divide\parskip by 2
+% reset the current fonts
+\textfonts
+\rm
+} % end of 10pt text font size definitions
+% We provide the user-level command
+%   @fonttextsize 10
+% (or 11) to redefine the text font size.  pt is assumed.
+%
+\def\xword{10}
+\def\xiword{11}
+%
+\parseargdef\fonttextsize{%
+\def\textsizearg{#1}%
+\wlog{doing @fonttextsize \textsizearg}%
+%
+% Set \globaldefs so that documents can use this inside @tex, since
+% makeinfo 4.8 does not support it, but we need it nonetheless.
+%
+\begingroup \globaldefs=1
+\ifx\textsizearg\xword \definetextfontsizex
+\else \ifx\textsizearg\xiword \definetextfontsizexi
+\else
+\errhelp=\EMsimple
+\errmessage{@fonttextsize only supports `10' or `11', not `\textsizearg'}
+\fi\fi
+\endgroup
+}
 % In order for the font changes to affect most math symbols and letters,
 % we have to define the \textfont of the standard families.  Since
-% texinfo doesn't allow for producing subscripts and superscripts, we
+% texinfo doesn't allow for producing subscripts and superscripts except
-% don't bother to reset \scriptfont and \scriptscriptfont (which would
+% in the main text, we don't bother to reset \scriptfont and
-% also require loading a lot more fonts).
+% \scriptscriptfont (which would also require loading a lot more fonts).
 %
 \def\resetmathfonts{%
-\textfont0 = \tenrm \textfont1 = \teni \textfont2 = \tensy
+\textfont0=\tenrm \textfont1=\teni \textfont2=\tensy
-\textfont\itfam = \tenit \textfont\slfam = \tensl \textfont\bffam = \tenbf
+\textfont\itfam=\tenit \textfont\slfam=\tensl \textfont\bffam=\tenbf
-\textfont\ttfam = \tentt \textfont\sffam = \tensf
+\textfont\ttfam=\tentt \textfont\sffam=\tensf
 }
 % The font-changing commands redefine the meanings of \tenSTYLE, instead
-% of just \STYLE.  We do this so that font changes will continue to work
+% of just \STYLE.  We do this because \STYLE needs to also set the
-% in math mode, where it is the current \fam that is relevant in most
+% current \fam for math mode.  Our \STYLE (e.g., \rm) commands hardwire
-% cases, not the current font.  Plain TeX does \def\bf{\fam=\bffam
+% \tenSTYLE to set the current font.
-% \tenbf}, for example.  By redefining \tenbf, we obviate the need to
+%
-% redefine \bf itself.
+% Each font-changing command also sets the names \lsize (one size lower)
+% and \lllsize (three sizes lower).  These relative commands are used in
+% the LaTeX logo and acronyms.
+%
+% This all needs generalizing, badly.
+%
 \def\textfonts{%
 \let\tenrm=\textrm \let\tenit=\textit \let\tensl=\textsl
 \let\tenbf=\textbf \let\tentt=\texttt \let\smallcaps=\textsc
-\let\tensf=\textsf \let\teni=\texti \let\tensy=\textsy \let\tenttsl=\textttsl
+\let\tensf=\textsf \let\teni=\texti \let\tensy=\textsy
-\resetmathfonts}
+\let\tenttsl=\textttsl
+\def\curfontsize{text}%
+\def\lsize{reduced}\def\lllsize{smaller}%
+\resetmathfonts \setleading{\textleading}}
 \def\titlefonts{%
 \let\tenrm=\titlerm \let\tenit=\titleit \let\tensl=\titlesl
 \let\tenbf=\titlebf \let\tentt=\titlett \let\smallcaps=\titlesc
 \let\tensf=\titlesf \let\teni=\titlei \let\tensy=\titlesy
 \let\tenttsl=\titlettsl
+\def\curfontsize{title}%
+\def\lsize{chap}\def\lllsize{subsec}%
 \resetmathfonts \setleading{25pt}}
 \def\titlefont#1{{\titlefonts\rm #1}}
 \def\chapfonts{%
 \let\tenrm=\chaprm \let\tenit=\chapit \let\tensl=\chapsl
 \let\tenbf=\chapbf \let\tentt=\chaptt \let\smallcaps=\chapsc
-\let\tensf=\chapsf \let\teni=\chapi \let\tensy=\chapsy \let\tenttsl=\chapttsl
+\let\tensf=\chapsf \let\teni=\chapi \let\tensy=\chapsy
+\let\tenttsl=\chapttsl
+\def\curfontsize{chap}%
+\def\lsize{sec}\def\lllsize{text}%
 \resetmathfonts \setleading{19pt}}
 \def\secfonts{%
 \let\tenrm=\secrm \let\tenit=\secit \let\tensl=\secsl
 \let\tenbf=\secbf \let\tentt=\sectt \let\smallcaps=\secsc
-\let\tensf=\secsf \let\teni=\seci \let\tensy=\secsy \let\tenttsl=\secttsl
+\let\tensf=\secsf \let\teni=\seci \let\tensy=\secsy
+\let\tenttsl=\secttsl
+\def\curfontsize{sec}%
+\def\lsize{subsec}\def\lllsize{reduced}%
 \resetmathfonts \setleading{16pt}}
 \def\subsecfonts{%
 \let\tenrm=\ssecrm \let\tenit=\ssecit \let\tensl=\ssecsl
 \let\tenbf=\ssecbf \let\tentt=\ssectt \let\smallcaps=\ssecsc
-\let\tensf=\ssecsf \let\teni=\sseci \let\tensy=\ssecsy \let\tenttsl=\ssecttsl
+\let\tensf=\ssecsf \let\teni=\sseci \let\tensy=\ssecsy
+\let\tenttsl=\ssecttsl
+\def\curfontsize{ssec}%
+\def\lsize{text}\def\lllsize{small}%
 \resetmathfonts \setleading{15pt}}
-\let\subsubsecfonts = \subsecfonts % Maybe make sssec fonts scaled magstephalf?
+\let\subsubsecfonts = \subsecfonts
+\def\reducedfonts{%
+\let\tenrm=\reducedrm \let\tenit=\reducedit \let\tensl=\reducedsl
+\let\tenbf=\reducedbf \let\tentt=\reducedtt \let\reducedcaps=\reducedsc
+\let\tensf=\reducedsf \let\teni=\reducedi \let\tensy=\reducedsy
+\let\tenttsl=\reducedttsl
+\def\curfontsize{reduced}%
+\def\lsize{small}\def\lllsize{smaller}%
+\resetmathfonts \setleading{10.5pt}}
 \def\smallfonts{%
 \let\tenrm=\smallrm \let\tenit=\smallit \let\tensl=\smallsl
 \let\tenbf=\smallbf \let\tentt=\smalltt \let\smallcaps=\smallsc
 \let\tensf=\smallsf \let\teni=\smalli \let\tensy=\smallsy
 \let\tenttsl=\smallttsl
-\resetmathfonts \setleading{11pt}}
+\def\curfontsize{small}%
+\def\lsize{smaller}\def\lllsize{smaller}%
+\resetmathfonts \setleading{10.5pt}}
+\def\smallerfonts{%
+\let\tenrm=\smallerrm \let\tenit=\smallerit \let\tensl=\smallersl
+\let\tenbf=\smallerbf \let\tentt=\smallertt \let\smallcaps=\smallersc
+\let\tensf=\smallersf \let\teni=\smalleri \let\tensy=\smallersy
+\let\tenttsl=\smallerttsl
+\def\curfontsize{smaller}%
+\def\lsize{smaller}\def\lllsize{smaller}%
+\resetmathfonts \setleading{9.5pt}}
+% Set the fonts to use with the @small... environments.
+\let\smallexamplefonts = \smallfonts
+% About \smallexamplefonts.  If we use \smallfonts (9pt), @smallexample
+% can fit this many characters:
+%   8.5x11=86   smallbook=72  a4=90  a5=69
+% If we use \scriptfonts (8pt), then we can fit this many characters:
+%   8.5x11=90+  smallbook=80  a4=90+  a5=77
+% For me, subjectively, the few extra characters that fit aren't worth
+% the additional smallness of 8pt.  So I'm making the default 9pt.
+%
+% By the way, for comparison, here's what fits with @example (10pt):
+%   8.5x11=71  smallbook=60  a4=75  a5=58
+%
+% I wish the USA used A4 paper.
+% --karl, 24jan03.
 % Set up the default fonts, so we can use them for creating boxes.
 %
-\textfonts
+\definetextfontsizexi
 % Define these so they can be easily changed for other fonts.
 \def\angleleft{$\langle$}
 \def\angleright{$\rangle$}
 % Count depth in font-changes, for error checks
 \newcount\fontdepth \fontdepth=0
 % Fonts for short table of contents.
-\setfont\shortcontrm\rmshape{12}{1000}
+\setfont\shortcontrm\rmshape{12}{1000}{OT1}
-\setfont\shortcontbf\bxshape{12}{1000}
+\setfont\shortcontbf\bfshape{10}{\magstep1}{OT1}  % no cmb12
-\setfont\shortcontsl\slshape{12}{1000}
+\setfont\shortcontsl\slshape{12}{1000}{OT1}
+\setfont\shortconttt\ttshape{12}{1000}{OT1TT}
 %% Add scribe-like font environments, plus @l for inline lisp (usually sans
 %% serif) and @ii for TeX italic
 % \smartitalic{ARG} outputs arg in italics, followed by an italic correction
 % unless the following character is such as not to need one.
-\def\smartitalicx{\ifx\next,\else\ifx\next-\else\ifx\next.\else\/\fi\fi\fi}
+\def\smartitalicx{\ifx\next,\else\ifx\next-\else\ifx\next.\else
-\def\smartslanted#1{{\sl #1}\futurelet\next\smartitalicx}
+\ptexslash\fi\fi\fi}
-\def\smartitalic#1{{\it #1}\futurelet\next\smartitalicx}
+\def\smartslanted#1{{\ifusingtt\ttsl\sl #1}\futurelet\next\smartitalicx}
+\def\smartitalic#1{{\ifusingtt\ttsl\it #1}\futurelet\next\smartitalicx}
+% like \smartslanted except unconditionally uses \ttsl.
+% @var is set to this for defun arguments.
+\def\ttslanted#1{{\ttsl #1}\futurelet\next\smartitalicx}
+% like \smartslanted except unconditionally use \sl.  We never want
+% ttsl for book titles, do we?
+\def\cite#1{{\sl #1}\futurelet\next\smartitalicx}
 \let\i=\smartitalic
+\let\slanted=\smartslanted
 \let\var=\smartslanted
 \let\dfn=\smartslanted
 \let\emph=\smartitalic
-\let\cite=\smartslanted
+% @b, explicit bold.
 \def\b#1{{\bf #1}}
 \let\strong=\b
+% @sansserif, explicit sans.
+\def\sansserif#1{{\sf #1}}
 % We can't just use \exhyphenpenalty, because that only has effect at
 % the end of a paragraph.  Restore normal hyphenation at the end of the
 % group within which \nohyphenation is presumably called.
 %
 \def\nohyphenation{\hyphenchar\font = -1  \aftergroup\restorehyphenation}
 \def\restorehyphenation{\hyphenchar\font = `- }
+% Set sfcode to normal for the chars that usually have another value.
+% Can't use plain's \frenchspacing because it uses the `\x notation, and
+% sometimes \x has an active definition that messes things up.
+%
+\catcode`@=11
+\def\plainfrenchspacing{%
+\sfcode\dotChar  =\@m \sfcode\questChar=\@m \sfcode\exclamChar=\@m
+\sfcode\colonChar=\@m \sfcode\semiChar =\@m \sfcode\commaChar =\@m
+\def\endofsentencespacefactor{1000}% for @. and friends
+}
+\def\plainnonfrenchspacing{%
+\sfcode`\.3000\sfcode`\?3000\sfcode`\!3000
+\sfcode`\:2000\sfcode`\;1500\sfcode`\,1250
+\def\endofsentencespacefactor{3000}% for @. and friends
+}
+\catcode`@=\other
+\def\endofsentencespacefactor{3000}% default
 \def\t#1{%
-{\tt \rawbackslash \frenchspacing #1}%
+{\tt \rawbackslash \plainfrenchspacing #1}%
 \null
 }
-\let\ttfont=\t
 \def\samp#1{`\tclose{#1}'\null}
-\setfont\keyrm\rmshape{8}{1000}
+\setfont\keyrm\rmshape{8}{1000}{OT1}
 \font\keysy=cmsy9
 \def\key#1{{\keyrm\textfont2=\keysy \leavevmode\hbox{%
 \raise0.4pt\hbox{\angleleft}\kern-.08em\vtop{%
 \vbox{\hrule\kern-0.4pt
 \hbox{\raise0.4pt\hbox{\vphantom{\angleleft}}#1}}%
 \kern-0.4pt\hrule}%
 \kern-.06em\raise0.4pt\hbox{\angleright}}}}
+\def\key #1{{\nohyphenation \uppercase{#1}}\null}
 % The old definition, with no lozenge:
 %\def\key #1{{\ttsl \nohyphenation \uppercase{#1}}\null}
 \def\ctrl #1{{\tt \rawbackslash \hat}#1}
 % @file, @option are the same as @samp.
 %
 % Turn off hyphenation.
 \nohyphenation
 %
 \rawbackslash
-\frenchspacing
+\plainfrenchspacing
 #1%
 }%
 \null
 }
-% We *must* turn on hyphenation at `-' and `_' in \code.
+% We *must* turn on hyphenation at `-' and `_' in @code.
 % Otherwise, it is too hard to avoid overfull hboxes
 % in the Emacs manual, the Library manual, etc.
 % Unfortunately, TeX uses one parameter (\hyphenchar) to control
 % both hyphenation at - and hyphenation within words.
 % We must therefore turn them both off (\tclose does that)
 % and arrange explicitly to hyphenate at a dash.
 %  -- rms.
 {
-\catcode`\-=\active
+\catcode`\-=\active \catcode`\_=\active
-\catcode`\_=\active
+\catcode`\'=\active \catcode`\`=\active
 %
 \global\def\code{\begingroup
-\catcode`\-=\active \let-\codedash
+\catcode\rquoteChar=\active \catcode\lquoteChar=\active
-\catcode`\_=\active \let_\codeunder
+\let'\codequoteright \let`\codequoteleft
+%
+\catcode\dashChar=\active  \catcode\underChar=\active
+\ifallowcodebreaks
+\let-\codedash
+\let_\codeunder
+\else
+\let-\realdash
+\let_\realunder
+\fi
 \codex
 }
-%
-% If we end up with any active - characters when handling the index,
-% just treat them as a normal -.
-\global\def\indexbreaks{\catcode`\-=\active \let-\realdash}
 }
 \def\realdash{-}
 \def\codedash{-\discretionary{}{}{}}
-\def\codeunder{\ifusingtt{\normalunderscore\discretionary{}{}{}}{\_}}
+\def\codeunder{%
+% this is all so @math{@code{var_name}+1} can work.  In math mode, _
+% is "active" (mathcode"8000) and \normalunderscore (or \char95, etc.)
+% will therefore expand the active definition of _, which is us
+% (inside @code that is), therefore an endless loop.
+\ifusingtt{\ifmmode
+\mathchar"075F % class 0=ordinary, family 7=ttfam, pos 0x5F=_.
+\else\normalunderscore \fi
+\discretionary{}{}{}}%
+{\_}%
+}
 \def\codex #1{\tclose{#1}\endgroup}
-%\let\exp=\tclose  %Was temporary
+% An additional complication: the above will allow breaks after, e.g.,
+% each of the four underscores in __typeof__.  This is undesirable in
+% some manuals, especially if they don't have long identifiers in
+% general.  @allowcodebreaks provides a way to control this.
+%
+\newif\ifallowcodebreaks  \allowcodebreakstrue
+\def\keywordtrue{true}
+\def\keywordfalse{false}
+\parseargdef\allowcodebreaks{%
+\def\txiarg{#1}%
+\ifx\txiarg\keywordtrue
+\allowcodebreakstrue
+\else\ifx\txiarg\keywordfalse
+\allowcodebreaksfalse
+\else
+\errhelp = \EMsimple
+\errmessage{Unknown @allowcodebreaks option `\txiarg'}%
+\fi\fi
+}
 % @kbd is like @code, except that if the argument is just one @key command,
 % then @kbd has no effect.
 % @kbdinputstyle -- arg is `distinct' (@kbd uses slanted tty font always),
 %   `example' (@kbd uses ttsl only inside of @example and friends),
 %   or `code' (@kbd uses normal tty font always).
-\def\kbdinputstyle{\parsearg\kbdinputstylexxx}
+\parseargdef\kbdinputstyle{%
-\def\kbdinputstylexxx#1{%
+\def\txiarg{#1}%
-\def\arg{#1}%
+\ifx\txiarg\worddistinct
-\ifx\arg\worddistinct
 \gdef\kbdexamplefont{\ttsl}\gdef\kbdfont{\ttsl}%
-\else\ifx\arg\wordexample
+\else\ifx\txiarg\wordexample
 \gdef\kbdexamplefont{\ttsl}\gdef\kbdfont{\tt}%
-\else\ifx\arg\wordcode
+\else\ifx\txiarg\wordcode
 \gdef\kbdexamplefont{\tt}\gdef\kbdfont{\tt}%
+\else
+\errhelp = \EMsimple
+\errmessage{Unknown @kbdinputstyle option `\txiarg'}%
 \fi\fi\fi
 }
 \def\worddistinct{distinct}
 \def\wordexample{example}
 \def\wordcode{code}
-% Default is kbdinputdistinct.  (Too much of a hassle to call the macro,
+% Default is `distinct.'
-% the catcodes are wrong for parsearg to work.)
+\kbdinputstyle distinct
-\gdef\kbdexamplefont{\ttsl}\gdef\kbdfont{\ttsl}
 \def\xkey{\key}
 \def\kbdfoo#1#2#3\par{\def\one{#1}\def\three{#3}\def\threex{??}%
 \ifx\one\xkey\ifx\threex\three \key{#2}%
 \else{\tclose{\kbdfont\look}}\fi
 \else{\tclose{\kbdfont\look}}\fi}
-% For @url, @env, @command quotes seem unnecessary, so use \code.
+% For @indicateurl, @env, @command quotes seem unnecessary, so use \code.
-\let\url=\code
+\let\indicateurl=\code
 \let\env=\code
 \let\command=\code
+% @clicksequence{File @click{} Open ...}
+\def\clicksequence#1{\begingroup #1\endgroup}
+% @clickstyle @arrow   (by default)
+\parseargdef\clickstyle{\def\click{#1}}
+\def\click{\arrow}
 % @uref (abbreviation for `urlref') takes an optional (comma-separated)
 % second argument specifying the text to display and an optional third
 % arg as text to display instead of (rather than in addition to) the url
 % itself.  First (mandatory) arg is the url.  Perhaps eventually put in
 \fi
 \fi
 \endlink
 \endgroup}
+% @url synonym for @uref, since that's how everyone uses it.
+%
+\let\url=\uref
 % rms does not like angle brackets --karl, 17may97.
 % So now @email is just like @uref, unless we are pdf.
 %
 %\def\email#1{\angleleft{\tt #1}\angleright}
 \ifpdf
 \def\email#1{\doemail#1,,\finish}
 \def\doemail#1,#2,#3\finish{\begingroup
 \unsepspaces
 % Explicit font changes: @r, @sc, undocumented @ii.
 \def\r#1{{\rm #1}}              % roman font
 \def\sc#1{{\smallcaps#1}}       % smallcaps font
 \def\ii#1{{\it #1}}             % italic font
-% @acronym downcases the argument and prints in smallcaps.
+% @acronym for "FBI", "NATO", and the like.
-\def\acronym#1{{\smallcaps \lowercase{#1}}}
+% We print this one point size smaller, since it's intended for
+% all-uppercase.
-% @pounds{} is a sterling sign.
+%
+\def\acronym#1{\doacronym #1,,\finish}
+\def\doacronym#1,#2,#3\finish{%
+{\selectfonts\lsize #1}%
+\def\temp{#2}%
+\ifx\temp\empty \else
+\space ({\unsepspaces \ignorespaces \temp \unskip})%
+\fi
+}
+% @abbr for "Comput. J." and the like.
+% No font change, but don't do end-of-sentence spacing.
+%
+\def\abbr#1{\doabbr #1,,\finish}
+\def\doabbr#1,#2,#3\finish{%
+{\plainfrenchspacing #1}%
+\def\temp{#2}%
+\ifx\temp\empty \else
+\space ({\unsepspaces \ignorespaces \temp \unskip})%
+\fi
+}
+% @pounds{} is a sterling sign, which Knuth put in the CM italic font.
+%
 \def\pounds{{\it\$}}
+% @euro{} comes from a separate font, depending on the current style.
+% We use the free feym* fonts from the eurosym package by Henrik
+% Theiling, which support regular, slanted, bold and bold slanted (and
+% "outlined" (blackboard board, sort of) versions, which we don't need).
+% It is available from http://www.ctan.org/tex-archive/fonts/eurosym.
+%
+% Although only regular is the truly official Euro symbol, we ignore
+% that.  The Euro is designed to be slightly taller than the regular
+% font height.
+%
+% feymr - regular
+% feymo - slanted
+% feybr - bold
+% feybo - bold slanted
+%
+% There is no good (free) typewriter version, to my knowledge.
+% A feymr10 euro is ~7.3pt wide, while a normal cmtt10 char is ~5.25pt wide.
+% Hmm.
+%
+% Also doesn't work in math.  Do we need to do math with euro symbols?
+% Hope not.
+%
+%
+\def\euro{{\eurofont e}}
+\def\eurofont{%
+% We set the font at each command, rather than predefining it in
+% \textfonts and the other font-switching commands, so that
+% installations which never need the symbol don't have to have the
+% font installed.
+%
+% There is only one designed size (nominal 10pt), so we always scale
+% that to the current nominal size.
+%
+% By the way, simply using "at 1em" works for cmr10 and the like, but
+% does not work for cmbx10 and other extended/shrunken fonts.
+%
+\def\eurosize{\csname\curfontsize nominalsize\endcsname}%
+%
+\ifx\curfontstyle\bfstylename
+% bold:
+\font\thiseurofont = \ifusingit{feybo10}{feybr10} at \eurosize
+\else
+% regular:
+\font\thiseurofont = \ifusingit{feymo10}{feymr10} at \eurosize
+\fi
+\thiseurofont
+}
+% Hacks for glyphs from the EC fonts similar to \euro.  We don't
+% use \let for the aliases, because sometimes we redefine the original
+% macro, and the alias should reflect the redefinition.
+\def\guillemetleft{{\ecfont \char"13}}
+\def\guillemotleft{\guillemetleft}
+\def\guillemetright{{\ecfont \char"14}}
+\def\guillemotright{\guillemetright}
+\def\guilsinglleft{{\ecfont \char"0E}}
+\def\guilsinglright{{\ecfont \char"0F}}
+\def\quotedblbase{{\ecfont \char"12}}
+\def\quotesinglbase{{\ecfont \char"0D}}
+%
+\def\ecfont{%
+% We can't distinguish serif/sanserif and italic/slanted, but this
+% is used for crude hacks anyway (like adding French and German
+% quotes to documents typeset with CM, where we lose kerning), so
+% hopefully nobody will notice/care.
+\edef\ecsize{\csname\curfontsize ecsize\endcsname}%
+\edef\nominalsize{\csname\curfontsize nominalsize\endcsname}%
+\ifx\curfontstyle\bfstylename
+% bold:
+\font\thisecfont = ecb\ifusingit{i}{x}\ecsize \space at \nominalsize
+\else
+% regular:
+\font\thisecfont = ec\ifusingit{ti}{rm}\ecsize \space at \nominalsize
+\fi
+\thisecfont
+}
+% @registeredsymbol - R in a circle.  The font for the R should really
+% be smaller yet, but lllsize is the best we can do for now.
+% Adapted from the plain.tex definition of \copyright.
+%
+\def\registeredsymbol{%
+$^{{\ooalign{\hfil\raise.07ex\hbox{\selectfonts\lllsize R}%
+\hfil\crcr\Orb}}%
+}$%
+}
+% @textdegree - the normal degrees sign.
+%
+\def\textdegree{$^\circ$}
+% Laurent Siebenmann reports \Orb undefined with:
+%  Textures 1.7.7 (preloaded format=plain 93.10.14)  (68K)  16 APR 2004 02:38
+% so we'll define it if necessary.
+%
+\ifx\Orb\undefined
+\def\Orb{\mathhexbox20D}
+\fi
+% Quotes.
+\chardef\quotedblleft="5C
+\chardef\quotedblright=`\"
+\chardef\quoteleft=`\`
+\chardef\quoteright=`\'
 \message{page headings,}
 \newskip\titlepagetopglue \titlepagetopglue = 1.5in
 \newif\ifsetcontentsaftertitlepage
 \let\setcontentsaftertitlepage = \setcontentsaftertitlepagetrue
 \newif\ifsetshortcontentsaftertitlepage
 \let\setshortcontentsaftertitlepage = \setshortcontentsaftertitlepagetrue
-\def\shorttitlepage{\parsearg\shorttitlepagezzz}
+\parseargdef\shorttitlepage{\begingroup\hbox{}\vskip 1.5in \chaprm \centerline{#1}%
-\def\shorttitlepagezzz #1{\begingroup\hbox{}\vskip 1.5in \chaprm \centerline{#1}%
 \endgroup\page\hbox{}\page}
-\def\titlepage{\begingroup \parindent=0pt \textfonts
+\envdef\titlepage{%
-\let\subtitlerm=\tenrm
+% Open one extra group, as we want to close it in the middle of \Etitlepage.
-\def\subtitlefont{\subtitlerm \normalbaselineskip = 13pt \normalbaselines}%
+\begingroup
-%
+\parindent=0pt \textfonts
-\def\authorfont{\authorrm \normalbaselineskip = 16pt \normalbaselines}%
+% Leave some space at the very top of the page.
-%
+\vglue\titlepagetopglue
-% Leave some space at the very top of the page.
+% No rule at page bottom unless we print one at the top with @title.
-\vglue\titlepagetopglue
+\finishedtitlepagetrue
 %
-% Now you can print the title using @title.
+% Most title ``pages'' are actually two pages long, with space
-\def\title{\parsearg\titlezzz}%
+% at the top of the second.  We don't want the ragged left on the second.
-\def\titlezzz##1{\leftline{\titlefonts\rm ##1}
+\let\oldpage = \page
-% print a rule at the page bottom also.
+\def\page{%
-\finishedtitlepagefalse
-\vskip4pt \hrule height 4pt width \hsize \vskip4pt}%
-% No rule at page bottom unless we print one at the top with @title.
-\finishedtitlepagetrue
-%
-% Now you can put text using @subtitle.
-\def\subtitle{\parsearg\subtitlezzz}%
-\def\subtitlezzz##1{{\subtitlefont \rightline{##1}}}%
-%
-% @author should come last, but may come many times.
-\def\author{\parsearg\authorzzz}%
-\def\authorzzz##1{\ifseenauthor\else\vskip 0pt plus 1filll\seenauthortrue\fi
-{\authorfont \leftline{##1}}}%
-%
-% Most title ``pages'' are actually two pages long, with space
-% at the top of the second.  We don't want the ragged left on the second.
-\let\oldpage = \page
-\def\page{%
 \iffinishedtitlepage\else
-\finishtitlepage
+	 \finishtitlepage
 \fi
-\oldpage
 \let\page = \oldpage
-\hbox{}}%
+\page
-%   \def\page{\oldpage \hbox{}}
+\null
+}%
 }
 \def\Etitlepage{%
 \iffinishedtitlepage\else
-\finishtitlepage
+	\finishtitlepage
 \fi
 % It is important to do the page break before ending the group,
 % because the headline and footline are only empty inside the group.
 % If we use the new definition of \page, we always get a blank page
 % after the title page, which we certainly don't want.
 \oldpage
 \endgroup
 %
-% If they want short, they certainly want long too.
+% Need this before the \...aftertitlepage checks so that if they are
-\ifsetshortcontentsaftertitlepage
+% in effect the toc pages will come out with page numbers.
-\shortcontents
+\HEADINGSon
-\contents
+%
-\global\let\shortcontents = \relax
+% If they want short, they certainly want long too.
-\global\let\contents = \relax
+\ifsetshortcontentsaftertitlepage
-\fi
+\shortcontents
-%
+\contents
-\ifsetcontentsaftertitlepage
+\global\let\shortcontents = \relax
-\contents
+\global\let\contents = \relax
-\global\let\contents = \relax
+\fi
-\global\let\shortcontents = \relax
+%
-\fi
+\ifsetcontentsaftertitlepage
-%
+\contents
-\ifpdf \pdfmakepagedesttrue \fi
+\global\let\contents = \relax
-%
+\global\let\shortcontents = \relax
-\HEADINGSon
+\fi
 }
 \def\finishtitlepage{%
 \vskip4pt \hrule height 2pt width \hsize
 \vskip\titlepagebottomglue
 \finishedtitlepagetrue
 }
+%%% Macros to be used within @titlepage:
+\let\subtitlerm=\tenrm
+\def\subtitlefont{\subtitlerm \normalbaselineskip = 13pt \normalbaselines}
+\def\authorfont{\authorrm \normalbaselineskip = 16pt \normalbaselines
+		\let\tt=\authortt}
+\parseargdef\title{%
+\checkenv\titlepage
+\leftline{\titlefonts\rm #1}
+% print a rule at the page bottom also.
+\finishedtitlepagefalse
+\vskip4pt \hrule height 4pt width \hsize \vskip4pt
+}
+\parseargdef\subtitle{%
+\checkenv\titlepage
+{\subtitlefont \rightline{#1}}%
+}
+% @author should come last, but may come many times.
+% It can also be used inside @quotation.
+%
+\parseargdef\author{%
+\def\temp{\quotation}%
+\ifx\thisenv\temp
+\def\quotationauthor{#1}% printed in \Equotation.
+\else
+\checkenv\titlepage
+\ifseenauthor\else \vskip 0pt plus 1filll \seenauthortrue \fi
+{\authorfont \leftline{#1}}%
+\fi
+}
 %%% Set up page headings and footings.
 \let\thispage=\folio
 \newtoks\evenheadline    % headline on even pages
 \newtoks\oddheadline     % headline on odd pages
 \newtoks\evenfootline    % footline on even pages
 \newtoks\oddfootline     % footline on odd pages
-% Now make Tex use those variables
+% Now make TeX use those variables
 \headline={{\textfonts\rm \ifodd\pageno \the\oddheadline
 \else \the\evenheadline \fi}}
 \footline={{\textfonts\rm \ifodd\pageno \the\oddfootline
 \else \the\evenfootline \fi}\HEADINGShook}
 \let\HEADINGShook=\relax
 % @evenheading @thistitle|@thispage|@thischapter
 % @oddheading @thischapter|@thispage|@thistitle
 % @evenfooting @thisfile||
 % @oddfooting ||@thisfile
 \def\evenheading{\parsearg\evenheadingxxx}
+\def\evenheadingxxx #1{\evenheadingyyy #1\|\|\|\|\finish}
+\def\evenheadingyyy #1\|#2\|#3\|#4\finish{%
+\global\evenheadline={\rlap{\centerline{#2}}\line{#1\hfil#3}}}
 \def\oddheading{\parsearg\oddheadingxxx}
-\def\everyheading{\parsearg\everyheadingxxx}
+\def\oddheadingxxx #1{\oddheadingyyy #1\|\|\|\|\finish}
+\def\oddheadingyyy #1\|#2\|#3\|#4\finish{%
+\global\oddheadline={\rlap{\centerline{#2}}\line{#1\hfil#3}}}
+\parseargdef\everyheading{\oddheadingxxx{#1}\evenheadingxxx{#1}}%
 \def\evenfooting{\parsearg\evenfootingxxx}
+\def\evenfootingxxx #1{\evenfootingyyy #1\|\|\|\|\finish}
+\def\evenfootingyyy #1\|#2\|#3\|#4\finish{%
+\global\evenfootline={\rlap{\centerline{#2}}\line{#1\hfil#3}}}
 \def\oddfooting{\parsearg\oddfootingxxx}
-\def\everyfooting{\parsearg\everyfootingxxx}
+\def\oddfootingxxx #1{\oddfootingyyy #1\|\|\|\|\finish}
+\def\oddfootingyyy #1\|#2\|#3\|#4\finish{%
-{\catcode`\@=0 %
-\gdef\evenheadingxxx #1{\evenheadingyyy #1@|@|@|@|\finish}
-\gdef\evenheadingyyy #1@|#2@|#3@|#4\finish{%
-\global\evenheadline={\rlap{\centerline{#2}}\line{#1\hfil#3}}}
-\gdef\oddheadingxxx #1{\oddheadingyyy #1@|@|@|@|\finish}
-\gdef\oddheadingyyy #1@|#2@|#3@|#4\finish{%
-\global\oddheadline={\rlap{\centerline{#2}}\line{#1\hfil#3}}}
-\gdef\everyheadingxxx#1{\oddheadingxxx{#1}\evenheadingxxx{#1}}%
-\gdef\evenfootingxxx #1{\evenfootingyyy #1@|@|@|@|\finish}
-\gdef\evenfootingyyy #1@|#2@|#3@|#4\finish{%
-\global\evenfootline={\rlap{\centerline{#2}}\line{#1\hfil#3}}}
-\gdef\oddfootingxxx #1{\oddfootingyyy #1@|@|@|@|\finish}
-\gdef\oddfootingyyy #1@|#2@|#3@|#4\finish{%
 \global\oddfootline = {\rlap{\centerline{#2}}\line{#1\hfil#3}}%
 %
 % Leave some space for the footline.  Hopefully ok to assume
 % @evenfooting will not be used by itself.
-\global\advance\pageheight by -\baselineskip
+\global\advance\pageheight by -12pt
-\global\advance\vsize by -\baselineskip
+\global\advance\vsize by -12pt
 }
-\gdef\everyfootingxxx#1{\oddfootingxxx{#1}\evenfootingxxx{#1}}
+\parseargdef\everyfooting{\oddfootingxxx{#1}\evenfootingxxx{#1}}
-%
-}% unbind the catcode of @.
+% @evenheadingmarks top     \thischapter <- chapter at the top of a page
+% @evenheadingmarks bottom  \thischapter <- chapter at the bottom of a page
+%
+% The same set of arguments for:
+%
+% @oddheadingmarks
+% @evenfootingmarks
+% @oddfootingmarks
+% @everyheadingmarks
+% @everyfootingmarks
+\def\evenheadingmarks{\headingmarks{even}{heading}}
+\def\oddheadingmarks{\headingmarks{odd}{heading}}
+\def\evenfootingmarks{\headingmarks{even}{footing}}
+\def\oddfootingmarks{\headingmarks{odd}{footing}}
+\def\everyheadingmarks#1 {\headingmarks{even}{heading}{#1}
+\headingmarks{odd}{heading}{#1} }
+\def\everyfootingmarks#1 {\headingmarks{even}{footing}{#1}
+\headingmarks{odd}{footing}{#1} }
+% #1 = even/odd, #2 = heading/footing, #3 = top/bottom.
+\def\headingmarks#1#2#3 {%
+\expandafter\let\expandafter\temp \csname get#3headingmarks\endcsname
+\global\expandafter\let\csname get#1#2marks\endcsname \temp
+}
+\everyheadingmarks bottom
+\everyfootingmarks bottom
 % @headings double      turns headings on for double-sided printing.
 % @headings single      turns headings on for single-sided printing.
 % @headings off         turns them off.
 % @headings on          same as @headings double, retained for compatibility.
 % By default, they are off at the start of a document,
 % and turned `on' after @end titlepage.
 \def\headings #1 {\csname HEADINGS#1\endcsname}
-\def\HEADINGSoff{
+\def\HEADINGSoff{%
 \global\evenheadline={\hfil} \global\evenfootline={\hfil}
 \global\oddheadline={\hfil} \global\oddfootline={\hfil}}
 \HEADINGSoff
 % When we turn headings on, set the page number to 1.
 % For double-sided printing, put current file name in lower left corner,
 % chapter name on inside top of right hand pages, document
 % title on inside top of left hand pages, and page numbers on outside top
 % edge of all pages.
-\def\HEADINGSdouble{
+\def\HEADINGSdouble{%
 \global\pageno=1
 \global\evenfootline={\hfil}
 \global\oddfootline={\hfil}
 \global\evenheadline={\line{\folio\hfil\thistitle}}
 \global\oddheadline={\line{\thischapter\hfil\folio}}
 }
 \let\contentsalignmacro = \chappager
 % For single-sided printing, chapter title goes across top left of page,
 % page number on top right.
-\def\HEADINGSsingle{
+\def\HEADINGSsingle{%
 \global\pageno=1
 \global\evenfootline={\hfil}
 \global\oddfootline={\hfil}
 \global\evenheadline={\line{\thischapter\hfil\folio}}
 \global\oddheadline={\line{\thischapter\hfil\folio}}
 \global\oddheadline={\line{\thischapter\hfil\folio}}
 \global\let\contentsalignmacro = \chappager
 }
 % Subroutines used in generating headings
-% Produces Day Month Year style of output.
+% This produces Day Month Year style of output.
+% Only define if not already defined, in case a txi-??.tex file has set
+% up a different format (e.g., txi-cs.tex does this).
+\ifx\today\undefined
 \def\today{%
 \number\day\space
 \ifcase\month
 \or\putwordMJan\or\putwordMFeb\or\putwordMMar\or\putwordMApr
 \or\putwordMMay\or\putwordMJun\or\putwordMJul\or\putwordMAug
 \or\putwordMSep\or\putwordMOct\or\putwordMNov\or\putwordMDec
 \fi
 \space\number\year}
+\fi
 % @settitle line...  specifies the title of the document, for headings.
 % It generates no output of its own.
 \def\thistitle{\putwordNoTitle}
-\def\settitle{\parsearg\settitlezzz}
+\def\settitle{\parsearg{\gdef\thistitle}}
-\def\settitlezzz #1{\gdef\thistitle{#1}}
 \message{tables,}
-% Tables -- @table, @ftable, @vtable, @item(x), @kitem(x), @xitem(x).
+% Tables -- @table, @ftable, @vtable, @item(x).
 % default indentation of table text
 \newdimen\tableindent \tableindent=.8in
 % default indentation of @itemize and @enumerate text
 \newdimen\itemindent  \itemindent=.3in
 \newdimen\itemmargin  \itemmargin=.1in
 % used internally for \itemindent minus \itemmargin
 \newdimen\itemmax
-% Note @table, @vtable, and @vtable define @item, @itemx, etc., with
+% Note @table, @ftable, and @vtable define @item, @itemx, etc., with
 % these defs.
 % They also define \itemindex
 % to index the item name in whatever manner is desired (perhaps none).
 \newif\ifitemxneedsnegativevskip
 \def\itemxpar{\par\ifitemxneedsnegativevskip\nobreak\vskip-\parskip\nobreak\fi}
 \def\internalBitem{\smallbreak \parsearg\itemzzz}
 \def\internalBitemx{\itemxpar \parsearg\itemzzz}
-\def\internalBxitem "#1"{\def\xitemsubtopix{#1} \smallbreak \parsearg\xitemzzz}
-\def\internalBxitemx "#1"{\def\xitemsubtopix{#1} \itemxpar \parsearg\xitemzzz}
-\def\internalBkitem{\smallbreak \parsearg\kitemzzz}
-\def\internalBkitemx{\itemxpar \parsearg\kitemzzz}
-\def\kitemzzz #1{\dosubind {kw}{\code{#1}}{for {\bf \lastfunction}}%
-\itemzzz {#1}}
-\def\xitemzzz #1{\dosubind {kw}{\code{#1}}{for {\bf \xitemsubtopic}}%
-\itemzzz {#1}}
 \def\itemzzz #1{\begingroup %
 \advance\hsize by -\rightskip
 \advance\hsize by -\tableindent
-\setbox0=\hbox{\itemfont{#1}}%
+\setbox0=\hbox{\itemindicate{#1}}%
 \itemindex{#1}%
 \nobreak % This prevents a break before @itemx.
 %
 % If the item text does not fit in the space we have, put it on a line
 % by itself, and do not allow a page break either before or after that
 %
 % We're going to be starting a paragraph, but we don't want the
 % \parskip glue -- logically it's part of the @item we just started.
 \nobreak \vskip-\parskip
 %
-% Stop a page break at the \parskip glue coming up.  Unfortunately
+% Stop a page break at the \parskip glue coming up.  However, if
-% we can't prevent a possible page break at the following
+% what follows is an environment such as @example, there will be no
-% \baselineskip glue.
+% \parskip glue; then the negative vskip we just inserted would
-\nobreak
+% cause the example and the item to crash together.  So we use this
+% bizarre value of 10001 as a signal to \aboveenvbreak to insert
+% \parskip glue after all.  Section titles are handled this way also.
+%
+\penalty 10001
 \endgroup
 \itemxneedsnegativevskipfalse
 \else
 % The item text fits into the space.  Start a paragraph, so that the
 % following text (if any) will end up on the same line.
 \endgroup
 \itemxneedsnegativevskiptrue
 \fi
 }
-\def\item{\errmessage{@item while not in a table}}
+\def\item{\errmessage{@item while not in a list environment}}
-\def\itemx{\errmessage{@itemx while not in a table}}
+\def\itemx{\errmessage{@itemx while not in a list environment}}
-\def\kitem{\errmessage{@kitem while not in a table}}
-\def\kitemx{\errmessage{@kitemx while not in a table}}
-\def\xitem{\errmessage{@xitem while not in a table}}
-\def\xitemx{\errmessage{@xitemx while not in a table}}
-% Contains a kludge to get @end[description] to work.
-\def\description{\tablez{\dontindex}{1}{}{}{}{}}
 % @table, @ftable, @vtable.
-\def\table{\begingroup\inENV\obeylines\obeyspaces\tablex}
+\envdef\table{%
-{\obeylines\obeyspaces%
+\let\itemindex\gobble
-\gdef\tablex #1^^M{%
+\tablecheck{table}%
-\tabley\dontindex#1        \endtabley}}
+}
+\envdef\ftable{%
-\def\ftable{\begingroup\inENV\obeylines\obeyspaces\ftablex}
+\def\itemindex ##1{\doind {fn}{\code{##1}}}%
-{\obeylines\obeyspaces%
+\tablecheck{ftable}%
-\gdef\ftablex #1^^M{%
+}
-\tabley\fnitemindex#1        \endtabley
+\envdef\vtable{%
-\def\Eftable{\endgraf\afterenvbreak\endgroup}%
+\def\itemindex ##1{\doind {vr}{\code{##1}}}%
-\let\Etable=\relax}}
+\tablecheck{vtable}%
+}
-\def\vtable{\begingroup\inENV\obeylines\obeyspaces\vtablex}
+\def\tablecheck#1{%
-{\obeylines\obeyspaces%
+\ifnum \the\catcode`\^^M=\active
-\gdef\vtablex #1^^M{%
+\endgroup
-\tabley\vritemindex#1        \endtabley
+\errmessage{This command won't work in this context; perhaps the problem is
-\def\Evtable{\endgraf\afterenvbreak\endgroup}%
+that we are \inenvironment\thisenv}%
-\let\Etable=\relax}}
+\def\next{\doignore{#1}}%
+\else
-\def\dontindex #1{}
+\let\next\tablex
-\def\fnitemindex #1{\doind {fn}{\code{#1}}}%
+\fi
-\def\vritemindex #1{\doind {vr}{\code{#1}}}%
+\next
+}
-{\obeyspaces %
+\def\tablex#1{%
-\gdef\tabley#1#2 #3 #4 #5 #6 #7\endtabley{\endgroup%
+\def\itemindicate{#1}%
-\tablez{#1}{#2}{#3}{#4}{#5}{#6}}}
+\parsearg\tabley
+}
-\def\tablez #1#2#3#4#5#6{%
+\def\tabley#1{%
-\aboveenvbreak %
+{%
-\begingroup %
+\makevalueexpandable
-\def\Edescription{\Etable}% Necessary kludge.
+\edef\temp{\noexpand\tablez #1\space\space\space}%
-\let\itemindex=#1%
+\expandafter
-\ifnum 0#3>0 \advance \leftskip by #3\mil \fi %
+}\temp \endtablez
-\ifnum 0#4>0 \tableindent=#4\mil \fi %
+}
-\ifnum 0#5>0 \advance \rightskip by #5\mil \fi %
+\def\tablez #1 #2 #3 #4\endtablez{%
-\def\itemfont{#2}%
+\aboveenvbreak
-\itemmax=\tableindent %
+\ifnum 0#1>0 \advance \leftskip by #1\mil \fi
-\advance \itemmax by -\itemmargin %
+\ifnum 0#2>0 \tableindent=#2\mil \fi
-\advance \leftskip by \tableindent %
+\ifnum 0#3>0 \advance \rightskip by #3\mil \fi
-\exdentamount=\tableindent
+\itemmax=\tableindent
-\parindent = 0pt
+\advance \itemmax by -\itemmargin
-\parskip = \smallskipamount
+\advance \leftskip by \tableindent
-\ifdim \parskip=0pt \parskip=2pt \fi%
+\exdentamount=\tableindent
-\def\Etable{\endgraf\afterenvbreak\endgroup}%
+\parindent = 0pt
-\let\item = \internalBitem %
+\parskip = \smallskipamount
-\let\itemx = \internalBitemx %
+\ifdim \parskip=0pt \parskip=2pt \fi
-\let\kitem = \internalBkitem %
+\let\item = \internalBitem
-\let\kitemx = \internalBkitemx %
+\let\itemx = \internalBitemx
-\let\xitem = \internalBxitem %
+}
-\let\xitemx = \internalBxitemx %
+\def\Etable{\endgraf\afterenvbreak}
-}
+\let\Eftable\Etable
+\let\Evtable\Etable
+\let\Eitemize\Etable
+\let\Eenumerate\Etable
 % This is the counter used by @enumerate, which is really @itemize
 \newcount \itemno
-\def\itemize{\parsearg\itemizezzz}
+\envdef\itemize{\parsearg\doitemize}
-\def\itemizezzz #1{%
+\def\doitemize#1{%
-\begingroup % ended by the @end itemize
+\aboveenvbreak
-\itemizey {#1}{\Eitemize}
+\itemmax=\itemindent
-}
+\advance\itemmax by -\itemmargin
+\advance\leftskip by \itemindent
-\def\itemizey #1#2{%
+\exdentamount=\itemindent
-\aboveenvbreak %
+\parindent=0pt
-\itemmax=\itemindent %
+\parskip=\smallskipamount
-\advance \itemmax by -\itemmargin %
+\ifdim\parskip=0pt \parskip=2pt \fi
-\advance \leftskip by \itemindent %
+\def\itemcontents{#1}%
-\exdentamount=\itemindent
+% @itemize with no arg is equivalent to @itemize @bullet.
-\parindent = 0pt %
+\ifx\itemcontents\empty\def\itemcontents{\bullet}\fi
-\parskip = \smallskipamount %
+\let\item=\itemizeitem
-\ifdim \parskip=0pt \parskip=2pt \fi%
+}
-\def#2{\endgraf\afterenvbreak\endgroup}%
-\def\itemcontents{#1}%
+% Definition of @item while inside @itemize and @enumerate.
-\let\item=\itemizeitem}
+%
+\def\itemizeitem{%
-% Set sfcode to normal for the chars that usually have another value.
+\advance\itemno by 1  % for enumerations
-% These are `.?!:;,'
+{\let\par=\endgraf \smallbreak}% reasonable place to break
-\def\frenchspacing{\sfcode46=1000 \sfcode63=1000 \sfcode33=1000
+{%
-\sfcode58=1000 \sfcode59=1000 \sfcode44=1000 }
+% If the document has an @itemize directly after a section title, a
+% \nobreak will be last on the list, and \sectionheading will have
+% done a \vskip-\parskip.  In that case, we don't want to zero
+% parskip, or the item text will crash with the heading.  On the
+% other hand, when there is normal text preceding the item (as there
+% usually is), we do want to zero parskip, or there would be too much
+% space.  In that case, we won't have a \nobreak before.  At least
+% that's the theory.
+\ifnum\lastpenalty<10000 \parskip=0in \fi
+\noindent
+\hbox to 0pt{\hss \itemcontents \kern\itemmargin}%
+\vadjust{\penalty 1200}}% not good to break after first line of item.
+\flushcr
+}
 % \splitoff TOKENS\endmark defines \first to be the first token in
 % TOKENS, and \rest to be the remainder.
 %
 \def\splitoff#1#2\endmark{\def\first{#1}\def\rest{#2}}%
 % Allow an optional argument of an uppercase letter, lowercase letter,
 % or number, to specify the first label in the enumerated list.  No
 % argument is the same as `1'.
 %
-\def\enumerate{\parsearg\enumeratezzz}
+\envparseargdef\enumerate{\enumeratey #1  \endenumeratey}
-\def\enumeratezzz #1{\enumeratey #1  \endenumeratey}
 \def\enumeratey #1 #2\endenumeratey{%
-\begingroup % ended by the @end enumerate
-%
 % If we were given no argument, pretend we were given `1'.
 \def\thearg{#1}%
 \ifx\thearg\empty \def\thearg{1}\fi
 %
 % Detect if the argument is a single token.  If so, it might be a
 \fi
 \char\uccode\itemno
 }%
 }
-% Call itemizey, adding a period to the first argument and supplying the
+% Call \doitemize, adding a period to the first argument and supplying the
 % common last two arguments.  Also subtract one from the initial value in
 % \itemno, since @item increments \itemno.
 %
 \def\startenumeration#1{%
 \advance\itemno by -1
-\itemizey{#1.}\Eenumerate\flushcr
+\doitemize{#1.}\flushcr
 }
 % @alphaenumerate and @capsenumerate are abbreviations for giving an arg
 % to @enumerate.
 %
 \def\alphaenumerate{\enumerate{a}}
 \def\capsenumerate{\enumerate{A}}
 \def\Ealphaenumerate{\Eenumerate}
 \def\Ecapsenumerate{\Eenumerate}
-% Definition of @item while inside @itemize.
-\def\itemizeitem{%
-\advance\itemno by 1
-{\let\par=\endgraf \smallbreak}%
-\ifhmode \errmessage{In hmode at itemizeitem}\fi
-{\parskip=0in \hskip 0pt
-\hbox to 0pt{\hss \itemcontents\hskip \itemmargin}%
-\vadjust{\penalty 1200}}%
-\flushcr}
 % @multitable macros
 % Amy Hendrickson, 8/18/94, 3/6/96
 %
 % @multitable ... @end multitable will make as many columns as desired.
 % Or use a template:
 %   @multitable {Column 1 template} {Column 2 template} {Column 3 template}
 %   @item ...
 %   using the widest term desired in each column.
-%
-% For those who want to use more than one line's worth of words in
-% the preamble, break the line within one argument and it
-% will parse correctly, i.e.,
-%
-%     @multitable {Column 1 template} {Column 2 template} {Column 3
-%      template}
-% Not:
-%     @multitable {Column 1 template} {Column 2 template}
-%      {Column 3 template}
 % Each new table line starts with @item, each subsequent new column
 % starts with @tab. Empty columns may be produced by supplying @tab's
 % with nothing between them for as many times as empty columns are needed,
 % ie, @tab@tab@tab will produce two empty columns.
-% @item, @tab, @multitable or @end multitable do not need to be on their
+% @item, @tab do not need to be on their own lines, but it will not hurt
-% own lines, but it will not hurt if they are.
+% if they are.
 % Sample multitable:
 %   @multitable {Column 1 template} {Column 2 template} {Column 3 template}
 %   @item first col stuff @tab second col stuff @tab third col
 \def\xendsetuptable{\endsetuptable}
 \let\columnfractions\relax
 \def\xcolumnfractions{\columnfractions}
 \newif\ifsetpercent
-% #1 is the part of the @columnfraction before the decimal point, which
+% #1 is the @columnfraction, usually a decimal number like .5, but might
-% is presumably either 0 or the empty string (but we don't check, we
+% be just 1.  We just use it, whatever it is.
-% just throw it away).  #2 is the decimal part, which we use as the
+%
-% percent of \hsize for this column.
+\def\pickupwholefraction#1 {%
-\def\pickupwholefraction#1.#2 {%
 \global\advance\colcount by 1
-\expandafter\xdef\csname col\the\colcount\endcsname{.#2\hsize}%
+\expandafter\xdef\csname col\the\colcount\endcsname{#1\hsize}%
 \setuptable
 }
 \newcount\colcount
 \def\setuptable#1{%
 \else
 \ifsetpercent
 \let\go\pickupwholefraction
 \else
 \global\advance\colcount by 1
-\setbox0=\hbox{#1\unskip }% Add a normal word space as a separator;
+\setbox0=\hbox{#1\unskip\space}% Add a normal word space as a
-% typically that is always in the input, anyway.
+% separator; typically that is always in the input, anyway.
 \expandafter\xdef\csname col\the\colcount\endcsname{\the\wd0}%
 \fi
 \fi
 \ifx\go\pickupwholefraction
 % Put the argument back for the \pickupwholefraction call, so
 \fi%
 \fi
 \go
 }
-% This used to have \hskip1sp.  But then the space in a template line is
+% multitable-only commands.
-% not enough.  That is bad.  So let's go back to just & until we
+%
-% encounter the problem it was intended to solve again.
+% @headitem starts a heading row, which we typeset in bold.
-% --karl, nathan@acm.org, 20apr99.
+% Assignments have to be global since we are inside the implicit group
-\def\tab{&}
+% of an alignment entry.  Note that \everycr resets \everytab.
+\def\headitem{\checkenv\multitable \crcr \global\everytab={\bf}\the\everytab}%
+%
+% A \tab used to include \hskip1sp.  But then the space in a template
+% line is not enough.  That is bad.  So let's go back to just `&' until
+% we encounter the problem it was intended to solve again.
+%					--karl, nathan@acm.org, 20apr99.
+\def\tab{\checkenv\multitable &\the\everytab}%
 % @multitable ... @end multitable definitions:
 %
-\def\multitable{\parsearg\dotable}
+\newtoks\everytab  % insert after every tab.
-\def\dotable#1{\bgroup
+%
+\envdef\multitable{%
 \vskip\parskip
-\let\item\crcr
+\startsavinginserts
+%
+% @item within a multitable starts a normal row.
+% We use \def instead of \let so that if one of the multitable entries
+% contains an @itemize, we don't choke on the \item (seen as \crcr aka
+% \endtemplate) expanding \doitemize.
+\def\item{\crcr}%
+%
 \tolerance=9500
 \hbadness=9500
 \setmultitablespacing
 \parskip=\multitableparskip
 \parindent=\multitableparindent
 \overfullrule=0pt
 \global\colcount=0
-\def\Emultitable{\global\setpercentfalse\cr\egroup\egroup}%
+%
-%
+\everycr = {%
+\noalign{%
+\global\everytab={}%
+\global\colcount=0 % Reset the column counter.
+% Check for saved footnotes, etc.
+\checkinserts
+% Keeps underfull box messages off when table breaks over pages.
+%\filbreak
+	% Maybe so, but it also creates really weird page breaks when the
+	% table breaks over pages. Wouldn't \vfil be better?  Wait until the
+	% problem manifests itself, so it can be fixed for real --karl.
+}%
+}%
+%
+\parsearg\domultitable
+}
+\def\domultitable#1{%
 % To parse everything between @multitable and @item:
 \setuptable#1 \endsetuptable
-%
-% \everycr will reset column counter, \colcount, at the end of
-% each line. Every column entry will cause \colcount to advance by one.
-% The table preamble
-% looks at the current \colcount to find the correct column width.
-\everycr{\noalign{%
-%
-% \filbreak%% keeps underfull box messages off when table breaks over pages.
-% Maybe so, but it also creates really weird page breaks when the table
-% breaks over pages. Wouldn't \vfil be better?  Wait until the problem
-% manifests itself, so it can be fixed for real --karl.
-\global\colcount=0\relax}}%
 %
 % This preamble sets up a generic column definition, which will
 % be used as many times as user calls for columns.
 % \vtop will set a single line and will also let text wrap and
 % continue for many paragraphs if desired.
-\halign\bgroup&\global\advance\colcount by 1\relax
+\halign\bgroup &%
-\multistrut\vtop{\hsize=\expandafter\csname col\the\colcount\endcsname
+\global\advance\colcount by 1
-%
+\multistrut
-% In order to keep entries from bumping into each other
+\vtop{%
-% we will add a \leftskip of \multitablecolspace to all columns after
+% Use the current \colcount to find the correct column width:
-% the first one.
+\hsize=\expandafter\csname col\the\colcount\endcsname
 %
-% If a template has been used, we will add \multitablecolspace
+% In order to keep entries from bumping into each other
-% to the width of each template entry.
+% we will add a \leftskip of \multitablecolspace to all columns after
-%
+% the first one.
-% If the user has set preamble in terms of percent of \hsize we will
+%
-% use that dimension as the width of the column, and the \leftskip
+% If a template has been used, we will add \multitablecolspace
-% will keep entries from bumping into each other.  Table will start at
+% to the width of each template entry.
-% left margin and final column will justify at right margin.
+%
-%
+% If the user has set preamble in terms of percent of \hsize we will
-% Make sure we don't inherit \rightskip from the outer environment.
+% use that dimension as the width of the column, and the \leftskip
-\rightskip=0pt
+% will keep entries from bumping into each other.  Table will start at
-\ifnum\colcount=1
+% left margin and final column will justify at right margin.
-% The first column will be indented with the surrounding text.
+%
-\advance\hsize by\leftskip
+% Make sure we don't inherit \rightskip from the outer environment.
-\else
+\rightskip=0pt
-\ifsetpercent \else
+\ifnum\colcount=1
-% If user has not set preamble in terms of percent of \hsize
+	% The first column will be indented with the surrounding text.
-% we will advance \hsize by \multitablecolspace.
+	\advance\hsize by\leftskip
-\advance\hsize by \multitablecolspace
+\else
-\fi
+	\ifsetpercent \else
-% In either case we will make \leftskip=\multitablecolspace:
+	  % If user has not set preamble in terms of percent of \hsize
-\leftskip=\multitablecolspace
+	  % we will advance \hsize by \multitablecolspace.
-\fi
+	  \advance\hsize by \multitablecolspace
-% Ignoring space at the beginning and end avoids an occasional spurious
+	\fi
-% blank line, when TeX decides to break the line at the space before the
+% In either case we will make \leftskip=\multitablecolspace:
-% box from the multistrut, so the strut ends up on a line by itself.
+\leftskip=\multitablecolspace
-% For example:
+\fi
-% @multitable @columnfractions .11 .89
+% Ignoring space at the beginning and end avoids an occasional spurious
-% @item @code{#}
+% blank line, when TeX decides to break the line at the space before the
-% @tab Legal holiday which is valid in major parts of the whole country.
+% box from the multistrut, so the strut ends up on a line by itself.
-% Is automatically provided with highlighting sequences respectively marking
+% For example:
-% characters.
+% @multitable @columnfractions .11 .89
-\noindent\ignorespaces##\unskip\multistrut}\cr
+% @item @code{#}
-}
+% @tab Legal holiday which is valid in major parts of the whole country.
+% Is automatically provided with highlighting sequences respectively
-\def\setmultitablespacing{% test to see if user has set \multitablelinespace.
+% marking characters.
-% If so, do nothing. If not, give it an appropriate dimension based on
+\noindent\ignorespaces##\unskip\multistrut
-% current baselineskip.
+}\cr
+}
+\def\Emultitable{%
+\crcr
+\egroup % end the \halign
+\global\setpercentfalse
+}
+\def\setmultitablespacing{%
+\def\multistrut{\strut}% just use the standard line spacing
+%
+% Compute \multitablelinespace (if not defined by user) for use in
+% \multitableparskip calculation.  We used define \multistrut based on
+% this, but (ironically) that caused the spacing to be off.
+% See bug-texinfo report from Werner Lemberg, 31 Oct 2004 12:52:20 +0100.
 \ifdim\multitablelinespace=0pt
 \setbox0=\vbox{X}\global\multitablelinespace=\the\baselineskip
 \global\advance\multitablelinespace by-\ht0
-%% strut to put in table in case some entry doesn't have descenders,
+\fi
-%% to keep lines equally spaced
-\let\multistrut = \strut
-\else
-%% FIXME: what is \box0 supposed to be?
-\gdef\multistrut{\vrule height\multitablelinespace depth\dp0
-width0pt\relax} \fi
 %% Test to see if parskip is larger than space between lines of
 %% table. If not, do nothing.
 %%        If so, set to same dimension as multitablelinespace.
 \ifdim\multitableparskip>\multitablelinespace
 \global\multitableparskip=\multitablelinespace
 %% than skip between lines in the table.
 \fi}
 \message{conditionals,}
-% Prevent errors for section commands.
-% Used in @ignore and in failing conditionals.
+% @iftex, @ifnotdocbook, @ifnothtml, @ifnotinfo, @ifnotplaintext,
-\def\ignoresections{%
+% @ifnotxml always succeed.  They currently do nothing; we don't
-\let\chapter=\relax
+% attempt to check whether the conditionals are properly nested.  But we
-\let\unnumbered=\relax
+% have to remember that they are conditionals, so that @end doesn't
-\let\top=\relax
+% attempt to close an environment group.
-\let\unnumberedsec=\relax
+%
-\let\unnumberedsection=\relax
+\def\makecond#1{%
-\let\unnumberedsubsec=\relax
+\expandafter\let\csname #1\endcsname = \relax
-\let\unnumberedsubsection=\relax
+\expandafter\let\csname iscond.#1\endcsname = 1
-\let\unnumberedsubsubsec=\relax
+}
-\let\unnumberedsubsubsection=\relax
+\makecond{iftex}
-\let\section=\relax
+\makecond{ifnotdocbook}
-\let\subsec=\relax
+\makecond{ifnothtml}
-\let\subsubsec=\relax
+\makecond{ifnotinfo}
-\let\subsection=\relax
+\makecond{ifnotplaintext}
-\let\subsubsection=\relax
+\makecond{ifnotxml}
-\let\appendix=\relax
-\let\appendixsec=\relax
+% Ignore @ignore, @ifhtml, @ifinfo, and the like.
-\let\appendixsection=\relax
+%
-\let\appendixsubsec=\relax
+\def\direntry{\doignore{direntry}}
-\let\appendixsubsection=\relax
+\def\documentdescription{\doignore{documentdescription}}
-\let\appendixsubsubsec=\relax
+\def\docbook{\doignore{docbook}}
-\let\appendixsubsubsection=\relax
+\def\html{\doignore{html}}
-\let\contents=\relax
+\def\ifdocbook{\doignore{ifdocbook}}
-\let\smallbook=\relax
+\def\ifhtml{\doignore{ifhtml}}
-\let\titlepage=\relax
+\def\ifinfo{\doignore{ifinfo}}
-}
+\def\ifnottex{\doignore{ifnottex}}
+\def\ifplaintext{\doignore{ifplaintext}}
-% Used in nested conditionals, where we have to parse the Texinfo source
+\def\ifxml{\doignore{ifxml}}
-% and so want to turn off most commands, in case they are used
-% incorrectly.
-%
-\def\ignoremorecommands{%
-\let\defcodeindex = \relax
-\let\defcv = \relax
-\let\deffn = \relax
-\let\deffnx = \relax
-\let\defindex = \relax
-\let\defivar = \relax
-\let\defmac = \relax
-\let\defmethod = \relax
-\let\defop = \relax
-\let\defopt = \relax
-\let\defspec = \relax
-\let\deftp = \relax
-\let\deftypefn = \relax
-\let\deftypefun = \relax
-\let\deftypeivar = \relax
-\let\deftypeop = \relax
-\let\deftypevar = \relax
-\let\deftypevr = \relax
-\let\defun = \relax
-\let\defvar = \relax
-\let\defvr = \relax
-\let\ref = \relax
-\let\xref = \relax
-\let\printindex = \relax
-\let\pxref = \relax
-\let\settitle = \relax
-\let\setchapternewpage = \relax
-\let\setchapterstyle = \relax
-\let\everyheading = \relax
-\let\evenheading = \relax
-\let\oddheading = \relax
-\let\everyfooting = \relax
-\let\evenfooting = \relax
-\let\oddfooting = \relax
-\let\headings = \relax
-\let\include = \relax
-\let\lowersections = \relax
-\let\down = \relax
-\let\raisesections = \relax
-\let\up = \relax
-\let\set = \relax
-\let\clear = \relax
-\let\item = \relax
-}
-% Ignore @ignore ... @end ignore.
-%
 \def\ignore{\doignore{ignore}}
-% Ignore @ifinfo, @ifhtml, @ifnottex, @html, @menu, and @direntry text.
-%
-\def\ifinfo{\doignore{ifinfo}}
-\def\ifhtml{\doignore{ifhtml}}
-\def\ifnottex{\doignore{ifnottex}}
-\def\html{\doignore{html}}
 \def\menu{\doignore{menu}}
-\def\direntry{\doignore{direntry}}
+\def\xml{\doignore{xml}}
-% @dircategory CATEGORY  -- specify a category of the dir file
+% Ignore text until a line `@end #1', keeping track of nested conditionals.
-% which this file should belong to.  Ignore this in TeX.
+%
-\let\dircategory = \comment
+% A count to remember the depth of nesting.
+\newcount\doignorecount
-% Ignore text until a line `@end #1'.
-%
 \def\doignore#1{\begingroup
-% Don't complain about control sequences we have declared \outer.
+% Scan in ``verbatim'' mode:
-\ignoresections
+\obeylines
-%
+\catcode`\@ = \other
-% Define a command to swallow text until we reach `@end #1'.
+\catcode`\{ = \other
-% This @ is a catcode 12 token (that is the normal catcode of @ in
+\catcode`\} = \other
-% this texinfo.tex file).  We change the catcode of @ below to match.
-\long\def\doignoretext##1@end #1{\enddoignore}%
 %
 % Make sure that spaces turn into tokens that match what \doignoretext wants.
-\catcode32 = 10
+\spaceisspace
 %
-% Ignore braces, too, so mismatched braces don't cause trouble.
+% Count number of #1's that we've seen.
-\catcode`\{ = 9
+\doignorecount = 0
-\catcode`\} = 9
+%
-%
+% Swallow text until we reach the matching `@end #1'.
-% We must not have @c interpreted as a control sequence.
+\dodoignore{#1}%
-\catcode`\@ = 12
+}
-%
-% Make the letter c a comment character so that the rest of the line
+{ \catcode`_=11 % We want to use \_STOP_ which cannot appear in texinfo source.
-% will be ignored. This way, the document can have (for example)
+\obeylines %
-%   @c @end ifinfo
+%
-% and the @end ifinfo will be properly ignored.
+\gdef\dodoignore#1{%
-% (We've just changed @ to catcode 12.)
+% #1 contains the command name as a string, e.g., `ifinfo'.
-\catcode`\c = 14
-%
-% And now expand that command.
-\doignoretext
-}
-% What we do to finish off ignored text.
-%
-\def\enddoignore{\endgroup\ignorespaces}%
-\newif\ifwarnedobs\warnedobsfalse
-\def\obstexwarn{%
-\ifwarnedobs\relax\else
-% We need to warn folks that they may have trouble with TeX 3.0.
-% This uses \immediate\write16 rather than \message to get newlines.
-\immediate\write16{}
-\immediate\write16{WARNING: for users of Unix TeX 3.0!}
-\immediate\write16{This manual trips a bug in TeX version 3.0 (tex hangs).}
-\immediate\write16{If you are running another version of TeX, relax.}
-\immediate\write16{If you are running Unix TeX 3.0, kill this TeX process.}
-\immediate\write16{  Then upgrade your TeX installation if you can.}
-\immediate\write16{  (See ftp://ftp.gnu.org/pub/gnu/TeX.README.)}
-\immediate\write16{If you are stuck with version 3.0, run the}
-\immediate\write16{  script ``tex3patch'' from the Texinfo distribution}
-\immediate\write16{  to use a workaround.}
-\immediate\write16{}
-\global\warnedobstrue
-\fi
-}
-% **In TeX 3.0, setting text in \nullfont hangs tex.  For a
-% workaround (which requires the file ``dummy.tfm'' to be installed),
-% uncomment the following line:
-%%%%%\font\nullfont=dummy\let\obstexwarn=\relax
-% Ignore text, except that we keep track of conditional commands for
-% purposes of nesting, up to an `@end #1' command.
-%
-\def\nestedignore#1{%
-\obstexwarn
-% We must actually expand the ignored text to look for the @end
-% command, so that nested ignore constructs work.  Thus, we put the
-% text into a \vbox and then do nothing with the result.  To minimize
-% the change of memory overflow, we follow the approach outlined on
-% page 401 of the TeXbook: make the current font be a dummy font.
-%
-\setbox0 = \vbox\bgroup
-% Don't complain about control sequences we have declared \outer.
-\ignoresections
 %
-% Define `@end #1' to end the box, which will in turn undefine the
+% Define a command to find the next `@end #1'.
-% @end command again.
+\long\def\doignoretext##1^^M@end #1{%
-\expandafter\def\csname E#1\endcsname{\egroup\ignorespaces}%
+\doignoretextyyy##1^^M@#1\_STOP_}%
 %
-% We are going to be parsing Texinfo commands.  Most cause no
+% And this command to find another #1 command, at the beginning of a
-% trouble when they are used incorrectly, but some commands do
+% line.  (Otherwise, we would consider a line `@c @ifset', for
-% complicated argument parsing or otherwise get confused, so we
+% example, to count as an @ifset for nesting.)
-% undefine them.
+\long\def\doignoretextyyy##1^^M@#1##2\_STOP_{\doignoreyyy{##2}\_STOP_}%
 %
-% We can't do anything about stray @-signs, unfortunately;
+% And now expand that command.
-% they'll produce `undefined control sequence' errors.
+\doignoretext ^^M%
-\ignoremorecommands
+}%
-%
+}
-% Set the current font to be \nullfont, a TeX primitive, and define
-% all the font commands to also use \nullfont.  We don't use
+\def\doignoreyyy#1{%
-% dummy.tfm, as suggested in the TeXbook, because not all sites
+\def\temp{#1}%
-% might have that installed.  Therefore, math mode will still
+\ifx\temp\empty			% Nothing found.
-% produce output, but that should be an extremely small amount of
+\let\next\doignoretextzzz
-% stuff compared to the main input.
+\else					% Found a nested condition, ...
-%
+\advance\doignorecount by 1
-\nullfont
+\let\next\doignoretextyyy		% ..., look for another.
-\let\tenrm=\nullfont \let\tenit=\nullfont \let\tensl=\nullfont
+% If we're here, #1 ends with ^^M\ifinfo (for example).
-\let\tenbf=\nullfont \let\tentt=\nullfont \let\smallcaps=\nullfont
+\fi
-\let\tensf=\nullfont
+\next #1% the token \_STOP_ is present just after this macro.
-% Similarly for index fonts (mostly for their use in smallexample).
+}
-\let\smallrm=\nullfont \let\smallit=\nullfont \let\smallsl=\nullfont
-\let\smallbf=\nullfont \let\smalltt=\nullfont \let\smallsc=\nullfont
+% We have to swallow the remaining "\_STOP_".
-\let\smallsf=\nullfont
+%
-%
+\def\doignoretextzzz#1{%
-% Don't complain when characters are missing from the fonts.
+\ifnum\doignorecount = 0	% We have just found the outermost @end.
-\tracinglostchars = 0
+\let\next\enddoignore
-%
+\else				% Still inside a nested condition.
-% Don't bother to do space factor calculations.
+\advance\doignorecount by -1
-\frenchspacing
+\let\next\doignoretext      % Look for the next @end.
-%
+\fi
-% Don't report underfull hboxes.
+\next
-\hbadness = 10000
+}
-%
-% Do minimal line-breaking.
+% Finish off ignored text.
-\pretolerance = 10000
+{ \obeylines%
-%
+% Ignore anything after the last `@end #1'; this matters in verbatim
-% Do not execute instructions in @tex
+% environments, where otherwise the newline after an ignored conditional
-\def\tex{\doignore{tex}}%
+% would result in a blank line in the output.
-% Do not execute macro definitions.
+\gdef\enddoignore#1^^M{\endgroup\ignorespaces}%
-% `c' is a comment character, so the word `macro' will get cut off.
+}
-\def\macro{\doignore{ma}}%
-}
 % @set VAR sets the variable VAR to an empty value.
 % @set VAR REST-OF-LINE sets VAR to the value REST-OF-LINE.
 %
 % Since we want to separate VAR from REST-OF-LINE (which might be
 % empty), we can't just use \parsearg; we have to insert a space of our
 % own to delimit the rest of the line, and then take it out again if we
-% didn't need it.  Make sure the catcode of space is correct to avoid
+% didn't need it.
-% losing inside @example, for instance.
+% We rely on the fact that \parsearg sets \catcode`\ =10.
 %
-\def\set{\begingroup\catcode` =10
+\parseargdef\set{\setyyy#1 \endsetyyy}
-\catcode`\-=12 \catcode`\_=12 % Allow - and _ in VAR.
-\parsearg\setxxx}
-\def\setxxx#1{\setyyy#1 \endsetyyy}
 \def\setyyy#1 #2\endsetyyy{%
-\def\temp{#2}%
+{%
-\ifx\temp\empty \global\expandafter\let\csname SET#1\endcsname = \empty
+\makevalueexpandable
-\else \setzzz{#1}#2\endsetzzz % Remove the trailing space \setxxx inserted.
+\def\temp{#2}%
-\fi
+\edef\next{\gdef\makecsname{SET#1}}%
-\endgroup
+\ifx\temp\empty
-}
+\next{}%
-% Can't use \xdef to pre-expand #2 and save some time, since \temp or
+\else
-% \next or other control sequences that we've defined might get us into
+\setzzz#2\endsetzzz
-% an infinite loop. Consider `@set foo @cite{bar}'.
+\fi
-\def\setzzz#1#2 \endsetzzz{\expandafter\gdef\csname SET#1\endcsname{#2}}
+}%
+}
+% Remove the trailing space \setxxx inserted.
+\def\setzzz#1 \endsetzzz{\next{#1}}
 % @clear VAR clears (i.e., unsets) the variable VAR.
 %
-\def\clear{\parsearg\clearxxx}
+\parseargdef\clear{%
-\def\clearxxx#1{\global\expandafter\let\csname SET#1\endcsname=\relax}
+{%
+\makevalueexpandable
+\global\expandafter\let\csname SET#1\endcsname=\relax
+}%
+}
 % @value{foo} gets the text saved in variable foo.
+\def\value{\begingroup\makevalueexpandable\valuexxx}
+\def\valuexxx#1{\expandablevalue{#1}\endgroup}
 {
-\catcode`\_ = \active
+\catcode`\- = \active \catcode`\_ = \active
 %
-% We might end up with active _ or - characters in the argument if
+\gdef\makevalueexpandable{%
-% we're called from @code, as @code{@value{foo-bar_}}.  So \let any
+\let\value = \expandablevalue
-% such active characters to their normal equivalents.
+% We don't want these characters active, ...
-\gdef\value{\begingroup
+\catcode`\-=\other \catcode`\_=\other
-\catcode`\-=12 \catcode`\_=12
+% ..., but we might end up with active ones in the argument if
-\indexbreaks \let_\normalunderscore
+% we're called from @code, as @code{@value{foo-bar_}}, though.
-\valuexxx}
+% So \let them to their normal equivalents.
-}
+\let-\realdash \let_\normalunderscore
-\def\valuexxx#1{\expandablevalue{#1}\endgroup}
+}
+}
 % We have this subroutine so that we can handle at least some @value's
-% properly in indexes (we \let\value to this in \indexdummies).  Ones
+% properly in indexes (we call \makevalueexpandable in \indexdummies).
-% whose names contain - or _ still won't work, but we can't do anything
+% The command has to be fully expandable (if the variable is set), since
-% about that.  The command has to be fully expandable, since the result
+% the result winds up in the index file.  This means that if the
-% winds up in the index file.  This means that if the variable's value
+% variable's value contains other Texinfo commands, it's almost certain
-% contains other Texinfo commands, it's almost certain it will fail
+% it will fail (although perhaps we could fix that with sufficient work
-% (although perhaps we could fix that with sufficient work to do a
+% to do a one-level expansion on the result, instead of complete).
-% one-level expansion on the result, instead of complete).
 %
 \def\expandablevalue#1{%
 \expandafter\ifx\csname SET#1\endcsname\relax
 {[No value for ``#1'']}%
+\message{Variable `#1', used in @value, is not set.}%
 \else
 \csname SET#1\endcsname
 \fi
 }
 % @ifset VAR ... @end ifset reads the `...' iff VAR has been defined
 % with @set.
 %
-\def\ifset{\parsearg\ifsetxxx}
+% To get special treatment of `@end ifset,' call \makeond and the redefine.
-\def\ifsetxxx #1{%
+%
-\expandafter\ifx\csname SET#1\endcsname\relax
+\makecond{ifset}
-\expandafter\ifsetfail
+\def\ifset{\parsearg{\doifset{\let\next=\ifsetfail}}}
-\else
+\def\doifset#1#2{%
-\expandafter\ifsetsucceed
+{%
-\fi
+\makevalueexpandable
-}
+\let\next=\empty
-\def\ifsetsucceed{\conditionalsucceed{ifset}}
+\expandafter\ifx\csname SET#2\endcsname\relax
-\def\ifsetfail{\nestedignore{ifset}}
+#1% If not set, redefine \next.
-\defineunmatchedend{ifset}
+\fi
+\expandafter
+}\next
+}
+\def\ifsetfail{\doignore{ifset}}
 % @ifclear VAR ... @end ifclear reads the `...' iff VAR has never been
 % defined with @set, or has been undefined with @clear.
 %
-\def\ifclear{\parsearg\ifclearxxx}
+% The `\else' inside the `\doifset' parameter is a trick to reuse the
-\def\ifclearxxx #1{%
+% above code: if the variable is not set, do nothing, if it is set,
-\expandafter\ifx\csname SET#1\endcsname\relax
+% then redefine \next to \ifclearfail.
-\expandafter\ifclearsucceed
+%
-\else
+\makecond{ifclear}
-\expandafter\ifclearfail
+\def\ifclear{\parsearg{\doifset{\else \let\next=\ifclearfail}}}
-\fi
+\def\ifclearfail{\doignore{ifclear}}
-}
-\def\ifclearsucceed{\conditionalsucceed{ifclear}}
+% @dircategory CATEGORY  -- specify a category of the dir file
-\def\ifclearfail{\nestedignore{ifclear}}
+% which this file should belong to.  Ignore this in TeX.
-\defineunmatchedend{ifclear}
+\let\dircategory=\comment
-% @iftex, @ifnothtml, @ifnotinfo always succeed; we read the text
-% following, through the first @end iftex (etc.).  Make `@end iftex'
-% (etc.) valid only after an @iftex.
-%
-\def\iftex{\conditionalsucceed{iftex}}
-\def\ifnothtml{\conditionalsucceed{ifnothtml}}
-\def\ifnotinfo{\conditionalsucceed{ifnotinfo}}
-\defineunmatchedend{iftex}
-\defineunmatchedend{ifnothtml}
-\defineunmatchedend{ifnotinfo}
-% We can't just want to start a group at @iftex (for example) and end it
-% at @end iftex, since then @set commands inside the conditional have no
-% effect (they'd get reverted at the end of the group).  So we must
-% define \Eiftex to redefine itself to be its previous value.  (We can't
-% just define it to fail again with an ``unmatched end'' error, since
-% the @ifset might be nested.)
-%
-\def\conditionalsucceed#1{%
-\edef\temp{%
-% Remember the current value of \E#1.
-\let\nece{prevE#1} = \nece{E#1}%
-%
-% At the `@end #1', redefine \E#1 to be its previous value.
-\def\nece{E#1}{\let\nece{E#1} = \nece{prevE#1}}%
-}%
-\temp
-}
-% We need to expand lots of \csname's, but we don't want to expand the
-% control sequences after we've constructed them.
-%
-\def\nece#1{\expandafter\noexpand\csname#1\endcsname}
 % @defininfoenclose.
 \let\definfoenclose=\comment
 \message{indexing,}
 % Index generation facilities
 % Define \newwrite to be identical to plain tex's \newwrite
-% except not \outer, so it can be used within \newindex.
+% except not \outer, so it can be used within macros and \if's.
-{\catcode`\@=11
+\edef\newwrite{\makecsname{ptexnewwrite}}
-\gdef\newwrite{\alloc@7\write\chardef\sixt@@n}}
 % \newindex {foo} defines an index named foo.
 % It automatically defines \fooindex such that
 % \fooindex ...rest of line... puts an entry in the index foo.
 % It also defines \fooindfile to be the number of the output channel for
 \expandafter\xdef\csname#1index\endcsname{%     % Define @#1index
 \noexpand\doindex{#1}}
 }
 % @defindex foo  ==  \newindex{foo}
+%
 \def\defindex{\parsearg\newindex}
 % Define @defcodeindex, like @defindex except put all entries in @code.
+%
+\def\defcodeindex{\parsearg\newcodeindex}
+%
 \def\newcodeindex#1{%
 \iflinks
 \expandafter\newwrite \csname#1indfile\endcsname
 \openout \csname#1indfile\endcsname \jobname.#1
 \fi
 \expandafter\xdef\csname#1index\endcsname{%
-\noexpand\docodeindex{#1}}
+\noexpand\docodeindex{#1}}%
 }
-\def\defcodeindex{\parsearg\newcodeindex}
 % @synindex foo bar    makes index foo feed into index bar.
 % Do this instead of @defindex foo if you don't want it as a separate index.
-% The \closeout helps reduce unnecessary open files; the limit on the
+%
-% Acorn RISC OS is a mere 16 files.
-\def\synindex#1 #2 {%
-\expandafter\let\expandafter\synindexfoo\expandafter=\csname#2indfile\endcsname
-\expandafter\closeout\csname#1indfile\endcsname
-\expandafter\let\csname#1indfile\endcsname=\synindexfoo
-\expandafter\xdef\csname#1index\endcsname{% define \xxxindex
-\noexpand\doindex{#2}}%
-}
 % @syncodeindex foo bar   similar, but put all entries made for index foo
 % inside @code.
-\def\syncodeindex#1 #2 {%
+%
-\expandafter\let\expandafter\synindexfoo\expandafter=\csname#2indfile\endcsname
+\def\synindex#1 #2 {\dosynindex\doindex{#1}{#2}}
-\expandafter\closeout\csname#1indfile\endcsname
+\def\syncodeindex#1 #2 {\dosynindex\docodeindex{#1}{#2}}
-\expandafter\let\csname#1indfile\endcsname=\synindexfoo
-\expandafter\xdef\csname#1index\endcsname{% define \xxxindex
+% #1 is \doindex or \docodeindex, #2 the index getting redefined (foo),
-\noexpand\docodeindex{#2}}%
+% #3 the target index (bar).
+\def\dosynindex#1#2#3{%
+% Only do \closeout if we haven't already done it, else we'll end up
+% closing the target index.
+\expandafter \ifx\csname donesynindex#2\endcsname \undefined
+% The \closeout helps reduce unnecessary open files; the limit on the
+% Acorn RISC OS is a mere 16 files.
+\expandafter\closeout\csname#2indfile\endcsname
+\expandafter\let\csname\donesynindex#2\endcsname = 1
+\fi
+% redefine \fooindfile:
+\expandafter\let\expandafter\temp\expandafter=\csname#3indfile\endcsname
+\expandafter\let\csname#2indfile\endcsname=\temp
+% redefine \fooindex:
+\expandafter\xdef\csname#2index\endcsname{\noexpand#1{#3}}%
 }
 % Define \doindex, the driver for all \fooindex macros.
 % Argument #1 is generated by the calling \fooindex macro,
 %  and it is "foo", the name of the index.
 % like the previous two, but they put @code around the argument.
 \def\docodeindex#1{\edef\indexname{#1}\parsearg\singlecodeindexer}
 \def\singlecodeindexer #1{\doind{\indexname}{\code{#1}}}
+% Take care of Texinfo commands that can appear in an index entry.
+% Since there are some commands we want to expand, and others we don't,
+% we have to laboriously prevent expansion for those that we don't.
+%
 \def\indexdummies{%
-\def\ { }%
+\escapechar = `\\     % use backslash in output files.
-% Take care of the plain tex accent commands.
+\def\@{@}% change to @@ when we switch to @ as escape char in index files.
-\def\"{\realbackslash "}%
+\def\ {\realbackslash\space }%
-\def\`{\realbackslash `}%
+%
-\def\'{\realbackslash '}%
+% Need these in case \tex is in effect and \{ is a \delimiter again.
-\def\^{\realbackslash ^}%
+% But can't use \lbracecmd and \rbracecmd because texindex assumes
-\def\~{\realbackslash ~}%
+% braces and backslashes are used only as delimiters.
-\def\={\realbackslash =}%
+\let\{ = \mylbrace
-\def\b{\realbackslash b}%
+\let\} = \myrbrace
-\def\c{\realbackslash c}%
+%
-\def\d{\realbackslash d}%
+% I don't entirely understand this, but when an index entry is
-\def\u{\realbackslash u}%
+% generated from a macro call, the \endinput which \scanmacro inserts
-\def\v{\realbackslash v}%
+% causes processing to be prematurely terminated.  This is,
-\def\H{\realbackslash H}%
+% apparently, because \indexsorttmp is fully expanded, and \endinput
-% Take care of the plain tex special European modified letters.
+% is an expandable command.  The redefinition below makes \endinput
-\def\oe{\realbackslash oe}%
+% disappear altogether for that purpose -- although logging shows that
-\def\ae{\realbackslash ae}%
+% processing continues to some further point.  On the other hand, it
-\def\aa{\realbackslash aa}%
+% seems \endinput does not hurt in the printed index arg, since that
-\def\OE{\realbackslash OE}%
+% is still getting written without apparent harm.
-\def\AE{\realbackslash AE}%
+%
-\def\AA{\realbackslash AA}%
+% Sample source (mac-idx3.tex, reported by Graham Percival to
-\def\o{\realbackslash o}%
+% help-texinfo, 22may06):
-\def\O{\realbackslash O}%
+% @macro funindex {WORD}
-\def\l{\realbackslash l}%
+% @findex xyz
-\def\L{\realbackslash L}%
+% @end macro
-\def\ss{\realbackslash ss}%
+% ...
-% Take care of texinfo commands likely to appear in an index entry.
+% @funindex commtest
-% (Must be a way to avoid doing expansion at all, and thus not have to
+%
-% laboriously list every single command here.)
+% The above is not enough to reproduce the bug, but it gives the flavor.
-\def\@{@}% will be @@ when we switch to @ as escape char.
+%
-% Need these in case \tex is in effect and \{ is a \delimiter again.
+% Sample whatsit resulting:
-% But can't use \lbracecmd and \rbracecmd because texindex assumes
+% .@write3{\entry{xyz}{@folio }{@code {xyz@endinput }}}
-% braces and backslashes are used only as delimiters.
+%
-\let\{ = \mylbrace
+% So:
-\let\} = \myrbrace
+\let\endinput = \empty
-\def\_{{\realbackslash _}}%
+%
-\def\w{\realbackslash w }%
+% Do the redefinitions.
-\def\bf{\realbackslash bf }%
+\commondummies
-%\def\rm{\realbackslash rm }%
+}
-\def\sl{\realbackslash sl }%
-\def\sf{\realbackslash sf}%
+% For the aux and toc files, @ is the escape character.  So we want to
-\def\tt{\realbackslash tt}%
+% redefine everything using @ as the escape character (instead of
-\def\gtr{\realbackslash gtr}%
+% \realbackslash, still used for index files).  When everything uses @,
-\def\less{\realbackslash less}%
+% this will be simpler.
-\def\hat{\realbackslash hat}%
+%
-\def\TeX{\realbackslash TeX}%
+\def\atdummies{%
-\def\dots{\realbackslash dots }%
+\def\@{@@}%
-\def\result{\realbackslash result}%
+\def\ {@ }%
-\def\equiv{\realbackslash equiv}%
+\let\{ = \lbraceatcmd
-\def\expansion{\realbackslash expansion}%
+\let\} = \rbraceatcmd
-\def\print{\realbackslash print}%
+%
-\def\error{\realbackslash error}%
+% Do the redefinitions.
-\def\point{\realbackslash point}%
+\commondummies
-\def\copyright{\realbackslash copyright}%
+\otherbackslash
-\def\tclose##1{\realbackslash tclose {##1}}%
+}
-\def\code##1{\realbackslash code {##1}}%
-\def\uref##1{\realbackslash uref {##1}}%
+% Called from \indexdummies and \atdummies.
-\def\url##1{\realbackslash url {##1}}%
+%
-\def\env##1{\realbackslash env {##1}}%
+\def\commondummies{%
-\def\command##1{\realbackslash command {##1}}%
+%
-\def\option##1{\realbackslash option {##1}}%
+% \definedummyword defines \#1 as \string\#1\space, thus effectively
-\def\dotless##1{\realbackslash dotless {##1}}%
+% preventing its expansion.  This is used only for control% words,
-\def\samp##1{\realbackslash samp {##1}}%
+% not control letters, because the \space would be incorrect for
-\def\,##1{\realbackslash ,{##1}}%
+% control characters, but is needed to separate the control word
-\def\t##1{\realbackslash t {##1}}%
+% from whatever follows.
-\def\r##1{\realbackslash r {##1}}%
+%
-\def\i##1{\realbackslash i {##1}}%
+% For control letters, we have \definedummyletter, which omits the
-\def\b##1{\realbackslash b {##1}}%
+% space.
-\def\sc##1{\realbackslash sc {##1}}%
+%
-\def\cite##1{\realbackslash cite {##1}}%
+% These can be used both for control words that take an argument and
-\def\key##1{\realbackslash key {##1}}%
+% those that do not.  If it is followed by {arg} in the input, then
-\def\file##1{\realbackslash file {##1}}%
+% that will dutifully get written to the index (or wherever).
-\def\var##1{\realbackslash var {##1}}%
+%
-\def\kbd##1{\realbackslash kbd {##1}}%
+\def\definedummyword  ##1{\def##1{\string##1\space}}%
-\def\dfn##1{\realbackslash dfn {##1}}%
+\def\definedummyletter##1{\def##1{\string##1}}%
-\def\emph##1{\realbackslash emph {##1}}%
+\let\definedummyaccent\definedummyletter
-\def\acronym##1{\realbackslash acronym {##1}}%
+%
-%
+\commondummiesnofonts
-% Handle some cases of @value -- where the variable name does not
+%
-% contain - or _, and the value does not contain any
+\definedummyletter\_%
-% (non-fully-expandable) commands.
+%
-\let\value = \expandablevalue
+% Non-English letters.
-%
+\definedummyword\AA
-\unsepspaces
+\definedummyword\AE
-% Turn off macro expansion
+\definedummyword\L
-\turnoffmacros
+\definedummyword\OE
-}
+\definedummyword\O
+\definedummyword\aa
-% If an index command is used in an @example environment, any spaces
+\definedummyword\ae
-% therein should become regular spaces in the raw index file, not the
+\definedummyword\l
-% expansion of \tie (\\leavevmode \penalty \@M \ ).
+\definedummyword\oe
-{\obeyspaces
+\definedummyword\o
-\gdef\unsepspaces{\obeyspaces\let =\space}}
+\definedummyword\ss
+\definedummyword\exclamdown
-% \indexnofonts no-ops all font-change commands.
+\definedummyword\questiondown
-% This is used when outputting the strings to sort the index by.
+\definedummyword\ordf
-\def\indexdummyfont#1{#1}
+\definedummyword\ordm
-\def\indexdummytex{TeX}
+%
-\def\indexdummydots{...}
+% Although these internal commands shouldn't show up, sometimes they do.
+\definedummyword\bf
+\definedummyword\gtr
+\definedummyword\hat
+\definedummyword\less
+\definedummyword\sf
+\definedummyword\sl
+\definedummyword\tclose
+\definedummyword\tt
+%
+\definedummyword\LaTeX
+\definedummyword\TeX
+%
+% Assorted special characters.
+\definedummyword\bullet
+\definedummyword\comma
+\definedummyword\copyright
+\definedummyword\registeredsymbol
+\definedummyword\dots
+\definedummyword\enddots
+\definedummyword\equiv
+\definedummyword\error
+\definedummyword\euro
+\definedummyword\guillemetleft
+\definedummyword\guillemetright
+\definedummyword\guilsinglleft
+\definedummyword\guilsinglright
+\definedummyword\expansion
+\definedummyword\minus
+\definedummyword\pounds
+\definedummyword\point
+\definedummyword\print
+\definedummyword\quotedblbase
+\definedummyword\quotedblleft
+\definedummyword\quotedblright
+\definedummyword\quoteleft
+\definedummyword\quoteright
+\definedummyword\quotesinglbase
+\definedummyword\result
+\definedummyword\textdegree
+%
+% We want to disable all macros so that they are not expanded by \write.
+\macrolist
+%
+\normalturnoffactive
+%
+% Handle some cases of @value -- where it does not contain any
+% (non-fully-expandable) commands.
+\makevalueexpandable
+}
+% \commondummiesnofonts: common to \commondummies and \indexnofonts.
+%
+\def\commondummiesnofonts{%
+% Control letters and accents.
+\definedummyletter\!%
+\definedummyaccent\"%
+\definedummyaccent\'%
+\definedummyletter\*%
+\definedummyaccent\,%
+\definedummyletter\.%
+\definedummyletter\/%
+\definedummyletter\:%
+\definedummyaccent\=%
+\definedummyletter\?%
+\definedummyaccent\^%
+\definedummyaccent\`%
+\definedummyaccent\~%
+\definedummyword\u
+\definedummyword\v
+\definedummyword\H
+\definedummyword\dotaccent
+\definedummyword\ringaccent
+\definedummyword\tieaccent
+\definedummyword\ubaraccent
+\definedummyword\udotaccent
+\definedummyword\dotless
+%
+% Texinfo font commands.
+\definedummyword\b
+\definedummyword\i
+\definedummyword\r
+\definedummyword\sc
+\definedummyword\t
+%
+% Commands that take arguments.
+\definedummyword\acronym
+\definedummyword\cite
+\definedummyword\code
+\definedummyword\command
+\definedummyword\dfn
+\definedummyword\emph
+\definedummyword\env
+\definedummyword\file
+\definedummyword\kbd
+\definedummyword\key
+\definedummyword\math
+\definedummyword\option
+\definedummyword\pxref
+\definedummyword\ref
+\definedummyword\samp
+\definedummyword\strong
+\definedummyword\tie
+\definedummyword\uref
+\definedummyword\url
+\definedummyword\var
+\definedummyword\verb
+\definedummyword\w
+\definedummyword\xref
+}
+% \indexnofonts is used when outputting the strings to sort the index
+% by, and when constructing control sequence names.  It eliminates all
+% control sequences and just writes whatever the best ASCII sort string
+% would be for a given command (usually its argument).
+%
 \def\indexnofonts{%
-% Just ignore accents.
+% Accent commands should become @asis.
-\let\,=\indexdummyfont
+\def\definedummyaccent##1{\let##1\asis}%
-\let\"=\indexdummyfont
+% We can just ignore other control letters.
-\let\`=\indexdummyfont
+\def\definedummyletter##1{\let##1\empty}%
-\let\'=\indexdummyfont
+% Hopefully, all control words can become @asis.
-\let\^=\indexdummyfont
+\let\definedummyword\definedummyaccent
-\let\~=\indexdummyfont
+%
-\let\==\indexdummyfont
+\commondummiesnofonts
-\let\b=\indexdummyfont
+%
-\let\c=\indexdummyfont
+% Don't no-op \tt, since it isn't a user-level command
-\let\d=\indexdummyfont
+% and is used in the definitions of the active chars like <, >, |, etc.
-\let\u=\indexdummyfont
+% Likewise with the other plain tex font commands.
-\let\v=\indexdummyfont
+%\let\tt=\asis
-\let\H=\indexdummyfont
+%
-\let\dotless=\indexdummyfont
+\def\ { }%
-% Take care of the plain tex special European modified letters.
+\def\@{@}%
-\def\oe{oe}%
+% how to handle braces?
-\def\ae{ae}%
+\def\_{\normalunderscore}%
-\def\aa{aa}%
+%
-\def\OE{OE}%
+% Non-English letters.
-\def\AE{AE}%
+\def\AA{AA}%
-\def\AA{AA}%
+\def\AE{AE}%
-\def\o{o}%
+\def\L{L}%
-\def\O{O}%
+\def\OE{OE}%
-\def\l{l}%
+\def\O{O}%
-\def\L{L}%
+\def\aa{aa}%
-\def\ss{ss}%
+\def\ae{ae}%
-\let\w=\indexdummyfont
+\def\l{l}%
-\let\t=\indexdummyfont
+\def\oe{oe}%
-\let\r=\indexdummyfont
+\def\o{o}%
-\let\i=\indexdummyfont
+\def\ss{ss}%
-\let\b=\indexdummyfont
+\def\exclamdown{!}%
-\let\emph=\indexdummyfont
+\def\questiondown{?}%
-\let\strong=\indexdummyfont
+\def\ordf{a}%
-\let\cite=\indexdummyfont
+\def\ordm{o}%
-\let\sc=\indexdummyfont
+%
-%Don't no-op \tt, since it isn't a user-level command
+\def\LaTeX{LaTeX}%
-% and is used in the definitions of the active chars like <, >, |...
+\def\TeX{TeX}%
-%\let\tt=\indexdummyfont
+%
-\let\tclose=\indexdummyfont
+% Assorted special characters.
-\let\code=\indexdummyfont
+% (The following {} will end up in the sort string, but that's ok.)
-\let\url=\indexdummyfont
+\def\bullet{bullet}%
-\let\uref=\indexdummyfont
+\def\comma{,}%
-\let\env=\indexdummyfont
+\def\copyright{copyright}%
-\let\acronym=\indexdummyfont
+\def\registeredsymbol{R}%
-\let\command=\indexdummyfont
+\def\dots{...}%
-\let\option=\indexdummyfont
+\def\enddots{...}%
-\let\file=\indexdummyfont
+\def\equiv{==}%
-\let\samp=\indexdummyfont
+\def\error{error}%
-\let\kbd=\indexdummyfont
+\def\euro{euro}%
-\let\key=\indexdummyfont
+\def\guillemetleft{<<}%
-\let\var=\indexdummyfont
+\def\guillemetright{>>}%
-\let\TeX=\indexdummytex
+\def\guilsinglleft{<}%
-\let\dots=\indexdummydots
+\def\guilsinglright{>}%
-\def\@{@}%
+\def\expansion{==>}%
-}
+\def\minus{-}%
+\def\pounds{pounds}%
-% To define \realbackslash, we must make \ not be an escape.
+\def\point{.}%
-% We must first make another character (@) an escape
+\def\print{-|}%
-% so we do not become unable to do a definition.
+\def\quotedblbase{"}%
+\def\quotedblleft{"}%
-{\catcode`\@=0 \catcode`\\=\other
+\def\quotedblright{"}%
-@gdef@realbackslash{\}}
+\def\quoteleft{`}%
+\def\quoteright{'}%
+\def\quotesinglbase{,}%
+\def\result{=>}%
+\def\textdegree{degrees}%
+%
+% We need to get rid of all macros, leaving only the arguments (if present).
+% Of course this is not nearly correct, but it is the best we can do for now.
+% makeinfo does not expand macros in the argument to @deffn, which ends up
+% writing an index entry, and texindex isn't prepared for an index sort entry
+% that starts with \.
+%
+% Since macro invocations are followed by braces, we can just redefine them
+% to take a single TeX argument.  The case of a macro invocation that
+% goes to end-of-line is not handled.
+%
+\macrolist
+}
 \let\indexbackslash=0  %overridden during \printindex.
 \let\SETmarginindex=\relax % put index entries in margin (undocumented)?
-% For \ifx comparisons.
-\def\emptymacro{\empty}
 % Most index entries go through here, but \dosubind is the general case.
-%
+% #1 is the index name, #2 is the entry text.
-\def\doind#1#2{\dosubind{#1}{#2}\empty}
+\def\doind#1#2{\dosubind{#1}{#2}{}}
 % Workhorse for all \fooindexes.
 % #1 is name of index, #2 is stuff to put there, #3 is subentry --
-% \empty if called from \doind, as we usually are.  The main exception
+% empty if called from \doind, as we usually are (the main exception
-% is with defuns, which call us directly.
+% is with most defuns, which call us directly).
 %
 \def\dosubind#1#2#3{%
+\iflinks
+{%
+% Store the main index entry text (including the third arg).
+\toks0 = {#2}%
+% If third arg is present, precede it with a space.
+\def\thirdarg{#3}%
+\ifx\thirdarg\empty \else
+\toks0 = \expandafter{\the\toks0 \space #3}%
+\fi
+%
+\edef\writeto{\csname#1indfile\endcsname}%
+%
+\safewhatsit\dosubindwrite
+}%
+\fi
+}
+% Write the entry in \toks0 to the index file:
+%
+\def\dosubindwrite{%
 % Put the index entry in the margin if desired.
 \ifx\SETmarginindex\relax\else
-\insert\margin{\hbox{\vrule height8pt depth3pt width0pt #2}}%
+\insert\margin{\hbox{\vrule height8pt depth3pt width0pt \the\toks0}}%
 \fi
-{%
+%
-\count255=\lastpenalty
+% Remember, we are within a group.
-{%
+\indexdummies % Must do this here, since \bf, etc expand at this stage
-\indexdummies % Must do this here, since \bf, etc expand at this stage
+\def\backslashcurfont{\indexbackslash}% \indexbackslash isn't defined now
-\escapechar=`\\
+% so it will be output as is; and it will print as backslash.
-{%
+%
-\let\folio = 0% We will expand all macros now EXCEPT \folio.
+% Process the index entry with all font commands turned off, to
-\def\rawbackslashxx{\indexbackslash}% \indexbackslash isn't defined now
+% get the string to sort by.
-% so it will be output as is; and it will print as backslash.
+{\indexnofonts
-%
+\edef\temp{\the\toks0}% need full expansion
-\def\thirdarg{#3}%
+\xdef\indexsorttmp{\temp}%
-%
-% If third arg is present, precede it with space in sort key.
-\ifx\thirdarg\emptymacro
-\let\subentry = \empty
-\else
-\def\subentry{ #3}%
-\fi
-%
-% First process the index entry with all font commands turned
-% off to get the string to sort by.
-{\indexnofonts \xdef\indexsorttmp{#2\subentry}}%
-%
-% Now the real index entry with the fonts.
-\toks0 = {#2}%
-%
-% If third (subentry) arg is present, add it to the index
-% string.  And include a space.
-\ifx\thirdarg\emptymacro \else
-\toks0 = \expandafter{\the\toks0 \space #3}%
-\fi
-%
-% Set up the complete index entry, with both the sort key
-% and the original text, including any font commands.  We write
-% three arguments to \entry to the .?? file, texindex reduces to
-% two when writing the .??s sorted result.
-\edef\temp{%
-\write\csname#1indfile\endcsname{%
-\realbackslash entry{\indexsorttmp}{\folio}{\the\toks0}}%
-}%
-%
-% If a skip is the last thing on the list now, preserve it
-% by backing up by \lastskip, doing the \write, then inserting
-% the skip again.  Otherwise, the whatsit generated by the
-% \write will make \lastskip zero.  The result is that sequences
-% like this:
-% @end defun
-% @tindex whatever
-% @defun ...
-% will have extra space inserted, because the \medbreak in the
-% start of the @defun won't see the skip inserted by the @end of
-% the previous defun.
-%
-% But don't do any of this if we're not in vertical mode.  We
-% don't want to do a \vskip and prematurely end a paragraph.
-%
-% Avoid page breaks due to these extra skips, too.
-%
-\iflinks
-\ifvmode
-\skip0 = \lastskip
-\ifdim\lastskip = 0pt \else \nobreak\vskip-\lastskip \fi
-\fi
-%
-\temp % do the write
-%
-%
-\ifvmode \ifdim\skip0 = 0pt \else \nobreak\vskip\skip0 \fi \fi
-\fi
-}%
-}%
-\penalty\count255
 }%
+%
+% Set up the complete index entry, with both the sort key and
+% the original text, including any font commands.  We write
+% three arguments to \entry to the .?? file (four in the
+% subentry case), texindex reduces to two when writing the .??s
+% sorted result.
+\edef\temp{%
+\write\writeto{%
+\string\entry{\indexsorttmp}{\noexpand\folio}{\the\toks0}}%
+}%
+\temp
+}
+% Take care of unwanted page breaks/skips around a whatsit:
+%
+% If a skip is the last thing on the list now, preserve it
+% by backing up by \lastskip, doing the \write, then inserting
+% the skip again.  Otherwise, the whatsit generated by the
+% \write or \pdfdest will make \lastskip zero.  The result is that
+% sequences like this:
+% @end defun
+% @tindex whatever
+% @defun ...
+% will have extra space inserted, because the \medbreak in the
+% start of the @defun won't see the skip inserted by the @end of
+% the previous defun.
+%
+% But don't do any of this if we're not in vertical mode.  We
+% don't want to do a \vskip and prematurely end a paragraph.
+%
+% Avoid page breaks due to these extra skips, too.
+%
+% But wait, there is a catch there:
+% We'll have to check whether \lastskip is zero skip.  \ifdim is not
+% sufficient for this purpose, as it ignores stretch and shrink parts
+% of the skip.  The only way seems to be to check the textual
+% representation of the skip.
+%
+% The following is almost like \def\zeroskipmacro{0.0pt} except that
+% the ``p'' and ``t'' characters have catcode \other, not 11 (letter).
+%
+\edef\zeroskipmacro{\expandafter\the\csname z@skip\endcsname}
+%
+\newskip\whatsitskip
+\newcount\whatsitpenalty
+%
+% ..., ready, GO:
+%
+\def\safewhatsit#1{%
+\ifhmode
+#1%
+\else
+% \lastskip and \lastpenalty cannot both be nonzero simultaneously.
+\whatsitskip = \lastskip
+\edef\lastskipmacro{\the\lastskip}%
+\whatsitpenalty = \lastpenalty
+%
+% If \lastskip is nonzero, that means the last item was a
+% skip.  And since a skip is discardable, that means this
+% -\whatsitskip glue we're inserting is preceded by a
+% non-discardable item, therefore it is not a potential
+% breakpoint, therefore no \nobreak needed.
+\ifx\lastskipmacro\zeroskipmacro
+\else
+\vskip-\whatsitskip
+\fi
+%
+#1%
+%
+\ifx\lastskipmacro\zeroskipmacro
+% If \lastskip was zero, perhaps the last item was a penalty, and
+% perhaps it was >=10000, e.g., a \nobreak.  In that case, we want
+% to re-insert the same penalty (values >10000 are used for various
+% signals); since we just inserted a non-discardable item, any
+% following glue (such as a \parskip) would be a breakpoint.  For example:
+%
+%   @deffn deffn-whatever
+%   @vindex index-whatever
+%   Description.
+% would allow a break between the index-whatever whatsit
+% and the "Description." paragraph.
+\ifnum\whatsitpenalty>9999 \penalty\whatsitpenalty \fi
+\else
+% On the other hand, if we had a nonzero \lastskip,
+% this make-up glue would be preceded by a non-discardable item
+% (the whatsit from the \write), so we must insert a \nobreak.
+\nobreak\vskip\whatsitskip
+\fi
+\fi
 }
 % The index entry written in the file actually looks like
 %  \entry {sortstring}{page}{topic}
 % or
 % Define the macros used in formatting output of the sorted index material.
 % @printindex causes a particular index (the ??s file) to get printed.
 % It does not print any chapter heading (usually an @unnumbered).
 %
-\def\printindex{\parsearg\doprintindex}
+\parseargdef\printindex{\begingroup
-\def\doprintindex#1{\begingroup
 \dobreak \chapheadingskip{10000}%
 %
 \smallfonts \rm
 \tolerance = 9500
-\indexbreaks
+\plainfrenchspacing
+\everypar = {}% don't want the \kern\-parindent from indentation suppression.
 %
 % See if the index file exists and is nonempty.
 % Change catcode of @ here so that if the index file contains
 % \initial {@}
 % as its first line, TeX doesn't complain about mismatched braces
 \putwordIndexIsEmpty
 \else
 % Index files are almost Texinfo source, but we use \ as the escape
 % character.  It would be better to use @, but that's too big a change
 % to make right now.
-\def\indexbackslash{\rawbackslashxx}%
+\def\indexbackslash{\backslashcurfont}%
 \catcode`\\ = 0
 \escapechar = `\\
 \begindoublecolumns
 \input \jobname.#1s
 \enddoublecolumns
 %
 % Remove any glue we may have, we'll be inserting our own.
 \removelastskip
 %
 % We like breaks before the index initials, so insert a bonus.
-\penalty -300
+\nobreak
+\vskip 0pt plus 3\baselineskip
+\penalty 0
+\vskip 0pt plus -3\baselineskip
 %
 % Typeset the initial.  Making this add up to a whole number of
 % baselineskips increases the chance of the dots lining up from column
 % to column.  It still won't often be perfect, because of the stretch
 % we need before each entry, but it's better.
 %
 % No shrink because it confuses \balancecolumns.
 \vskip 1.67\baselineskip plus .5\baselineskip
 \leftline{\secbf #1}%
-\vskip .33\baselineskip plus .1\baselineskip
-%
 % Do our best not to break after the initial.
 \nobreak
+\vskip .33\baselineskip plus .1\baselineskip
 }}
-% This typesets a paragraph consisting of #1, dot leaders, and then #2
+% \entry typesets a paragraph consisting of the text (#1), dot leaders, and
-% flush to the right margin.  It is used for index and table of contents
+% then page number (#2) flushed to the right margin.  It is used for index
-% entries.  The paragraph is indented by \leftskip.
+% and table of contents entries.  The paragraph is indented by \leftskip.
 %
-\def\entry#1#2{\begingroup
+% A straightforward implementation would start like this:
-%
+%	\def\entry#1#2{...
-% Start a new paragraph if necessary, so our assignments below can't
+% But this freezes the catcodes in the argument, and can cause problems to
-% affect previous text.
+% @code, which sets - active.  This problem was fixed by a kludge---
+% ``-'' was active throughout whole index, but this isn't really right.
+%
+% The right solution is to prevent \entry from swallowing the whole text.
+%                                 --kasal, 21nov03
+\def\entry{%
+\begingroup
+%
+% Start a new paragraph if necessary, so our assignments below can't
+% affect previous text.
+\par
+%
+% Do not fill out the last line with white space.
+\parfillskip = 0in
+%
+% No extra space above this paragraph.
+\parskip = 0in
+%
+% Do not prefer a separate line ending with a hyphen to fewer lines.
+\finalhyphendemerits = 0
+%
+% \hangindent is only relevant when the entry text and page number
+% don't both fit on one line.  In that case, bob suggests starting the
+% dots pretty far over on the line.  Unfortunately, a large
+% indentation looks wrong when the entry text itself is broken across
+% lines.  So we use a small indentation and put up with long leaders.
+%
+% \hangafter is reset to 1 (which is the value we want) at the start
+% of each paragraph, so we need not do anything with that.
+\hangindent = 2em
+%
+% When the entry text needs to be broken, just fill out the first line
+% with blank space.
+\rightskip = 0pt plus1fil
+%
+% A bit of stretch before each entry for the benefit of balancing
+% columns.
+\vskip 0pt plus1pt
+%
+% Swallow the left brace of the text (first parameter):
+\afterassignment\doentry
+\let\temp =
+}
+\def\doentry{%
+\bgroup % Instead of the swallowed brace.
+\noindent
+\aftergroup\finishentry
+% And now comes the text of the entry.
+}
+\def\finishentry#1{%
+% #1 is the page number.
+%
+% The following is kludged to not output a line of dots in the index if
+% there are no page numbers.  The next person who breaks this will be
+% cursed by a Unix daemon.
+\setbox\boxA = \hbox{#1}%
+\ifdim\wd\boxA = 0pt
+\ %
+\else
+%
+% If we must, put the page number on a line of its own, and fill out
+% this line with blank space.  (The \hfil is overwhelmed with the
+% fill leaders glue in \indexdotfill if the page number does fit.)
+\hfil\penalty50
+\null\nobreak\indexdotfill % Have leaders before the page number.
+%
+% The `\ ' here is removed by the implicit \unskip that TeX does as
+% part of (the primitive) \par.  Without it, a spurious underfull
+% \hbox ensues.
+\ifpdf
+	\pdfgettoks#1.%
+	\ \the\toksA
+\else
+	\ #1%
+\fi
+\fi
+\par
+\endgroup
+}
+% Like plain.tex's \dotfill, except uses up at least 1 em.
+\def\indexdotfill{\cleaders
+\hbox{$\mathsurround=0pt \mkern1.5mu.\mkern1.5mu$}\hskip 1em plus 1fill}
+\def\primary #1{\line{#1\hfil}}
+\newskip\secondaryindent \secondaryindent=0.5cm
+\def\secondary#1#2{{%
+\parfillskip=0in
+\parskip=0in
+\hangindent=1in
+\hangafter=1
+\noindent\hskip\secondaryindent\hbox{#1}\indexdotfill
+\ifpdf
+\pdfgettoks#2.\ \the\toksA % The page number ends the paragraph.
+\else
+#2
+\fi
 \par
-%
-% Do not fill out the last line with white space.
-\parfillskip = 0in
-%
-% No extra space above this paragraph.
-\parskip = 0in
-%
-% Do not prefer a separate line ending with a hyphen to fewer lines.
-\finalhyphendemerits = 0
-%
-% \hangindent is only relevant when the entry text and page number
-% don't both fit on one line.  In that case, bob suggests starting the
-% dots pretty far over on the line.  Unfortunately, a large
-% indentation looks wrong when the entry text itself is broken across
-% lines.  So we use a small indentation and put up with long leaders.
-%
-% \hangafter is reset to 1 (which is the value we want) at the start
-% of each paragraph, so we need not do anything with that.
-\hangindent = 2em
-%
-% When the entry text needs to be broken, just fill out the first line
-% with blank space.
-\rightskip = 0pt plus1fil
-%
-% A bit of stretch before each entry for the benefit of balancing columns.
-\vskip 0pt plus1pt
-%
-% Start a ``paragraph'' for the index entry so the line breaking
-% parameters we've set above will have an effect.
-\noindent
-%
-% Insert the text of the index entry.  TeX will do line-breaking on it.
-#1%
-% The following is kludged to not output a line of dots in the index if
-% there are no page numbers.  The next person who breaks this will be
-% cursed by a Unix daemon.
-\def\tempa{{\rm }}%
-\def\tempb{#2}%
-\edef\tempc{\tempa}%
-\edef\tempd{\tempb}%
-\ifx\tempc\tempd\ \else%
-%
-% If we must, put the page number on a line of its own, and fill out
-% this line with blank space.  (The \hfil is overwhelmed with the
-% fill leaders glue in \indexdotfill if the page number does fit.)
-\hfil\penalty50
-\null\nobreak\indexdotfill % Have leaders before the page number.
-%
-% The `\ ' here is removed by the implicit \unskip that TeX does as
-% part of (the primitive) \par.  Without it, a spurious underfull
-% \hbox ensues.
-\ifpdf
-\pdfgettoks#2.\ \the\toksA % The page number ends the paragraph.
-\else
-\ #2% The page number ends the paragraph.
-\fi
-\fi%
-\par
-\endgroup}
-% Like \dotfill except takes at least 1 em.
-\def\indexdotfill{\cleaders
-\hbox{$\mathsurround=0pt \mkern1.5mu ${\it .}$ \mkern1.5mu$}\hskip 1em plus 1fill}
-\def\primary #1{\line{#1\hfil}}
-\newskip\secondaryindent \secondaryindent=0.5cm
-\def\secondary #1#2{
-{\parfillskip=0in \parskip=0in
-\hangindent =1in \hangafter=1
-\noindent\hskip\secondaryindent\hbox{#1}\indexdotfill #2\par
 }}
 % Define two-column mode, which we use to typeset indexes.
 % Adapted from the TeXbook, page 416, which is to say,
 % the manmac.tex format used to print the TeXbook itself.
 \divide\doublecolumnhsize by 2
 \hsize = \doublecolumnhsize
 %
 % Double the \vsize as well.  (We don't need a separate register here,
 % since nobody clobbers \vsize.)
-\advance\vsize by -\ht\partialpage
 \vsize = 2\vsize
 }
 % The double-column output routine for all double-column pages except
 % the last.
 % Get the available space for the double columns -- the normal
 % (undoubled) page height minus any material left over from the
 % previous page.
 \dimen@ = \vsize
 \divide\dimen@ by 2
+\advance\dimen@ by -\ht\partialpage
 %
 % box0 will be the left-hand column, box2 the right.
 \setbox0=\vsplit255 to\dimen@ \setbox2=\vsplit255 to\dimen@
 \onepageout\pagesofar
 \unvbox255
 \penalty\outputpenalty
 }
+%
+% Re-output the contents of the output page -- any previous material,
+% followed by the two boxes we just split, in box0 and box2.
 \def\pagesofar{%
-% Re-output the contents of the output page -- any previous material,
-% followed by the two boxes we just split, in box0 and box2.
 \unvbox\partialpage
 %
 \hsize = \doublecolumnhsize
 \wd0=\hsize \wd2=\hsize
 \hbox to\pagewidth{\box0\hfil\box2}%
 }
+%
+% All done with double columns.
 \def\enddoublecolumns{%
+% The following penalty ensures that the page builder is exercised
+% _before_ we change the output routine.  This is necessary in the
+% following situation:
+%
+% The last section of the index consists only of a single entry.
+% Before this section, \pagetotal is less than \pagegoal, so no
+% break occurs before the last section starts.  However, the last
+% section, consisting of \initial and the single \entry, does not
+% fit on the page and has to be broken off.  Without the following
+% penalty the page builder will not be exercised until \eject
+% below, and by that time we'll already have changed the output
+% routine to the \balancecolumns version, so the next-to-last
+% double-column page will be processed with \balancecolumns, which
+% is wrong:  The two columns will go to the main vertical list, with
+% the broken-off section in the recent contributions.  As soon as
+% the output routine finishes, TeX starts reconsidering the page
+% break.  The two columns and the broken-off section both fit on the
+% page, because the two columns now take up only half of the page
+% goal.  When TeX sees \eject from below which follows the final
+% section, it invokes the new output routine that we've set after
+% \balancecolumns below; \onepageout will try to fit the two columns
+% and the final section into the vbox of \pageheight (see
+% \pagebody), causing an overfull box.
+%
+% Note that glue won't work here, because glue does not exercise the
+% page builder, unlike penalties (see The TeXbook, pp. 280-281).
+\penalty0
+%
 \output = {%
 % Split the last of the double-column material.  Leave it on the
 % current page, no automatic page break.
 \balancecolumns
 %
 % the current page.  We're now back to normal single-column
 % typesetting, so reset \pagegoal to the normal \vsize (after the
 % \endgroup where \vsize got restored).
 \pagegoal = \vsize
 }
+%
+% Called at the end of the double column material.
 \def\balancecolumns{%
-% Called at the end of the double column material.
 \setbox0 = \vbox{\unvbox255}% like \box255 but more efficient, see p.120.
 \dimen@ = \ht0
 \advance\dimen@ by \topskip
 \advance\dimen@ by-\baselineskip
 \divide\dimen@ by 2 % target to split to
 \message{sectioning,}
 % Chapters, sections, etc.
+% \unnumberedno is an oxymoron, of course.  But we count the unnumbered
+% sections so that we can refer to them unambiguously in the pdf
+% outlines by their "section number".  We avoid collisions with chapter
+% numbers by starting them at 10000.  (If a document ever has 10000
+% chapters, we're in trouble anyway, I'm sure.)
+\newcount\unnumberedno \unnumberedno = 10000
 \newcount\chapno
 \newcount\secno        \secno=0
 \newcount\subsecno     \subsecno=0
 \newcount\subsubsecno  \subsubsecno=0
 % This counter is funny since it counts through charcodes of letters A, B, ...
 \newcount\appendixno  \appendixno = `\@
+%
 % \def\appendixletter{\char\the\appendixno}
-% We do the following for the sake of pdftex, which needs the actual
+% We do the following ugly conditional instead of the above simple
+% construct for the sake of pdftex, which needs the actual
 % letter in the expansion, not just typeset.
+%
 \def\appendixletter{%
 \ifnum\appendixno=`A A%
 \else\ifnum\appendixno=`B B%
 \else\ifnum\appendixno=`C C%
 \else\ifnum\appendixno=`D D%
 % with the same letter (or @) in the toc without it.
 \else\char\the\appendixno
 \fi\fi\fi\fi\fi\fi\fi\fi\fi\fi\fi\fi\fi
 \fi\fi\fi\fi\fi\fi\fi\fi\fi\fi\fi\fi\fi}
-% Each @chapter defines this as the name of the chapter.
+% Each @chapter defines these (using marks) as the number+name, number
-% page headings and footings can use it.  @section does likewise.
+% and name of the chapter.  Page headings and footings can use
+% these.  @section does likewise.
 \def\thischapter{}
+\def\thischapternum{}
+\def\thischaptername{}
 \def\thissection{}
+\def\thissectionnum{}
+\def\thissectionname{}
 \newcount\absseclevel % used to calculate proper heading level
-\newcount\secbase\secbase=0 % @raise/lowersections modify this count
+\newcount\secbase\secbase=0 % @raisesections/@lowersections modify this count
 % @raisesections: treat @section as chapter, @subsection as section, etc.
 \def\raisesections{\global\advance\secbase by -1}
 \let\up=\raisesections % original BFox name
 % @lowersections: treat @chapter as section, @section as subsection, etc.
 \def\lowersections{\global\advance\secbase by 1}
 \let\down=\lowersections % original BFox name
-% Choose a numbered-heading macro
+% we only have subsub.
-% #1 is heading level if unmodified by @raisesections or @lowersections
+\chardef\maxseclevel = 3
-% #2 is text for heading
+%
-\def\numhead#1#2{\absseclevel=\secbase\advance\absseclevel by #1
+% A numbered section within an unnumbered changes to unnumbered too.
-\ifcase\absseclevel
+% To achive this, remember the "biggest" unnum. sec. we are currently in:
-\chapterzzz{#2}
+\chardef\unmlevel = \maxseclevel
-\or
+%
-\seczzz{#2}
+% Trace whether the current chapter is an appendix or not:
-\or
+% \chapheadtype is "N" or "A", unnumbered chapters are ignored.
-\numberedsubseczzz{#2}
+\def\chapheadtype{N}
-\or
-\numberedsubsubseczzz{#2}
+% Choose a heading macro
-\else
+% #1 is heading type
-\ifnum \absseclevel<0
+% #2 is heading level
-\chapterzzz{#2}
+% #3 is text for heading
+\def\genhead#1#2#3{%
+% Compute the abs. sec. level:
+\absseclevel=#2
+\advance\absseclevel by \secbase
+% Make sure \absseclevel doesn't fall outside the range:
+\ifnum \absseclevel < 0
+\absseclevel = 0
 \else
-\numberedsubsubseczzz{#2}
+\ifnum \absseclevel > 3
+\absseclevel = 3
+\fi
 \fi
-\fi
+% The heading type:
-}
+\def\headtype{#1}%
+\if \headtype U%
-% like \numhead, but chooses appendix heading levels
+\ifnum \absseclevel < \unmlevel
-\def\apphead#1#2{\absseclevel=\secbase\advance\absseclevel by #1
+\chardef\unmlevel = \absseclevel
-\ifcase\absseclevel
+\fi
-\appendixzzz{#2}
-\or
-\appendixsectionzzz{#2}
-\or
-\appendixsubseczzz{#2}
-\or
-\appendixsubsubseczzz{#2}
-\else
-\ifnum \absseclevel<0
-\appendixzzz{#2}
 \else
-\appendixsubsubseczzz{#2}
+% Check for appendix sections:
+\ifnum \absseclevel = 0
+\edef\chapheadtype{\headtype}%
+\else
+\if \headtype A\if \chapheadtype N%
+	\errmessage{@appendix... within a non-appendix chapter}%
+\fi\fi
+\fi
+% Check for numbered within unnumbered:
+\ifnum \absseclevel > \unmlevel
+\def\headtype{U}%
+\else
+\chardef\unmlevel = 3
+\fi
 \fi
-\fi
+% Now print the heading:
-}
+\if \headtype U%
+\ifcase\absseclevel
-% like \numhead, but chooses numberless heading levels
+	\unnumberedzzz{#3}%
-\def\unnmhead#1#2{\absseclevel=\secbase\advance\absseclevel by #1
+\or \unnumberedseczzz{#3}%
-\ifcase\absseclevel
+\or \unnumberedsubseczzz{#3}%
-\unnumberedzzz{#2}
+\or \unnumberedsubsubseczzz{#3}%
-\or
+\fi
-\unnumberedseczzz{#2}
-\or
-\unnumberedsubseczzz{#2}
-\or
-\unnumberedsubsubseczzz{#2}
-\else
-\ifnum \absseclevel<0
-\unnumberedzzz{#2}
 \else
-\unnumberedsubsubseczzz{#2}
+\if \headtype A%
+\ifcase\absseclevel
+	  \appendixzzz{#3}%
+\or \appendixsectionzzz{#3}%
+\or \appendixsubseczzz{#3}%
+\or \appendixsubsubseczzz{#3}%
+\fi
+\else
+\ifcase\absseclevel
+	  \chapterzzz{#3}%
+\or \seczzz{#3}%
+\or \numberedsubseczzz{#3}%
+\or \numberedsubsubseczzz{#3}%
+\fi
+\fi
 \fi
-\fi
+\suppressfirstparagraphindent
 }
-% @chapter, @appendix, @unnumbered.
+% an interface:
-\def\thischaptername{No Chapter Title}
+\def\numhead{\genhead N}
-\outer\def\chapter{\parsearg\chapteryyy}
+\def\apphead{\genhead A}
-\def\chapteryyy #1{\numhead0{#1}} % normally numhead0 calls chapterzzz
+\def\unnmhead{\genhead U}
-\def\chapterzzz #1{%
-\secno=0 \subsecno=0 \subsubsecno=0
+% @chapter, @appendix, @unnumbered.  Increment top-level counter, reset
-\global\advance \chapno by 1 \message{\putwordChapter\space \the\chapno}%
+% all lower-level sectioning counters to zero.
-\chapmacro {#1}{\the\chapno}%
+%
-\gdef\thissection{#1}%
+% Also set \chaplevelprefix, which we prepend to @float sequence numbers
-\gdef\thischaptername{#1}%
+% (e.g., figures), q.v.  By default (before any chapter), that is empty.
-% We don't substitute the actual chapter name into \thischapter
+\let\chaplevelprefix = \empty
-% because we don't want its macros evaluated now.
+%
-\xdef\thischapter{\putwordChapter{} \the\chapno: \noexpand\thischaptername}%
+\outer\parseargdef\chapter{\numhead0{#1}} % normally numhead0 calls chapterzzz
-\toks0 = {#1}%
+\def\chapterzzz#1{%
-\edef\temp{\noexpand\writetocentry{\realbackslash chapentry{\the\toks0}%
+% section resetting is \global in case the chapter is in a group, such
-{\the\chapno}}}%
+% as an @include file.
-\temp
+\global\secno=0 \global\subsecno=0 \global\subsubsecno=0
-\donoderef
+\global\advance\chapno by 1
-\global\let\section = \numberedsec
+%
-\global\let\subsection = \numberedsubsec
+% Used for \float.
-\global\let\subsubsection = \numberedsubsubsec
+\gdef\chaplevelprefix{\the\chapno.}%
-}
+\resetallfloatnos
+%
-\outer\def\appendix{\parsearg\appendixyyy}
+\message{\putwordChapter\space \the\chapno}%
-\def\appendixyyy #1{\apphead0{#1}} % normally apphead0 calls appendixzzz
+%
-\def\appendixzzz #1{%
+% Write the actual heading.
-\secno=0 \subsecno=0 \subsubsecno=0
+\chapmacro{#1}{Ynumbered}{\the\chapno}%
-\global\advance \appendixno by 1
+%
-\message{\putwordAppendix\space \appendixletter}%
+% So @section and the like are numbered underneath this chapter.
-\chapmacro {#1}{\putwordAppendix{} \appendixletter}%
+\global\let\section = \numberedsec
-\gdef\thissection{#1}%
+\global\let\subsection = \numberedsubsec
-\gdef\thischaptername{#1}%
+\global\let\subsubsection = \numberedsubsubsec
-\xdef\thischapter{\putwordAppendix{} \appendixletter: \noexpand\thischaptername}%
+}
-\toks0 = {#1}%
-\edef\temp{\noexpand\writetocentry{\realbackslash chapentry{\the\toks0}%
+\outer\parseargdef\appendix{\apphead0{#1}} % normally apphead0 calls appendixzzz
-{\putwordAppendix{} \appendixletter}}}%
+\def\appendixzzz#1{%
-\temp
+\global\secno=0 \global\subsecno=0 \global\subsubsecno=0
-\appendixnoderef
+\global\advance\appendixno by 1
-\global\let\section = \appendixsec
+\gdef\chaplevelprefix{\appendixletter.}%
-\global\let\subsection = \appendixsubsec
+\resetallfloatnos
-\global\let\subsubsection = \appendixsubsubsec
+%
+\def\appendixnum{\putwordAppendix\space \appendixletter}%
+\message{\appendixnum}%
+%
+\chapmacro{#1}{Yappendix}{\appendixletter}%
+%
+\global\let\section = \appendixsec
+\global\let\subsection = \appendixsubsec
+\global\let\subsubsection = \appendixsubsubsec
+}
+\outer\parseargdef\unnumbered{\unnmhead0{#1}} % normally unnmhead0 calls unnumberedzzz
+\def\unnumberedzzz#1{%
+\global\secno=0 \global\subsecno=0 \global\subsubsecno=0
+\global\advance\unnumberedno by 1
+%
+% Since an unnumbered has no number, no prefix for figures.
+\global\let\chaplevelprefix = \empty
+\resetallfloatnos
+%
+% This used to be simply \message{#1}, but TeX fully expands the
+% argument to \message.  Therefore, if #1 contained @-commands, TeX
+% expanded them.  For example, in `@unnumbered The @cite{Book}', TeX
+% expanded @cite (which turns out to cause errors because \cite is meant
+% to be executed, not expanded).
+%
+% Anyway, we don't want the fully-expanded definition of @cite to appear
+% as a result of the \message, we just want `@cite' itself.  We use
+% \the<toks register> to achieve this: TeX expands \the<toks> only once,
+% simply yielding the contents of <toks register>.  (We also do this for
+% the toc entries.)
+\toks0 = {#1}%
+\message{(\the\toks0)}%
+%
+\chapmacro{#1}{Ynothing}{\the\unnumberedno}%
+%
+\global\let\section = \unnumberedsec
+\global\let\subsection = \unnumberedsubsec
+\global\let\subsubsection = \unnumberedsubsubsec
 }
 % @centerchap is like @unnumbered, but the heading is centered.
-\outer\def\centerchap{\parsearg\centerchapyyy}
+\outer\parseargdef\centerchap{%
-\def\centerchapyyy #1{{\let\unnumbchapmacro=\centerchapmacro \unnumberedyyy{#1}}}
+% Well, we could do the following in a group, but that would break
+% an assumption that \chapmacro is called at the outermost level.
+% Thus we are safer this way:		--kasal, 24feb04
+\let\centerparametersmaybe = \centerparameters
+\unnmhead0{#1}%
+\let\centerparametersmaybe = \relax
+}
 % @top is like @unnumbered.
-\outer\def\top{\parsearg\unnumberedyyy}
+\let\top\unnumbered
-\outer\def\unnumbered{\parsearg\unnumberedyyy}
-\def\unnumberedyyy #1{\unnmhead0{#1}} % normally unnmhead0 calls unnumberedzzz
-\def\unnumberedzzz #1{%
-\secno=0 \subsecno=0 \subsubsecno=0
-%
-% This used to be simply \message{#1}, but TeX fully expands the
-% argument to \message.  Therefore, if #1 contained @-commands, TeX
-% expanded them.  For example, in `@unnumbered The @cite{Book}', TeX
-% expanded @cite (which turns out to cause errors because \cite is meant
-% to be executed, not expanded).
-%
-% Anyway, we don't want the fully-expanded definition of @cite to appear
-% as a result of the \message, we just want `@cite' itself.  We use
-% \the<toks register> to achieve this: TeX expands \the<toks> only once,
-% simply yielding the contents of <toks register>.  (We also do this for
-% the toc entries.)
-\toks0 = {#1}\message{(\the\toks0)}%
-%
-\unnumbchapmacro {#1}%
-\gdef\thischapter{#1}\gdef\thissection{#1}%
-\toks0 = {#1}%
-\edef\temp{\noexpand\writetocentry{\realbackslash unnumbchapentry{\the\toks0}}}%
-\temp
-\unnumbnoderef
-\global\let\section = \unnumberedsec
-\global\let\subsection = \unnumberedsubsec
-\global\let\subsubsection = \unnumberedsubsubsec
-}
 % Sections.
-\outer\def\numberedsec{\parsearg\secyyy}
+\outer\parseargdef\numberedsec{\numhead1{#1}} % normally calls seczzz
-\def\secyyy #1{\numhead1{#1}} % normally calls seczzz
+\def\seczzz#1{%
-\def\seczzz #1{%
+\global\subsecno=0 \global\subsubsecno=0  \global\advance\secno by 1
-\subsecno=0 \subsubsecno=0 \global\advance \secno by 1 %
+\sectionheading{#1}{sec}{Ynumbered}{\the\chapno.\the\secno}%
-\gdef\thissection{#1}\secheading {#1}{\the\chapno}{\the\secno}%
+}
-\toks0 = {#1}%
-\edef\temp{\noexpand\writetocentry{\realbackslash secentry{\the\toks0}%
+\outer\parseargdef\appendixsection{\apphead1{#1}} % normally calls appendixsectionzzz
-{\the\chapno}{\the\secno}}}%
+\def\appendixsectionzzz#1{%
-\temp
+\global\subsecno=0 \global\subsubsecno=0  \global\advance\secno by 1
-\donoderef
+\sectionheading{#1}{sec}{Yappendix}{\appendixletter.\the\secno}%
-\nobreak
+}
-}
+\let\appendixsec\appendixsection
-\outer\def\appendixsection{\parsearg\appendixsecyyy}
+\outer\parseargdef\unnumberedsec{\unnmhead1{#1}} % normally calls unnumberedseczzz
-\outer\def\appendixsec{\parsearg\appendixsecyyy}
+\def\unnumberedseczzz#1{%
-\def\appendixsecyyy #1{\apphead1{#1}} % normally calls appendixsectionzzz
+\global\subsecno=0 \global\subsubsecno=0  \global\advance\secno by 1
-\def\appendixsectionzzz #1{%
+\sectionheading{#1}{sec}{Ynothing}{\the\unnumberedno.\the\secno}%
-\subsecno=0 \subsubsecno=0 \global\advance \secno by 1 %
-\gdef\thissection{#1}\secheading {#1}{\appendixletter}{\the\secno}%
-\toks0 = {#1}%
-\edef\temp{\noexpand\writetocentry{\realbackslash secentry{\the\toks0}%
-{\appendixletter}{\the\secno}}}%
-\temp
-\appendixnoderef
-\nobreak
-}
-\outer\def\unnumberedsec{\parsearg\unnumberedsecyyy}
-\def\unnumberedsecyyy #1{\unnmhead1{#1}} % normally calls unnumberedseczzz
-\def\unnumberedseczzz #1{%
-\plainsecheading {#1}\gdef\thissection{#1}%
-\toks0 = {#1}%
-\edef\temp{\noexpand\writetocentry{\realbackslash unnumbsecentry{\the\toks0}}}%
-\temp
-\unnumbnoderef
-\nobreak
 }
 % Subsections.
-\outer\def\numberedsubsec{\parsearg\numberedsubsecyyy}
+\outer\parseargdef\numberedsubsec{\numhead2{#1}} % normally calls numberedsubseczzz
-\def\numberedsubsecyyy #1{\numhead2{#1}} % normally calls numberedsubseczzz
+\def\numberedsubseczzz#1{%
-\def\numberedsubseczzz #1{%
+\global\subsubsecno=0  \global\advance\subsecno by 1
-\gdef\thissection{#1}\subsubsecno=0 \global\advance \subsecno by 1 %
+\sectionheading{#1}{subsec}{Ynumbered}{\the\chapno.\the\secno.\the\subsecno}%
-\subsecheading {#1}{\the\chapno}{\the\secno}{\the\subsecno}%
+}
-\toks0 = {#1}%
-\edef\temp{\noexpand\writetocentry{\realbackslash subsecentry{\the\toks0}%
+\outer\parseargdef\appendixsubsec{\apphead2{#1}} % normally calls appendixsubseczzz
-{\the\chapno}{\the\secno}{\the\subsecno}}}%
+\def\appendixsubseczzz#1{%
-\temp
+\global\subsubsecno=0  \global\advance\subsecno by 1
-\donoderef
+\sectionheading{#1}{subsec}{Yappendix}%
-\nobreak
+{\appendixletter.\the\secno.\the\subsecno}%
 }
-\outer\def\appendixsubsec{\parsearg\appendixsubsecyyy}
+\outer\parseargdef\unnumberedsubsec{\unnmhead2{#1}} %normally calls unnumberedsubseczzz
-\def\appendixsubsecyyy #1{\apphead2{#1}} % normally calls appendixsubseczzz
+\def\unnumberedsubseczzz#1{%
-\def\appendixsubseczzz #1{%
+\global\subsubsecno=0  \global\advance\subsecno by 1
-\gdef\thissection{#1}\subsubsecno=0 \global\advance \subsecno by 1 %
+\sectionheading{#1}{subsec}{Ynothing}%
-\subsecheading {#1}{\appendixletter}{\the\secno}{\the\subsecno}%
+{\the\unnumberedno.\the\secno.\the\subsecno}%
-\toks0 = {#1}%
-\edef\temp{\noexpand\writetocentry{\realbackslash subsecentry{\the\toks0}%
-{\appendixletter}{\the\secno}{\the\subsecno}}}%
-\temp
-\appendixnoderef
-\nobreak
-}
-\outer\def\unnumberedsubsec{\parsearg\unnumberedsubsecyyy}
-\def\unnumberedsubsecyyy #1{\unnmhead2{#1}} %normally calls unnumberedsubseczzz
-\def\unnumberedsubseczzz #1{%
-\plainsubsecheading {#1}\gdef\thissection{#1}%
-\toks0 = {#1}%
-\edef\temp{\noexpand\writetocentry{\realbackslash unnumbsubsecentry%
-{\the\toks0}}}%
-\temp
-\unnumbnoderef
-\nobreak
 }
 % Subsubsections.
-\outer\def\numberedsubsubsec{\parsearg\numberedsubsubsecyyy}
+\outer\parseargdef\numberedsubsubsec{\numhead3{#1}} % normally numberedsubsubseczzz
-\def\numberedsubsubsecyyy #1{\numhead3{#1}} % normally numberedsubsubseczzz
+\def\numberedsubsubseczzz#1{%
-\def\numberedsubsubseczzz #1{%
+\global\advance\subsubsecno by 1
-\gdef\thissection{#1}\global\advance \subsubsecno by 1 %
+\sectionheading{#1}{subsubsec}{Ynumbered}%
-\subsubsecheading {#1}
+{\the\chapno.\the\secno.\the\subsecno.\the\subsubsecno}%
-{\the\chapno}{\the\secno}{\the\subsecno}{\the\subsubsecno}%
+}
-\toks0 = {#1}%
-\edef\temp{\noexpand\writetocentry{\realbackslash subsubsecentry{\the\toks0}%
+\outer\parseargdef\appendixsubsubsec{\apphead3{#1}} % normally appendixsubsubseczzz
-{\the\chapno}{\the\secno}{\the\subsecno}{\the\subsubsecno}}}%
+\def\appendixsubsubseczzz#1{%
-\temp
+\global\advance\subsubsecno by 1
-\donoderef
+\sectionheading{#1}{subsubsec}{Yappendix}%
-\nobreak
+{\appendixletter.\the\secno.\the\subsecno.\the\subsubsecno}%
 }
-\outer\def\appendixsubsubsec{\parsearg\appendixsubsubsecyyy}
+\outer\parseargdef\unnumberedsubsubsec{\unnmhead3{#1}} %normally unnumberedsubsubseczzz
-\def\appendixsubsubsecyyy #1{\apphead3{#1}} % normally appendixsubsubseczzz
+\def\unnumberedsubsubseczzz#1{%
-\def\appendixsubsubseczzz #1{%
+\global\advance\subsubsecno by 1
-\gdef\thissection{#1}\global\advance \subsubsecno by 1 %
+\sectionheading{#1}{subsubsec}{Ynothing}%
-\subsubsecheading {#1}
+{\the\unnumberedno.\the\secno.\the\subsecno.\the\subsubsecno}%
-{\appendixletter}{\the\secno}{\the\subsecno}{\the\subsubsecno}%
+}
-\toks0 = {#1}%
-\edef\temp{\noexpand\writetocentry{\realbackslash subsubsecentry{\the\toks0}%
-{\appendixletter}{\the\secno}{\the\subsecno}{\the\subsubsecno}}}%
-\temp
-\appendixnoderef
-\nobreak
-}
-\outer\def\unnumberedsubsubsec{\parsearg\unnumberedsubsubsecyyy}
-\def\unnumberedsubsubsecyyy #1{\unnmhead3{#1}} %normally unnumberedsubsubseczzz
-\def\unnumberedsubsubseczzz #1{%
-\plainsubsubsecheading {#1}\gdef\thissection{#1}%
-\toks0 = {#1}%
-\edef\temp{\noexpand\writetocentry{\realbackslash unnumbsubsubsecentry%
-{\the\toks0}}}%
-\temp
-\unnumbnoderef
-\nobreak
-}
-% These are variants which are not "outer", so they can appear in @ifinfo.
-% Actually, they should now be obsolete; ordinary section commands should work.
-\def\infotop{\parsearg\unnumberedzzz}
-\def\infounnumbered{\parsearg\unnumberedzzz}
-\def\infounnumberedsec{\parsearg\unnumberedseczzz}
-\def\infounnumberedsubsec{\parsearg\unnumberedsubseczzz}
-\def\infounnumberedsubsubsec{\parsearg\unnumberedsubsubseczzz}
-\def\infoappendix{\parsearg\appendixzzz}
-\def\infoappendixsec{\parsearg\appendixseczzz}
-\def\infoappendixsubsec{\parsearg\appendixsubseczzz}
-\def\infoappendixsubsubsec{\parsearg\appendixsubsubseczzz}
-\def\infochapter{\parsearg\chapterzzz}
-\def\infosection{\parsearg\sectionzzz}
-\def\infosubsection{\parsearg\subsectionzzz}
-\def\infosubsubsection{\parsearg\subsubsectionzzz}
 % These macros control what the section commands do, according
 % to what kind of chapter we are in (ordinary, appendix, or unnumbered).
 % Define them by default for a numbered chapter.
-\global\let\section = \numberedsec
+\let\section = \numberedsec
-\global\let\subsection = \numberedsubsec
+\let\subsection = \numberedsubsec
-\global\let\subsubsection = \numberedsubsubsec
+\let\subsubsection = \numberedsubsubsec
 % Define @majorheading, @heading and @subheading
 % NOTE on use of \vbox for chapter headings, section headings, and such:
 %       1) We use \vbox rather than the earlier \line to permit
 %          heading is obnoxious; this forbids it.
 %       3) Likewise, headings look best if no \parindent is used, and
 %          if justification is not attempted.  Hence \raggedright.
-\def\majorheading{\parsearg\majorheadingzzz}
+\def\majorheading{%
-\def\majorheadingzzz #1{%
+{\advance\chapheadingskip by 10pt \chapbreak }%
-{\advance\chapheadingskip by 10pt \chapbreak }%
+\parsearg\chapheadingzzz
-{\chapfonts \vbox{\hyphenpenalty=10000\tolerance=5000
+}
-\parindent=0pt\raggedright
-\rm #1\hfill}}\bigskip \par\penalty 200}
+\def\chapheading{\chapbreak \parsearg\chapheadingzzz}
+\def\chapheadingzzz#1{%
-\def\chapheading{\parsearg\chapheadingzzz}
+{\chapfonts \vbox{\hyphenpenalty=10000\tolerance=5000
-\def\chapheadingzzz #1{\chapbreak %
+\parindent=0pt\raggedright
-{\chapfonts \vbox{\hyphenpenalty=10000\tolerance=5000
+\rm #1\hfill}}%
-\parindent=0pt\raggedright
+\bigskip \par\penalty 200\relax
-\rm #1\hfill}}\bigskip \par\penalty 200}
+\suppressfirstparagraphindent
+}
 % @heading, @subheading, @subsubheading.
-\def\heading{\parsearg\plainsecheading}
+\parseargdef\heading{\sectionheading{#1}{sec}{Yomitfromtoc}{}
-\def\subheading{\parsearg\plainsubsecheading}
+\suppressfirstparagraphindent}
-\def\subsubheading{\parsearg\plainsubsubsecheading}
+\parseargdef\subheading{\sectionheading{#1}{subsec}{Yomitfromtoc}{}
+\suppressfirstparagraphindent}
+\parseargdef\subsubheading{\sectionheading{#1}{subsubsec}{Yomitfromtoc}{}
+\suppressfirstparagraphindent}
 % These macros generate a chapter, section, etc. heading only
 % (including whitespace, linebreaking, etc. around it),
 % given all the information in convenient, parsed form.
 %%% Args are the skip and penalty (usually negative)
 \def\dobreak#1#2{\par\ifdim\lastskip<#1\removelastskip\penalty#2\vskip#1\fi}
-\def\setchapterstyle #1 {\csname CHAPF#1\endcsname}
 %%% Define plain chapter starts, and page on/off switching for it
 % Parameter controlling skip before chapter headings (if needed)
 \newskip\chapheadingskip
 \def\chapbreak{\dobreak \chapheadingskip {-4000}}
 \def\chappager{\par\vfill\supereject}
-\def\chapoddpage{\chappager \ifodd\pageno \else \hbox to 0pt{} \chappager\fi}
+% Because \domark is called before \chapoddpage, the filler page will
+% get the headings for the next chapter, which is wrong.  But we don't
+% care -- we just disable all headings on the filler page.
+\def\chapoddpage{%
+\chappager
+\ifodd\pageno \else
+\begingroup
+\evenheadline={\hfil}\evenfootline={\hfil}%
+\oddheadline={\hfil}\oddfootline={\hfil}%
+\hbox to 0pt{}%
+\chappager
+\endgroup
+\fi
+}
 \def\setchapternewpage #1 {\csname CHAPPAG#1\endcsname}
 \def\CHAPPAGoff{%
 \global\let\contentsalignmacro = \chappager
 \global\let\contentsalignmacro = \chappager
 \global\let\pchapsepmacro=\chappager
 \global\let\pagealignmacro=\chappager
 \global\def\HEADINGSon{\HEADINGSsingle}}
-\def\CHAPPAGodd{
+\def\CHAPPAGodd{%
 \global\let\contentsalignmacro = \chapoddpage
 \global\let\pchapsepmacro=\chapoddpage
 \global\let\pagealignmacro=\chapoddpage
 \global\def\HEADINGSon{\HEADINGSdouble}}
 \CHAPPAGon
-\def\CHAPFplain{
+% Chapter opening.
-\global\let\chapmacro=\chfplain
+%
-\global\let\unnumbchapmacro=\unnchfplain
+% #1 is the text, #2 is the section type (Ynumbered, Ynothing,
-\global\let\centerchapmacro=\centerchfplain}
+% Yappendix, Yomitfromtoc), #3 the chapter number.
+%
-% Plain chapter opening.
+% To test against our argument.
-% #1 is the text, #2 the chapter number or empty if unnumbered.
+\def\Ynothingkeyword{Ynothing}
-\def\chfplain#1#2{%
+\def\Yomitfromtockeyword{Yomitfromtoc}
+\def\Yappendixkeyword{Yappendix}
+%
+\def\chapmacro#1#2#3{%
+% Insert the first mark before the heading break (see notes for \domark).
+\let\prevchapterdefs=\lastchapterdefs
+\let\prevsectiondefs=\lastsectiondefs
+\gdef\lastsectiondefs{\gdef\thissectionname{}\gdef\thissectionnum{}%
+\gdef\thissection{}}%
+%
+\def\temptype{#2}%
+\ifx\temptype\Ynothingkeyword
+\gdef\lastchapterdefs{\gdef\thischaptername{#1}\gdef\thischapternum{}%
+\gdef\thischapter{\thischaptername}}%
+\else\ifx\temptype\Yomitfromtockeyword
+\gdef\lastchapterdefs{\gdef\thischaptername{#1}\gdef\thischapternum{}%
+\gdef\thischapter{}}%
+\else\ifx\temptype\Yappendixkeyword
+\toks0={#1}%
+\xdef\lastchapterdefs{%
+\gdef\noexpand\thischaptername{\the\toks0}%
+\gdef\noexpand\thischapternum{\appendixletter}%
+\gdef\noexpand\thischapter{\putwordAppendix{} \noexpand\thischapternum:
+\noexpand\thischaptername}%
+}%
+\else
+\toks0={#1}%
+\xdef\lastchapterdefs{%
+\gdef\noexpand\thischaptername{\the\toks0}%
+\gdef\noexpand\thischapternum{\the\chapno}%
+\gdef\noexpand\thischapter{\putwordChapter{} \noexpand\thischapternum:
+\noexpand\thischaptername}%
+}%
+\fi\fi\fi
+%
+% Output the mark.  Pass it through \safewhatsit, to take care of
+% the preceding space.
+\safewhatsit\domark
+%
+% Insert the chapter heading break.
 \pchapsepmacro
+%
+% Now the second mark, after the heading break.  No break points
+% between here and the heading.
+\let\prevchapterdefs=\lastchapterdefs
+\let\prevsectiondefs=\lastsectiondefs
+\domark
+%
 {%
 \chapfonts \rm
-\def\chapnum{#2}%
+%
-\setbox0 = \hbox{#2\ifx\chapnum\empty\else\enspace\fi}%
+% Have to define \lastsection before calling \donoderef, because the
+% xref code eventually uses it.  On the other hand, it has to be called
+% after \pchapsepmacro, or the headline will change too soon.
+\gdef\lastsection{#1}%
+%
+% Only insert the separating space if we have a chapter/appendix
+% number, and don't print the unnumbered ``number''.
+\ifx\temptype\Ynothingkeyword
+\setbox0 = \hbox{}%
+\def\toctype{unnchap}%
+\else\ifx\temptype\Yomitfromtockeyword
+\setbox0 = \hbox{}% contents like unnumbered, but no toc entry
+\def\toctype{omit}%
+\else\ifx\temptype\Yappendixkeyword
+\setbox0 = \hbox{\putwordAppendix{} #3\enspace}%
+\def\toctype{app}%
+\else
+\setbox0 = \hbox{#3\enspace}%
+\def\toctype{numchap}%
+\fi\fi\fi
+%
+% Write the toc entry for this chapter.  Must come before the
+% \donoderef, because we include the current node name in the toc
+% entry, and \donoderef resets it to empty.
+\writetocentry{\toctype}{#1}{#3}%
+%
+% For pdftex, we have to write out the node definition (aka, make
+% the pdfdest) after any page break, but before the actual text has
+% been typeset.  If the destination for the pdf outline is after the
+% text, then jumping from the outline may wind up with the text not
+% being visible, for instance under high magnification.
+\donoderef{#2}%
+%
+% Typeset the actual heading.
+\nobreak % Avoid page breaks at the interline glue.
 \vbox{\hyphenpenalty=10000 \tolerance=5000 \parindent=0pt \raggedright
-\hangindent = \wd0 \centerparametersmaybe
+\hangindent=\wd0 \centerparametersmaybe
 \unhbox0 #1\par}%
 }%
 \nobreak\bigskip % no page break after a chapter title
 \nobreak
 }
-% Plain opening for unnumbered.
-\def\unnchfplain#1{\chfplain{#1}{}}
 % @centerchap -- centered and unnumbered.
 \let\centerparametersmaybe = \relax
-\def\centerchfplain#1{{%
+\def\centerparameters{%
-\def\centerparametersmaybe{%
+\advance\rightskip by 3\rightskip
-\advance\rightskip by 3\rightskip
+\leftskip = \rightskip
-\leftskip = \rightskip
+\parfillskip = 0pt
-\parfillskip = 0pt
+}
-}%
-\chfplain{#1}{}%
-}}
+% I don't think this chapter style is supported any more, so I'm not
+% updating it with the new noderef stuff.  We'll see.  --karl, 11aug03.
-\CHAPFplain % The default
+%
+\def\setchapterstyle #1 {\csname CHAPF#1\endcsname}
+%
 \def\unnchfopen #1{%
 \chapoddpage {\chapfonts \vbox{\hyphenpenalty=10000\tolerance=5000
 \parindent=0pt\raggedright
 \rm #1\hfill}}\bigskip \par\nobreak
 }
 \def\chfopen #1#2{\chapoddpage {\chapfonts
 \vbox to 3in{\vfil \hbox to\hsize{\hfil #2} \hbox to\hsize{\hfil #1} \vfil}}%
 \par\penalty 5000 %
 }
 \def\centerchfopen #1{%
 \chapoddpage {\chapfonts \vbox{\hyphenpenalty=10000\tolerance=5000
 \parindent=0pt
 \hfill {\rm #1}\hfill}}\bigskip \par\nobreak
 }
+\def\CHAPFopen{%
-\def\CHAPFopen{
+\global\let\chapmacro=\chfopen
-\global\let\chapmacro=\chfopen
+\global\let\centerchapmacro=\centerchfopen}
-\global\let\unnumbchapmacro=\unnchfopen
-\global\let\centerchapmacro=\centerchfopen}
+% Section titles.  These macros combine the section number parts and
+% call the generic \sectionheading to do the printing.
-% Section titles.
+%
 \newskip\secheadingskip
-\def\secheadingbreak{\dobreak \secheadingskip {-1000}}
+\def\secheadingbreak{\dobreak \secheadingskip{-1000}}
-\def\secheading#1#2#3{\sectionheading{sec}{#2.#3}{#1}}
-\def\plainsecheading#1{\sectionheading{sec}{}{#1}}
 % Subsection titles.
-\newskip \subsecheadingskip
+\newskip\subsecheadingskip
-\def\subsecheadingbreak{\dobreak \subsecheadingskip {-500}}
+\def\subsecheadingbreak{\dobreak \subsecheadingskip{-500}}
-\def\subsecheading#1#2#3#4{\sectionheading{subsec}{#2.#3.#4}{#1}}
-\def\plainsubsecheading#1{\sectionheading{subsec}{}{#1}}
 % Subsubsection titles.
-\let\subsubsecheadingskip = \subsecheadingskip
+\def\subsubsecheadingskip{\subsecheadingskip}
-\let\subsubsecheadingbreak = \subsecheadingbreak
+\def\subsubsecheadingbreak{\subsecheadingbreak}
-\def\subsubsecheading#1#2#3#4#5{\sectionheading{subsubsec}{#2.#3.#4.#5}{#1}}
-\def\plainsubsubsecheading#1{\sectionheading{subsubsec}{}{#1}}
+% Print any size, any type, section title.
+%
-% Print any size section title.
+% #1 is the text, #2 is the section level (sec/subsec/subsubsec), #3 is
-%
+% the section type for xrefs (Ynumbered, Ynothing, Yappendix), #4 is the
-% #1 is the section type (sec/subsec/subsubsec), #2 is the section
+% section number.
-% number (maybe empty), #3 the text.
+%
-\def\sectionheading#1#2#3{%
+\def\seckeyword{sec}
-{%
+%
-\expandafter\advance\csname #1headingskip\endcsname by \parskip
+\def\sectionheading#1#2#3#4{%
-\csname #1headingbreak\endcsname
-}%
 {%
 % Switch to the right set of fonts.
-\csname #1fonts\endcsname \rm
+\csname #2fonts\endcsname \rm
 %
-% Only insert the separating space if we have a section number.
+\def\sectionlevel{#2}%
-\def\secnum{#2}%
+\def\temptype{#3}%
-\setbox0 = \hbox{#2\ifx\secnum\empty\else\enspace\fi}%
 %
+% Insert first mark before the heading break (see notes for \domark).
+\let\prevsectiondefs=\lastsectiondefs
+\ifx\temptype\Ynothingkeyword
+\ifx\sectionlevel\seckeyword
+\gdef\lastsectiondefs{\gdef\thissectionname{#1}\gdef\thissectionnum{}%
+\gdef\thissection{\thissectionname}}%
+\fi
+\else\ifx\temptype\Yomitfromtockeyword
+% Don't redefine \thissection.
+\else\ifx\temptype\Yappendixkeyword
+\ifx\sectionlevel\seckeyword
+\toks0={#1}%
+\xdef\lastsectiondefs{%
+\gdef\noexpand\thissectionname{\the\toks0}%
+\gdef\noexpand\thissectionnum{#4}%
+\gdef\noexpand\thissection{\putwordSection{} \noexpand\thissectionnum:
+\noexpand\thissectionname}%
+}%
+\fi
+\else
+\ifx\sectionlevel\seckeyword
+\toks0={#1}%
+\xdef\lastsectiondefs{%
+\gdef\noexpand\thissectionname{\the\toks0}%
+\gdef\noexpand\thissectionnum{#4}%
+\gdef\noexpand\thissection{\putwordSection{} \noexpand\thissectionnum:
+\noexpand\thissectionname}%
+}%
+\fi
+\fi\fi\fi
+%
+% Output the mark.  Pass it through \safewhatsit, to take care of
+% the preceding space.
+\safewhatsit\domark
+%
+% Insert space above the heading.
+\csname #2headingbreak\endcsname
+%
+% Now the second mark, after the heading break.  No break points
+% between here and the heading.
+\let\prevsectiondefs=\lastsectiondefs
+\domark
+%
+% Only insert the space after the number if we have a section number.
+\ifx\temptype\Ynothingkeyword
+\setbox0 = \hbox{}%
+\def\toctype{unn}%
+\gdef\lastsection{#1}%
+\else\ifx\temptype\Yomitfromtockeyword
+% for @headings -- no section number, don't include in toc,
+% and don't redefine \lastsection.
+\setbox0 = \hbox{}%
+\def\toctype{omit}%
+\let\sectionlevel=\empty
+\else\ifx\temptype\Yappendixkeyword
+\setbox0 = \hbox{#4\enspace}%
+\def\toctype{app}%
+\gdef\lastsection{#1}%
+\else
+\setbox0 = \hbox{#4\enspace}%
+\def\toctype{num}%
+\gdef\lastsection{#1}%
+\fi\fi\fi
+%
+% Write the toc entry (before \donoderef).  See comments in \chapmacro.
+\writetocentry{\toctype\sectionlevel}{#1}{#4}%
+%
+% Write the node reference (= pdf destination for pdftex).
+% Again, see comments in \chapmacro.
+\donoderef{#3}%
+%
+% Interline glue will be inserted when the vbox is completed.
+% That glue will be a valid breakpoint for the page, since it'll be
+% preceded by a whatsit (usually from the \donoderef, or from the
+% \writetocentry if there was no node).  We don't want to allow that
+% break, since then the whatsits could end up on page n while the
+% section is on page n+1, thus toc/etc. are wrong.  Debian bug 276000.
+\nobreak
+%
+% Output the actual section heading.
 \vbox{\hyphenpenalty=10000 \tolerance=5000 \parindent=0pt \raggedright
-\hangindent = \wd0 % zero if no section number
+\hangindent=\wd0  % zero if no section number
-\unhbox0 #3}%
+\unhbox0 #1}%
 }%
-\ifdim\parskip<10pt \nobreak\kern10pt\nobreak\kern-\parskip\fi \nobreak
+% Add extra space after the heading -- half of whatever came above it.
+% Don't allow stretch, though.
+\kern .5 \csname #2headingskip\endcsname
+%
+% Do not let the kern be a potential breakpoint, as it would be if it
+% was followed by glue.
+\nobreak
+%
+% We'll almost certainly start a paragraph next, so don't let that
+% glue accumulate.  (Not a breakpoint because it's preceded by a
+% discardable item.)
+\vskip-\parskip
+%
+% This is purely so the last item on the list is a known \penalty >
+% 10000.  This is so \startdefun can avoid allowing breakpoints after
+% section headings.  Otherwise, it would insert a valid breakpoint between:
+%
+%   @section sec-whatever
+%   @deffn def-whatever
+\penalty 10001
 }
 \message{toc,}
 % Table of contents.
 \newwrite\tocfile
 % Write an entry to the toc file, opening it if necessary.
-% Called from @chapter, etc.  We supply {\folio} at the end of the
+% Called from @chapter, etc.
-% argument, which will end up as the last argument to the \...entry macro.
+%
-%
+% Example usage: \writetocentry{sec}{Section Name}{\the\chapno.\the\secno}
-% We open the .toc file here instead of at @setfilename or any other
+% We append the current node name (if any) and page number as additional
-% given time so that @contents can be put in the document anywhere.
+% arguments for the \{chap,sec,...}entry macros which will eventually
+% read this.  The node name is used in the pdf outlines as the
+% destination to jump to.
+%
+% We open the .toc file for writing here instead of at @setfilename (or
+% any other fixed time) so that @contents can be anywhere in the document.
+% But if #1 is `omit', then we don't do anything.  This is used for the
+% table of contents chapter openings themselves.
 %
 \newif\iftocfileopened
-\def\writetocentry#1{%
+\def\omitkeyword{omit}%
-\iftocfileopened\else
+%
-\immediate\openout\tocfile = \jobname.toc
+\def\writetocentry#1#2#3{%
-\global\tocfileopenedtrue
+\edef\writetoctype{#1}%
+\ifx\writetoctype\omitkeyword \else
+\iftocfileopened\else
+\immediate\openout\tocfile = \jobname.toc
+\global\tocfileopenedtrue
+\fi
+%
+\iflinks
+{\atdummies
+\edef\temp{%
+\write\tocfile{@#1entry{#2}{#3}{\lastnode}{\noexpand\folio}}}%
+\temp
+}%
+\fi
 \fi
-\iflinks \write\tocfile{#1{\folio}}\fi
+%
+% Tell \shipout to create a pdf destination on each page, if we're
+% writing pdf.  These are used in the table of contents.  We can't
+% just write one on every page because the title pages are numbered
+% 1 and 2 (the page numbers aren't printed), and so are the first
+% two pages of the document.  Thus, we'd have two destinations named
+% `1', and two named `2'.
+\ifpdf \global\pdfmakepagedesttrue \fi
+}
+% These characters do not print properly in the Computer Modern roman
+% fonts, so we must take special care.  This is more or less redundant
+% with the Texinfo input format setup at the end of this file.
+%
+\def\activecatcodes{%
+\catcode`\"=\active
+\catcode`\$=\active
+\catcode`\<=\active
+\catcode`\>=\active
+\catcode`\\=\active
+\catcode`\^=\active
+\catcode`\_=\active
+\catcode`\|=\active
+\catcode`\~=\active
+}
+% Read the toc file, which is essentially Texinfo input.
+\def\readtocfile{%
+\setupdatafile
+\activecatcodes
+\input \tocreadfilename
 }
 \newskip\contentsrightmargin \contentsrightmargin=1in
 \newcount\savepageno
 \newcount\lastnegativepageno \lastnegativepageno = -1
-% Finish up the main text and prepare to read what we've written
+% Prepare to read what we've written to \tocfile.
-% to \tocfile.
 %
 \def\startcontents#1{%
 % If @setchapternewpage on, and @headings double, the contents should
 % start on an odd page, unlike chapters.  Thus, we maintain
 % \contentsalignmacro in parallel with \pagealignmacro.
 % From: Torbjorn Granlund <tege@matematik.su.se>
 \contentsalignmacro
 \immediate\closeout\tocfile
 %
 % Don't need to put `Contents' or `Short Contents' in the headline.
 % It is abundantly clear what they are.
-\unnumbchapmacro{#1}\def\thischapter{}%
+\chapmacro{#1}{Yomitfromtoc}{}%
-\savepageno = \pageno
+%
-\begingroup                  % Set up to handle contents files properly.
+\savepageno = \pageno
-\catcode`\\=0  \catcode`\{=1  \catcode`\}=2  \catcode`\@=11
+\begingroup                  % Set up to handle contents files properly.
-% We can't do this, because then an actual ^ in a section
+\raggedbottom              % Worry more about breakpoints than the bottom.
-% title fails, e.g., @chapter ^ -- exponentiation.  --karl, 9jul97.
+\advance\hsize by -\contentsrightmargin % Don't use the full line length.
-%\catcode`\^=7 % to see ^^e4 as \"a etc. juha@piuha.ydi.vtt.fi
+%
-\raggedbottom             % Worry more about breakpoints than the bottom.
+% Roman numerals for page numbers.
-\advance\hsize by -\contentsrightmargin % Don't use the full line length.
+\ifnum \pageno>0 \global\pageno = \lastnegativepageno \fi
-%
+}
-% Roman numerals for page numbers.
-\ifnum \pageno>0 \pageno = \lastnegativepageno \fi
+% redefined for the two-volume lispref.  We always output on
-}
+% \jobname.toc even if this is redefined.
+%
+\def\tocreadfilename{\jobname.toc}
 % Normal (long) toc.
+%
 \def\contents{%
 \startcontents{\putwordTOC}%
-\openin 1 \jobname.toc
+\openin 1 \tocreadfilename\space
 \ifeof 1 \else
-\closein 1
+\readtocfile
-\input \jobname.toc
+\fi
-\fi
+\vfill \eject
-\vfill \eject
+\contentsalignmacro % in case @setchapternewpage odd is in effect
-\contentsalignmacro % in case @setchapternewpage odd is in effect
+\ifeof 1 \else
 \pdfmakeoutlines
-\endgroup
+\fi
-\lastnegativepageno = \pageno
+\closein 1
-\pageno = \savepageno
+\endgroup
+\lastnegativepageno = \pageno
+\global\pageno = \savepageno
 }
 % And just the chapters.
 \def\summarycontents{%
 \startcontents{\putwordShortTOC}%
 %
-\let\chapentry = \shortchapentry
+\let\numchapentry = \shortchapentry
-\let\unnumbchapentry = \shortunnumberedentry
+\let\appentry = \shortchapentry
-% We want a true roman here for the page numbers.
+\let\unnchapentry = \shortunnchapentry
-\secfonts
+% We want a true roman here for the page numbers.
-\let\rm=\shortcontrm \let\bf=\shortcontbf \let\sl=\shortcontsl
+\secfonts
-\rm
+\let\rm=\shortcontrm \let\bf=\shortcontbf
-\hyphenpenalty = 10000
+\let\sl=\shortcontsl \let\tt=\shortconttt
-\advance\baselineskip by 1pt % Open it up a little.
+\rm
-\def\secentry ##1##2##3##4{}
+\hyphenpenalty = 10000
-\def\unnumbsecentry ##1##2{}
+\advance\baselineskip by 1pt % Open it up a little.
-\def\subsecentry ##1##2##3##4##5{}
+\def\numsecentry##1##2##3##4{}
-\def\unnumbsubsecentry ##1##2{}
+\let\appsecentry = \numsecentry
-\def\subsubsecentry ##1##2##3##4##5##6{}
+\let\unnsecentry = \numsecentry
-\def\unnumbsubsubsecentry ##1##2{}
+\let\numsubsecentry = \numsecentry
-\openin 1 \jobname.toc
+\let\appsubsecentry = \numsecentry
-\ifeof 1 \else
+\let\unnsubsecentry = \numsecentry
-\closein 1
+\let\numsubsubsecentry = \numsecentry
-\input \jobname.toc
+\let\appsubsubsecentry = \numsecentry
-\fi
+\let\unnsubsubsecentry = \numsecentry
-\vfill \eject
+\openin 1 \tocreadfilename\space
-\contentsalignmacro % in case @setchapternewpage odd is in effect
+\ifeof 1 \else
-\endgroup
+\readtocfile
-\lastnegativepageno = \pageno
+\fi
-\pageno = \savepageno
+\closein 1
+\vfill \eject
+\contentsalignmacro % in case @setchapternewpage odd is in effect
+\endgroup
+\lastnegativepageno = \pageno
+\global\pageno = \savepageno
 }
 \let\shortcontents = \summarycontents
-\ifpdf
+% Typeset the label for a chapter or appendix for the short contents.
-\pdfcatalog{/PageMode /UseOutlines}%
+% The arg is, e.g., `A' for an appendix, or `3' for a chapter.
-\fi
+%
+\def\shortchaplabel#1{%
+% This space should be enough, since a single number is .5em, and the
+% widest letter (M) is 1em, at least in the Computer Modern fonts.
+% But use \hss just in case.
+% (This space doesn't include the extra space that gets added after
+% the label; that gets put in by \shortchapentry above.)
+%
+% We'd like to right-justify chapter numbers, but that looks strange
+% with appendix letters.  And right-justifying numbers and
+% left-justifying letters looks strange when there is less than 10
+% chapters.  Have to read the whole toc once to know how many chapters
+% there are before deciding ...
+\hbox to 1em{#1\hss}%
+}
 % These macros generate individual entries in the table of contents.
 % The first argument is the chapter or section name.
 % The last argument is the page number.
 % The arguments in between are the chapter number, section number, ...
-% Chapter-level things, for both the long and short contents.
+% Chapters, in the main contents.
-\def\chapentry#1#2#3{\dochapentry{#2\labelspace#1}{#3}}
+\def\numchapentry#1#2#3#4{\dochapentry{#2\labelspace#1}{#4}}
+%
-% See comments in \dochapentry re vbox and related settings
+% Chapters, in the short toc.
-\def\shortchapentry#1#2#3{%
+% See comments in \dochapentry re vbox and related settings.
-\tocentry{\shortchaplabel{#2}\labelspace #1}{\doshortpageno\bgroup#3\egroup}%
+\def\shortchapentry#1#2#3#4{%
-}
+\tocentry{\shortchaplabel{#2}\labelspace #1}{\doshortpageno\bgroup#4\egroup}%
+}
-% Typeset the label for a chapter or appendix for the short contents.
-% The arg is, e.g. `Appendix A' for an appendix, or `3' for a chapter.
+% Appendices, in the main contents.
-% We could simplify the code here by writing out an \appendixentry
+% Need the word Appendix, and a fixed-size box.
-% command in the toc file for appendices, instead of using \chapentry
+%
-% for both, but it doesn't seem worth it.
+\def\appendixbox#1{%
-%
+% We use M since it's probably the widest letter.
-\newdimen\shortappendixwidth
+\setbox0 = \hbox{\putwordAppendix{} M}%
-%
+\hbox to \wd0{\putwordAppendix{} #1\hss}}
-\def\shortchaplabel#1{%
+%
-% Compute width of word "Appendix", may change with language.
+\def\appentry#1#2#3#4{\dochapentry{\appendixbox{#2}\labelspace#1}{#4}}
-\setbox0 = \hbox{\shortcontrm \putwordAppendix}%
-\shortappendixwidth = \wd0
+% Unnumbered chapters.
-%
+\def\unnchapentry#1#2#3#4{\dochapentry{#1}{#4}}
-% We typeset #1 in a box of constant width, regardless of the text of
+\def\shortunnchapentry#1#2#3#4{\tocentry{#1}{\doshortpageno\bgroup#4\egroup}}
-% #1, so the chapter titles will come out aligned.
-\setbox0 = \hbox{#1}%
-\dimen0 = \ifdim\wd0 > \shortappendixwidth \shortappendixwidth \else 0pt \fi
-%
-% This space should be plenty, since a single number is .5em, and the
-% widest letter (M) is 1em, at least in the Computer Modern fonts.
-% (This space doesn't include the extra space that gets added after
-% the label; that gets put in by \shortchapentry above.)
-\advance\dimen0 by 1.1em
-\hbox to \dimen0{#1\hfil}%
-}
-\def\unnumbchapentry#1#2{\dochapentry{#1}{#2}}
-\def\shortunnumberedentry#1#2{\tocentry{#1}{\doshortpageno\bgroup#2\egroup}}
 % Sections.
-\def\secentry#1#2#3#4{\dosecentry{#2.#3\labelspace#1}{#4}}
+\def\numsecentry#1#2#3#4{\dosecentry{#2\labelspace#1}{#4}}
-\def\unnumbsecentry#1#2{\dosecentry{#1}{#2}}
+\let\appsecentry=\numsecentry
+\def\unnsecentry#1#2#3#4{\dosecentry{#1}{#4}}
 % Subsections.
-\def\subsecentry#1#2#3#4#5{\dosubsecentry{#2.#3.#4\labelspace#1}{#5}}
+\def\numsubsecentry#1#2#3#4{\dosubsecentry{#2\labelspace#1}{#4}}
-\def\unnumbsubsecentry#1#2{\dosubsecentry{#1}{#2}}
+\let\appsubsecentry=\numsubsecentry
+\def\unnsubsecentry#1#2#3#4{\dosubsecentry{#1}{#4}}
 % And subsubsections.
-\def\subsubsecentry#1#2#3#4#5#6{%
+\def\numsubsubsecentry#1#2#3#4{\dosubsubsecentry{#2\labelspace#1}{#4}}
-\dosubsubsecentry{#2.#3.#4.#5\labelspace#1}{#6}}
+\let\appsubsubsecentry=\numsubsubsecentry
-\def\unnumbsubsubsecentry#1#2{\dosubsubsecentry{#1}{#2}}
+\def\unnsubsubsecentry#1#2#3#4{\dosubsubsecentry{#1}{#4}}
 % This parameter controls the indentation of the various levels.
-\newdimen\tocindent \tocindent = 3pc
+% Same as \defaultparindent.
+\newdimen\tocindent \tocindent = 15pt
 % Now for the actual typesetting. In all these, #1 is the text and #2 is the
 % page number.
 %
 % If the toc has to be broken over pages, we want it to be at chapters
 \def\dosubsubsecentry#1#2{\begingroup
 \subsubsecentryfonts \leftskip=3\tocindent
 \tocentry{#1}{\dopageno\bgroup#2\egroup}%
 \endgroup}
-% Final typesetting of a toc entry; we use the same \entry macro as for
+% We use the same \entry macro as for the index entries.
-% the index entries, but we want to suppress hyphenation here.  (We
+\let\tocentry = \entry
-% can't do that in the \entry macro, since index entries might consist
-% of hyphenated-identifiers-that-do-not-fit-on-a-line-and-nothing-else.)
-\def\tocentry#1#2{\begingroup
-\vskip 0pt plus1pt % allow a little stretch for the sake of nice page breaks
-% Do not use \turnoffactive in these arguments.  Since the toc is
-% typeset in cmr, so characters such as _ would come out wrong; we
-% have to do the usual translation tricks.
-\entry{#1}{#2}%
-\endgroup}
 % Space between chapter (or whatever) number and the title.
 \def\labelspace{\hskip1em \relax}
 \def\dopageno#1{{\rm #1}}
 \def\doshortpageno#1{{\rm #1}}
 \def\chapentryfonts{\secfonts \rm}
 \def\secentryfonts{\textfonts}
-\let\subsecentryfonts = \textfonts
+\def\subsecentryfonts{\textfonts}
-\let\subsubsecentryfonts = \textfonts
+\def\subsubsecentryfonts{\textfonts}
 \message{environments,}
 % @foo ... @end foo.
-% Since these characters are used in examples, it should be an even number of
+% @point{}, @result{}, @expansion{}, @print{}, @equiv{}.
+%
+% Since these characters are used in examples, they should be an even number of
 % \tt widths. Each \tt character is 1en, so two makes it 1em.
-% Furthermore, these definitions must come after we define our fonts.
+%
-\newbox\dblarrowbox    \newbox\longdblarrowbox
-\newbox\pushcharbox    \newbox\bullbox
-\newbox\equivbox       \newbox\errorbox
-%{\tentt
-%\global\setbox\dblarrowbox = \hbox to 1em{\hfil$\Rightarrow$\hfil}
-%\global\setbox\longdblarrowbox = \hbox to 1em{\hfil$\mapsto$\hfil}
-%\global\setbox\pushcharbox = \hbox to 1em{\hfil$\dashv$\hfil}
-%\global\setbox\equivbox = \hbox to 1em{\hfil$\ptexequiv$\hfil}
-% Adapted from the manmac format (p.420 of TeXbook)
-%\global\setbox\bullbox = \hbox to 1em{\kern.15em\vrule height .75ex width .85ex
-%                                      depth .1ex\hfil}
-%}
-% @point{}, @result{}, @expansion{}, @print{}, @equiv{}.
 \def\point{$\star$}
-\def\result{\leavevmode\raise.15ex\hbox to 1em{\hfil$\Rightarrow$\hfil}}
+\def\arrow{\leavevmode\raise.05ex\hbox to 1em{\hfil$\rightarrow$\hfil}}
-\def\expansion{\leavevmode\raise.1ex\hbox to 1em{\hfil$\mapsto$\hfil}}
+\def\result{\leavevmode\raise.05ex\hbox to 1em{\hfil$\Rightarrow$\hfil}}
+\def\expansion{\leavevmode\hbox to 1em{\hfil$\mapsto$\hfil}}
 \def\print{\leavevmode\lower.1ex\hbox to 1em{\hfil$\dashv$\hfil}}
-\def\equiv{\leavevmode\lower.1ex\hbox to 1em{\hfil$\ptexequiv$\hfil}}
+\def\equiv{\leavevmode\hbox to 1em{\hfil$\ptexequiv$\hfil}}
+% The @error{} command.
 % Adapted from the TeXbook's \boxit.
+%
+\newbox\errorbox
+%
 {\tentt \global\dimen0 = 3em}% Width of the box.
 \dimen2 = .55pt % Thickness of rules
 % The text. (`r' is open on the right, `e' somewhat less so on the left.)
-\setbox0 = \hbox{\kern-.75pt \tensf error\kern-1.5pt}
+\setbox0 = \hbox{\kern-.75pt \reducedsf error\kern-1.5pt}
+%
-\global\setbox\errorbox=\hbox to \dimen0{\hfil
+\setbox\errorbox=\hbox to \dimen0{\hfil
 \hsize = \dimen0 \advance\hsize by -5.8pt % Space to left+right.
 \advance\hsize by -2\dimen2 % Rules.
-\vbox{
+\vbox{%
 \hrule height\dimen2
 \hbox{\vrule width\dimen2 \kern3pt          % Space to left of text.
 \vtop{\kern2.4pt \box0 \kern2.4pt}% Space above/below.
 \kern3pt\vrule width\dimen2}% Space to right.
 \hrule height\dimen2}
 \hfil}
+%
-% The @error{} command.
 \def\error{\leavevmode\lower.7ex\copy\errorbox}
 % @tex ... @end tex    escapes into raw Tex temporarily.
 % One exception: @ is still an escape character, so that @end tex works.
 % But \@ or @@ will get a plain tex @ character.
-\def\tex{\begingroup
+\envdef\tex{%
 \catcode `\\=0 \catcode `\{=1 \catcode `\}=2
 \catcode `\$=3 \catcode `\&=4 \catcode `\#=6
-\catcode `\^=7 \catcode `\_=8 \catcode `\~=13 \let~=\tie
+\catcode `\^=7 \catcode `\_=8 \catcode `\~=\active \let~=\tie
 \catcode `\%=14
-\catcode 43=12 % plus
+\catcode `\+=\other
-\catcode`\"=12
+\catcode `\"=\other
-\catcode`\==12
+\catcode `\|=\other
-\catcode`\|=12
+\catcode `\<=\other
-\catcode`\<=12
+\catcode `\>=\other
-\catcode`\>=12
 \escapechar=`\\
 %
 \let\b=\ptexb
 \let\bullet=\ptexbullet
 \let\c=\ptexc
 \let\.=\ptexdot
 \let\dots=\ptexdots
 \let\equiv=\ptexequiv
 \let\!=\ptexexclam
 \let\i=\ptexi
+\let\indent=\ptexindent
+\let\noindent=\ptexnoindent
 \let\{=\ptexlbrace
 \let\+=\tabalign
 \let\}=\ptexrbrace
+\let\/=\ptexslash
 \let\*=\ptexstar
 \let\t=\ptext
+\expandafter \let\csname top\endcsname=\ptextop  % outer
+\let\frenchspacing=\plainfrenchspacing
 %
 \def\endldots{\mathinner{\ldots\ldots\ldots\ldots}}%
 \def\enddots{\relax\ifmmode\endldots\else$\mathsurround=0pt \endldots\,$\fi}%
 \def\@{@}%
-\let\Etex=\endgroup}
+}
+% There is no need to define \Etex.
-% Define @lisp ... @endlisp.
-% @lisp does a \begingroup so it can rebind things,
+% Define @lisp ... @end lisp.
-% including the definition of @endlisp (which normally is erroneous).
+% @lisp environment forms a group so it can rebind things,
+% including the definition of @end lisp (which normally is erroneous).
 % Amount to narrow the margins by for @lisp.
 \newskip\lispnarrowing \lispnarrowing=0.4in
 % This is the definition that ^^M gets inside @lisp, @example, and other
 % such environments.  \null is better than a space, since it doesn't
 % have any width.
 \def\lisppar{\null\endgraf}
-% Make each space character in the input produce a normal interword
-% space in the output.  Don't allow a line break at this space, as this
-% is used only in environments like @example, where each line of input
-% should produce a line of output anyway.
-%
-{\obeyspaces %
-\gdef\sepspaces{\obeyspaces\let =\tie}}
-% Define \obeyedspace to be our active space, whatever it is.  This is
-% for use in \parsearg.
-{\sepspaces%
-\global\let\obeyedspace= }
 % This space is always present above and below environments.
 \newskip\envskipamount \envskipamount = 0pt
 % Make spacing and below environment symmetrical.  We use \parskip here
 % to help in doing that, since in @example-like environments \parskip
 % is reset to zero; thus the \afterenvbreak inserts no space -- but the
-% start of the next paragraph will insert \parskip
+% start of the next paragraph will insert \parskip.
 %
-\def\aboveenvbreak{{\advance\envskipamount by \parskip
+\def\aboveenvbreak{{%
-\endgraf \ifdim\lastskip<\envskipamount
+% =10000 instead of <10000 because of a special case in \itemzzz and
-\removelastskip \penalty-50 \vskip\envskipamount \fi}}
+% \sectionheading, q.v.
+\ifnum \lastpenalty=10000 \else
+\advance\envskipamount by \parskip
+\endgraf
+\ifdim\lastskip<\envskipamount
+\removelastskip
+% it's not a good place to break if the last penalty was \nobreak
+% or better ...
+\ifnum\lastpenalty<10000 \penalty-50 \fi
+\vskip\envskipamount
+\fi
+\fi
+}}
 \let\afterenvbreak = \aboveenvbreak
-% \nonarrowing is a flag.  If "set", @lisp etc don't narrow margins.
+% \nonarrowing is a flag.  If "set", @lisp etc don't narrow margins; it will
+% also clear it, so that its embedded environments do the narrowing again.
 \let\nonarrowing=\relax
 % @cartouche ... @end cartouche: draw rectangle w/rounded corners around
 % environment contents.
 \font\circle=lcircle10
 \cbl\leaders\hrule height\circthick\hfil\cbr
 \hskip\rskip}}
 %
 \newskip\lskip\newskip\rskip
-\long\def\cartouche{%
+\envdef\cartouche{%
-\begingroup
+\ifhmode\par\fi  % can't be in the midst of a paragraph.
-\lskip=\leftskip \rskip=\rightskip
+\startsavinginserts
-\leftskip=0pt\rightskip=0pt %we want these *outside*.
+\lskip=\leftskip \rskip=\rightskip
-\cartinner=\hsize \advance\cartinner by-\lskip
+\leftskip=0pt\rightskip=0pt % we want these *outside*.
-\advance\cartinner by-\rskip
+\cartinner=\hsize \advance\cartinner by-\lskip
-\cartouter=\hsize
+\advance\cartinner by-\rskip
-\advance\cartouter by 18.4pt % allow for 3pt kerns on either
+\cartouter=\hsize
-%                                    side, and for 6pt waste from
+\advance\cartouter by 18.4pt	% allow for 3pt kerns on either
-%                                    each corner char, and rule thickness
+				% side, and for 6pt waste from
-\normbskip=\baselineskip \normpskip=\parskip \normlskip=\lineskip
+				% each corner char, and rule thickness
-% Flag to tell @lisp, etc., not to narrow margin.
+\normbskip=\baselineskip \normpskip=\parskip \normlskip=\lineskip
-\let\nonarrowing=\comment
+% Flag to tell @lisp, etc., not to narrow margin.
-\vbox\bgroup
+\let\nonarrowing = t%
-\baselineskip=0pt\parskip=0pt\lineskip=0pt
+\vbox\bgroup
-\carttop
+\baselineskip=0pt\parskip=0pt\lineskip=0pt
-\hbox\bgroup
+\carttop
-\hskip\lskip
+\hbox\bgroup
-\vrule\kern3pt
+	  \hskip\lskip
-\vbox\bgroup
+	  \vrule\kern3pt
-\hsize=\cartinner
+	  \vbox\bgroup
-\kern3pt
+	      \kern3pt
-\begingroup
+	      \hsize=\cartinner
-\baselineskip=\normbskip
+	      \baselineskip=\normbskip
-\lineskip=\normlskip
+	      \lineskip=\normlskip
-\parskip=\normpskip
+	      \parskip=\normpskip
-\vskip -\parskip
+	      \vskip -\parskip
+	      \comment % For explanation, see the end of \def\group.
+}
 \def\Ecartouche{%
-\endgroup
+\ifhmode\par\fi
-\kern3pt
+	      \kern3pt
-\egroup
+	  \egroup
-\kern3pt\vrule
+	  \kern3pt\vrule
-\hskip\rskip
+	  \hskip\rskip
 \egroup
 \cartbot
 \egroup
-\endgroup
+\checkinserts
-}}
+}
 % This macro is called at the beginning of all the @example variants,
 % inside a group.
 \def\nonfillstart{%
 \aboveenvbreak
-\inENV % This group ends at the end of the body
 \hfuzz = 12pt % Don't be fussy
 \sepspaces % Make spaces be word-separators rather than space tokens.
-\singlespace
 \let\par = \lisppar % don't ignore blank lines
 \obeylines % each line of input is a line of output
 \parskip = 0pt
 \parindent = 0pt
 \emergencystretch = 0pt % don't try to avoid overfull boxes
-% @cartouche defines \nonarrowing to inhibit narrowing
-% at next level down.
 \ifx\nonarrowing\relax
 \advance \leftskip by \lispnarrowing
 \exdentamount=\lispnarrowing
-\let\exdent=\nofillexdent
+\else
-\let\nonarrowing=\relax
+\let\nonarrowing = \relax
 \fi
-}
+\let\exdent=\nofillexdent
+}
-% Define the \E... control sequence only if we are inside the particular
-% environment, so the error checking in \end will work.
+% If you want all examples etc. small: @set dispenvsize small.
-%
+% If you want even small examples the full size: @set dispenvsize nosmall.
-% To end an @example-like environment, we first end the paragraph (via
+% This affects the following displayed environments:
-% \afterenvbreak's vertical glue), and then the group.  That way we keep
+%    @example, @display, @format, @lisp
-% the zero \parskip that the environments set -- \parskip glue will be
+%
-% inserted at the beginning of the next paragraph in the document, after
+\def\smallword{small}
-% the environment.
+\def\nosmallword{nosmall}
-%
+\let\SETdispenvsize\relax
-\def\nonfillfinish{\afterenvbreak\endgroup}
+\def\setnormaldispenv{%
+\ifx\SETdispenvsize\smallword
-% @lisp: indented, narrowed, typewriter font.
+% end paragraph for sake of leading, in case document has no blank
-\def\lisp{\begingroup
+% line.  This is redundant with what happens in \aboveenvbreak, but
+% we need to do it before changing the fonts, and it's inconvenient
+% to change the fonts afterward.
+\ifnum \lastpenalty=10000 \else \endgraf \fi
+\smallexamplefonts \rm
+\fi
+}
+\def\setsmalldispenv{%
+\ifx\SETdispenvsize\nosmallword
+\else
+\ifnum \lastpenalty=10000 \else \endgraf \fi
+\smallexamplefonts \rm
+\fi
+}
+% We often define two environments, @foo and @smallfoo.
+% Let's do it by one command:
+\def\makedispenv #1#2{
+\expandafter\envdef\csname#1\endcsname {\setnormaldispenv #2}
+\expandafter\envdef\csname small#1\endcsname {\setsmalldispenv #2}
+\expandafter\let\csname E#1\endcsname \afterenvbreak
+\expandafter\let\csname Esmall#1\endcsname \afterenvbreak
+}
+% Define two synonyms:
+\def\maketwodispenvs #1#2#3{
+\makedispenv{#1}{#3}
+\makedispenv{#2}{#3}
+}
+% @lisp: indented, narrowed, typewriter font; @example: same as @lisp.
+%
+% @smallexample and @smalllisp: use smaller fonts.
+% Originally contributed by Pavel@xerox.
+%
+\maketwodispenvs {lisp}{example}{%
 \nonfillstart
-\let\Elisp = \nonfillfinish
+\tt\quoteexpand
-\tt
 \let\kbdfont = \kbdexamplefont % Allow @kbd to do something special.
 \gobble       % eat return
 }
+% @display/@smalldisplay: same as @lisp except keep current font.
-% @example: Same as @lisp.
+%
-\def\example{\begingroup \def\Eexample{\nonfillfinish\endgroup}\lisp}
+\makedispenv {display}{%
-% @small... is usually equivalent to the non-small (@smallbook
-% redefines).  We must call \example (or whatever) last in the
-% definition, since it reads the return following the @example (or
-% whatever) command.
-%
-% This actually allows (for example) @end display inside an
-% @smalldisplay.  Too bad, but makeinfo will catch the error anyway.
-%
-\def\smalldisplay{\begingroup\def\Esmalldisplay{\nonfillfinish\endgroup}\display}
-\def\smallexample{\begingroup\def\Esmallexample{\nonfillfinish\endgroup}\lisp}
-\def\smallformat{\begingroup\def\Esmallformat{\nonfillfinish\endgroup}\format}
-\def\smalllisp{\begingroup\def\Esmalllisp{\nonfillfinish\endgroup}\lisp}
-% Real @smallexample and @smalllisp (when @smallbook): use smaller fonts.
-% Originally contributed by Pavel@xerox.
-\def\smalllispx{\begingroup
-\def\Esmalllisp{\nonfillfinish\endgroup}%
-\def\Esmallexample{\nonfillfinish\endgroup}%
-\smallfonts
-\lisp
-}
-% @display: same as @lisp except keep current font.
-%
-\def\display{\begingroup
 \nonfillstart
-\let\Edisplay = \nonfillfinish
 \gobble
 }
-% @smalldisplay (when @smallbook): @display plus smaller fonts.
+% @format/@smallformat: same as @display except don't narrow margins.
 %
-\def\smalldisplayx{\begingroup
+\makedispenv{format}{%
-\def\Esmalldisplay{\nonfillfinish\endgroup}%
+\let\nonarrowing = t%
-\smallfonts \rm
-\display
-}
-% @format: same as @display except don't narrow margins.
-%
-\def\format{\begingroup
-\let\nonarrowing = t
 \nonfillstart
-\let\Eformat = \nonfillfinish
 \gobble
 }
-% @smallformat (when @smallbook): @format plus smaller fonts.
+% @flushleft: same as @format, but doesn't obey \SETdispenvsize.
-%
+\envdef\flushleft{%
-\def\smallformatx{\begingroup
+\let\nonarrowing = t%
-\def\Esmallformat{\nonfillfinish\endgroup}%
+\nonfillstart
-\smallfonts \rm
+\gobble
-\format
+}
-}
+\let\Eflushleft = \afterenvbreak
-% @flushleft (same as @format).
-%
-\def\flushleft{\begingroup \def\Eflushleft{\nonfillfinish\endgroup}\format}
 % @flushright.
 %
-\def\flushright{\begingroup
+\envdef\flushright{%
-\let\nonarrowing = t
+\let\nonarrowing = t%
 \nonfillstart
-\let\Eflushright = \nonfillfinish
 \advance\leftskip by 0pt plus 1fill
 \gobble
 }
+\let\Eflushright = \afterenvbreak
 % @quotation does normal linebreaking (hence we can't use \nonfillstart)
-% and narrows the margins.
+% and narrows the margins.  We keep \parskip nonzero in general, since
-%
+% we're doing normal filling.  So, when using \aboveenvbreak and
-\def\quotation{%
+% \afterenvbreak, temporarily make \parskip 0.
-\begingroup\inENV %This group ends at the end of the @quotation body
+%
+\envdef\quotation{%
 {\parskip=0pt \aboveenvbreak}% because \aboveenvbreak inserts \parskip
-\singlespace
 \parindent=0pt
-% We have retained a nonzero parskip for the environment, since we're
-% doing normal filling. So to avoid extra space below the environment...
-\def\Equotation{\parskip = 0pt \nonfillfinish}%
 %
 % @cartouche defines \nonarrowing to inhibit narrowing at next level down.
 \ifx\nonarrowing\relax
 \advance\leftskip by \lispnarrowing
 \advance\rightskip by \lispnarrowing
 \exdentamount = \lispnarrowing
+\else
 \let\nonarrowing = \relax
 \fi
+\parsearg\quotationlabel
+}
+% We have retained a nonzero parskip for the environment, since we're
+% doing normal filling.
+%
+\def\Equotation{%
+\par
+\ifx\quotationauthor\undefined\else
+% indent a bit.
+\leftline{\kern 2\leftskip \sl ---\quotationauthor}%
+\fi
+{\parskip=0pt \afterenvbreak}%
+}
+% If we're given an argument, typeset it in bold with a colon after.
+\def\quotationlabel#1{%
+\def\temp{#1}%
+\ifx\temp\empty \else
+{\bf #1: }%
+\fi
+}
+% LaTeX-like @verbatim...@end verbatim and @verb{<char>...<char>}
+% If we want to allow any <char> as delimiter,
+% we need the curly braces so that makeinfo sees the @verb command, eg:
+% `@verbx...x' would look like the '@verbx' command.  --janneke@gnu.org
+%
+% [Knuth]: Donald Ervin Knuth, 1996.  The TeXbook.
+%
+% [Knuth] p.344; only we need to do the other characters Texinfo sets
+% active too.  Otherwise, they get lost as the first character on a
+% verbatim line.
+\def\dospecials{%
+\do\ \do\\\do\{\do\}\do\$\do\&%
+\do\#\do\^\do\^^K\do\_\do\^^A\do\%\do\~%
+\do\<\do\>\do\|\do\@\do+\do\"%
+}
+%
+% [Knuth] p. 380
+\def\uncatcodespecials{%
+\def\do##1{\catcode`##1=\other}\dospecials}
+%
+% [Knuth] pp. 380,381,391
+% Disable Spanish ligatures ?` and !` of \tt font
+\begingroup
+\catcode`\`=\active\gdef`{\relax\lq}
+\endgroup
+%
+% Setup for the @verb command.
+%
+% Eight spaces for a tab
+\begingroup
+\catcode`\^^I=\active
+\gdef\tabeightspaces{\catcode`\^^I=\active\def^^I{\ \ \ \ \ \ \ \ }}
+\endgroup
+%
+\def\setupverb{%
+\tt  % easiest (and conventionally used) font for verbatim
+\def\par{\leavevmode\endgraf}%
+\catcode`\`=\active
+\tabeightspaces
+% Respect line breaks,
+% print special symbols as themselves, and
+% make each space count
+% must do in this order:
+\obeylines \uncatcodespecials \sepspaces
+}
+% Setup for the @verbatim environment
+%
+% Real tab expansion
+\newdimen\tabw \setbox0=\hbox{\tt\space} \tabw=8\wd0 % tab amount
+%
+\def\starttabbox{\setbox0=\hbox\bgroup}
+% Allow an option to not replace quotes with a regular directed right
+% quote/apostrophe (char 0x27), but instead use the undirected quote
+% from cmtt (char 0x0d).  The undirected quote is ugly, so don't make it
+% the default, but it works for pasting with more pdf viewers (at least
+% evince), the lilypond developers report.  xpdf does work with the
+% regular 0x27.
+%
+\def\codequoteright{%
+\expandafter\ifx\csname SETtxicodequoteundirected\endcsname\relax
+\expandafter\ifx\csname SETcodequoteundirected\endcsname\relax
+'%
+\else \char'15 \fi
+\else \char'15 \fi
+}
+%
+% and a similar option for the left quote char vs. a grave accent.
+% Modern fonts display ASCII 0x60 as a grave accent, so some people like
+% the code environments to do likewise.
+%
+\def\codequoteleft{%
+\expandafter\ifx\csname SETtxicodequotebacktick\endcsname\relax
+\expandafter\ifx\csname SETcodequotebacktick\endcsname\relax
+`%
+\else \char'22 \fi
+\else \char'22 \fi
+}
+%
+\begingroup
+\catcode`\^^I=\active
+\gdef\tabexpand{%
+\catcode`\^^I=\active
+\def^^I{\leavevmode\egroup
+\dimen0=\wd0 % the width so far, or since the previous tab
+\divide\dimen0 by\tabw
+\multiply\dimen0 by\tabw % compute previous multiple of \tabw
+\advance\dimen0 by\tabw  % advance to next multiple of \tabw
+\wd0=\dimen0 \box0 \starttabbox
+}%
+}
+\catcode`\'=\active
+\gdef\rquoteexpand{\catcode\rquoteChar=\active \def'{\codequoteright}}%
+%
+\catcode`\`=\active
+\gdef\lquoteexpand{\catcode\lquoteChar=\active \def`{\codequoteleft}}%
+%
+\gdef\quoteexpand{\rquoteexpand \lquoteexpand}%
+\endgroup
+% start the verbatim environment.
+\def\setupverbatim{%
+\let\nonarrowing = t%
+\nonfillstart
+% Easiest (and conventionally used) font for verbatim
+\tt
+\def\par{\leavevmode\egroup\box0\endgraf}%
+\catcode`\`=\active
+\tabexpand
+\quoteexpand
+% Respect line breaks,
+% print special symbols as themselves, and
+% make each space count
+% must do in this order:
+\obeylines \uncatcodespecials \sepspaces
+\everypar{\starttabbox}%
+}
+% Do the @verb magic: verbatim text is quoted by unique
+% delimiter characters.  Before first delimiter expect a
+% right brace, after last delimiter expect closing brace:
+%
+%    \def\doverb'{'<char>#1<char>'}'{#1}
+%
+% [Knuth] p. 382; only eat outer {}
+\begingroup
+\catcode`[=1\catcode`]=2\catcode`\{=\other\catcode`\}=\other
+\gdef\doverb{#1[\def\next##1#1}[##1\endgroup]\next]
+\endgroup
+%
+\def\verb{\begingroup\setupverb\doverb}
+%
+%
+% Do the @verbatim magic: define the macro \doverbatim so that
+% the (first) argument ends when '@end verbatim' is reached, ie:
+%
+%     \def\doverbatim#1@end verbatim{#1}
+%
+% For Texinfo it's a lot easier than for LaTeX,
+% because texinfo's \verbatim doesn't stop at '\end{verbatim}':
+% we need not redefine '\', '{' and '}'.
+%
+% Inspired by LaTeX's verbatim command set [latex.ltx]
+%
+\begingroup
+\catcode`\ =\active
+\obeylines %
+% ignore everything up to the first ^^M, that's the newline at the end
+% of the @verbatim input line itself.  Otherwise we get an extra blank
+% line in the output.
+\xdef\doverbatim#1^^M#2@end verbatim{#2\noexpand\end\gobble verbatim}%
+% We really want {...\end verbatim} in the body of the macro, but
+% without the active space; thus we have to use \xdef and \gobble.
+\endgroup
+%
+\envdef\verbatim{%
+\setupverbatim\doverbatim
+}
+\let\Everbatim = \afterenvbreak
+% @verbatiminclude FILE - insert text of file in verbatim environment.
+%
+\def\verbatiminclude{\parseargusing\filenamecatcodes\doverbatiminclude}
+%
+\def\doverbatiminclude#1{%
+{%
+\makevalueexpandable
+\setupverbatim
+\input #1
+\afterenvbreak
+}%
+}
+% @copying ... @end copying.
+% Save the text away for @insertcopying later.
+%
+% We save the uninterpreted tokens, rather than creating a box.
+% Saving the text in a box would be much easier, but then all the
+% typesetting commands (@smallbook, font changes, etc.) have to be done
+% beforehand -- and a) we want @copying to be done first in the source
+% file; b) letting users define the frontmatter in as flexible order as
+% possible is very desirable.
+%
+\def\copying{\checkenv{}\begingroup\scanargctxt\docopying}
+\def\docopying#1@end copying{\endgroup\def\copyingtext{#1}}
+%
+\def\insertcopying{%
+\begingroup
+\parindent = 0pt  % paragraph indentation looks wrong on title page
+\scanexp\copyingtext
+\endgroup
 }
 \message{defuns,}
 % @defun etc.
-% Allow user to change definition object font (\df) internally
-\def\setdeffont #1 {\csname DEF#1\endcsname}
 \newskip\defbodyindent \defbodyindent=.4in
 \newskip\defargsindent \defargsindent=50pt
-\newskip\deftypemargin \deftypemargin=12pt
 \newskip\deflastargmargin \deflastargmargin=18pt
+\newcount\defunpenalty
-\newcount\parencount
-% define \functionparens, which makes ( and ) and & do special things.
+% Start the processing of @deffn:
-% \functionparens affects the group it is contained in.
+\def\startdefun{%
+\ifnum\lastpenalty<10000
+\medbreak
+\defunpenalty=10003 % Will keep this @deffn together with the
+% following @def command, see below.
+\else
+% If there are two @def commands in a row, we'll have a \nobreak,
+% which is there to keep the function description together with its
+% header.  But if there's nothing but headers, we need to allow a
+% break somewhere.  Check specifically for penalty 10002, inserted
+% by \printdefunline, instead of 10000, since the sectioning
+% commands also insert a nobreak penalty, and we don't want to allow
+% a break between a section heading and a defun.
+%
+% As a minor refinement, we avoid "club" headers by signalling
+% with penalty of 10003 after the very first @deffn in the
+% sequence (see above), and penalty of 10002 after any following
+% @def command.
+\ifnum\lastpenalty=10002 \penalty2000 \else \defunpenalty=10002 \fi
+%
+% Similarly, after a section heading, do not allow a break.
+% But do insert the glue.
+\medskip  % preceded by discardable penalty, so not a breakpoint
+\fi
+%
+\parindent=0in
+\advance\leftskip by \defbodyindent
+\exdentamount=\defbodyindent
+}
+\def\dodefunx#1{%
+% First, check whether we are in the right environment:
+\checkenv#1%
+%
+% As above, allow line break if we have multiple x headers in a row.
+% It's not a great place, though.
+\ifnum\lastpenalty=10002 \penalty3000 \else \defunpenalty=10002 \fi
+%
+% And now, it's time to reuse the body of the original defun:
+\expandafter\gobbledefun#1%
+}
+\def\gobbledefun#1\startdefun{}
+% \printdefunline \deffnheader{text}
+%
+\def\printdefunline#1#2{%
+\begingroup
+% call \deffnheader:
+#1#2 \endheader
+% common ending:
+\interlinepenalty = 10000
+\advance\rightskip by 0pt plus 1fil
+\endgraf
+\nobreak\vskip -\parskip
+\penalty\defunpenalty  % signal to \startdefun and \dodefunx
+% Some of the @defun-type tags do not enable magic parentheses,
+% rendering the following check redundant.  But we don't optimize.
+\checkparencounts
+\endgroup
+}
+\def\Edefun{\endgraf\medbreak}
+% \makedefun{deffn} creates \deffn, \deffnx and \Edeffn;
+% the only thing remaining is to define \deffnheader.
+%
+\def\makedefun#1{%
+\expandafter\let\csname E#1\endcsname = \Edefun
+\edef\temp{\noexpand\domakedefun
+\makecsname{#1}\makecsname{#1x}\makecsname{#1header}}%
+\temp
+}
+% \domakedefun \deffn \deffnx \deffnheader
+%
+% Define \deffn and \deffnx, without parameters.
+% \deffnheader has to be defined explicitly.
+%
+\def\domakedefun#1#2#3{%
+\envdef#1{%
+\startdefun
+\parseargusing\activeparens{\printdefunline#3}%
+}%
+\def#2{\dodefunx#1}%
+\def#3%
+}
+%%% Untyped functions:
+% @deffn category name args
+\makedefun{deffn}{\deffngeneral{}}
+% @deffn category class name args
+\makedefun{defop}#1 {\defopon{#1\ \putwordon}}
+% \defopon {category on}class name args
+\def\defopon#1#2 {\deffngeneral{\putwordon\ \code{#2}}{#1\ \code{#2}} }
+% \deffngeneral {subind}category name args
+%
+\def\deffngeneral#1#2 #3 #4\endheader{%
+% Remember that \dosubind{fn}{foo}{} is equivalent to \doind{fn}{foo}.
+\dosubind{fn}{\code{#3}}{#1}%
+\defname{#2}{}{#3}\magicamp\defunargs{#4\unskip}%
+}
+%%% Typed functions:
+% @deftypefn category type name args
+\makedefun{deftypefn}{\deftypefngeneral{}}
+% @deftypeop category class type name args
+\makedefun{deftypeop}#1 {\deftypeopon{#1\ \putwordon}}
+% \deftypeopon {category on}class type name args
+\def\deftypeopon#1#2 {\deftypefngeneral{\putwordon\ \code{#2}}{#1\ \code{#2}} }
+% \deftypefngeneral {subind}category type name args
+%
+\def\deftypefngeneral#1#2 #3 #4 #5\endheader{%
+\dosubind{fn}{\code{#4}}{#1}%
+\defname{#2}{#3}{#4}\defunargs{#5\unskip}%
+}
+%%% Typed variables:
+% @deftypevr category type var args
+\makedefun{deftypevr}{\deftypecvgeneral{}}
+% @deftypecv category class type var args
+\makedefun{deftypecv}#1 {\deftypecvof{#1\ \putwordof}}
+% \deftypecvof {category of}class type var args
+\def\deftypecvof#1#2 {\deftypecvgeneral{\putwordof\ \code{#2}}{#1\ \code{#2}} }
+% \deftypecvgeneral {subind}category type var args
+%
+\def\deftypecvgeneral#1#2 #3 #4 #5\endheader{%
+\dosubind{vr}{\code{#4}}{#1}%
+\defname{#2}{#3}{#4}\defunargs{#5\unskip}%
+}
+%%% Untyped variables:
+% @defvr category var args
+\makedefun{defvr}#1 {\deftypevrheader{#1} {} }
+% @defcv category class var args
+\makedefun{defcv}#1 {\defcvof{#1\ \putwordof}}
+% \defcvof {category of}class var args
+\def\defcvof#1#2 {\deftypecvof{#1}#2 {} }
+%%% Type:
+% @deftp category name args
+\makedefun{deftp}#1 #2 #3\endheader{%
+\doind{tp}{\code{#2}}%
+\defname{#1}{}{#2}\defunargs{#3\unskip}%
+}
+% Remaining @defun-like shortcuts:
+\makedefun{defun}{\deffnheader{\putwordDeffunc} }
+\makedefun{defmac}{\deffnheader{\putwordDefmac} }
+\makedefun{defspec}{\deffnheader{\putwordDefspec} }
+\makedefun{deftypefun}{\deftypefnheader{\putwordDeffunc} }
+\makedefun{defvar}{\defvrheader{\putwordDefvar} }
+\makedefun{defopt}{\defvrheader{\putwordDefopt} }
+\makedefun{deftypevar}{\deftypevrheader{\putwordDefvar} }
+\makedefun{defmethod}{\defopon\putwordMethodon}
+\makedefun{deftypemethod}{\deftypeopon\putwordMethodon}
+\makedefun{defivar}{\defcvof\putwordInstanceVariableof}
+\makedefun{deftypeivar}{\deftypecvof\putwordInstanceVariableof}
+% \defname, which formats the name of the @def (not the args).
+% #1 is the category, such as "Function".
+% #2 is the return type, if any.
+% #3 is the function name.
+%
+% We are followed by (but not passed) the arguments, if any.
+%
+\def\defname#1#2#3{%
+% Get the values of \leftskip and \rightskip as they were outside the @def...
+\advance\leftskip by -\defbodyindent
+%
+% How we'll format the type name.  Putting it in brackets helps
+% distinguish it from the body text that may end up on the next line
+% just below it.
+\def\temp{#1}%
+\setbox0=\hbox{\kern\deflastargmargin \ifx\temp\empty\else [\rm\temp]\fi}
+%
+% Figure out line sizes for the paragraph shape.
+% The first line needs space for \box0; but if \rightskip is nonzero,
+% we need only space for the part of \box0 which exceeds it:
+\dimen0=\hsize  \advance\dimen0 by -\wd0  \advance\dimen0 by \rightskip
+% The continuations:
+\dimen2=\hsize  \advance\dimen2 by -\defargsindent
+% (plain.tex says that \dimen1 should be used only as global.)
+\parshape 2 0in \dimen0 \defargsindent \dimen2
+%
+% Put the type name to the right margin.
+\noindent
+\hbox to 0pt{%
+\hfil\box0 \kern-\hsize
+% \hsize has to be shortened this way:
+\kern\leftskip
+% Intentionally do not respect \rightskip, since we need the space.
+}%
+%
+% Allow all lines to be underfull without complaint:
+\tolerance=10000 \hbadness=10000
+\exdentamount=\defbodyindent
+{%
+% defun fonts. We use typewriter by default (used to be bold) because:
+% . we're printing identifiers, they should be in tt in principle.
+% . in languages with many accents, such as Czech or French, it's
+%   common to leave accents off identifiers.  The result looks ok in
+%   tt, but exceedingly strange in rm.
+% . we don't want -- and --- to be treated as ligatures.
+% . this still does not fix the ?` and !` ligatures, but so far no
+%   one has made identifiers using them :).
+\df \tt
+\def\temp{#2}% return value type
+\ifx\temp\empty\else \tclose{\temp} \fi
+#3% output function name
+}%
+{\rm\enskip}% hskip 0.5 em of \tenrm
+%
+\boldbrax
+% arguments will be output next, if any.
+}
+% Print arguments in slanted roman (not ttsl), inconsistently with using
+% tt for the name.  This is because literal text is sometimes needed in
+% the argument list (groff manual), and ttsl and tt are not very
+% distinguishable.  Prevent hyphenation at `-' chars.
+%
+\def\defunargs#1{%
+% use sl by default (not ttsl),
+% tt for the names.
+\df \sl \hyphenchar\font=0
+%
+% On the other hand, if an argument has two dashes (for instance), we
+% want a way to get ttsl.  Let's try @var for that.
+\let\var=\ttslanted
+#1%
+\sl\hyphenchar\font=45
+}
+% We want ()&[] to print specially on the defun line.
+%
 \def\activeparens{%
-\catcode`\(=\active \catcode`\)=\active \catcode`\&=\active
+\catcode`\(=\active \catcode`\)=\active
-\catcode`\[=\active \catcode`\]=\active}
+\catcode`\[=\active \catcode`\]=\active
+\catcode`\&=\active
+}
 % Make control sequences which act like normal parenthesis chars.
 \let\lparen = ( \let\rparen = )
-{\activeparens % Now, smart parens don't turn on until &foo (see \amprm)
 % Be sure that we always have a definition for `(', etc.  For example,
 % if the fn name has parens in it, \boldbrax will not be in effect yet,
 % so TeX would otherwise complain about undefined control sequence.
-\global\let(=\lparen \global\let)=\rparen
+{
-\global\let[=\lbrack \global\let]=\rbrack
+\activeparens
+\global\let(=\lparen \global\let)=\rparen
-\gdef\functionparens{\boldbrax\let&=\amprm\parencount=0 }
+\global\let[=\lbrack \global\let]=\rbrack
-\gdef\boldbrax{\let(=\opnr\let)=\clnr\let[=\lbrb\let]=\rbrb}
+\global\let& = \&
-% This is used to turn on special parens
-% but make & act ordinary (given that it's active).
+\gdef\boldbrax{\let(=\opnr\let)=\clnr\let[=\lbrb\let]=\rbrb}
-\gdef\boldbraxnoamp{\let(=\opnr\let)=\clnr\let[=\lbrb\let]=\rbrb\let&=\ampnr}
+\gdef\magicamp{\let&=\amprm}
+}
-% Definitions of (, ) and & used in args for functions.
-% This is the definition of ( outside of all parentheses.
+\newcount\parencount
-\gdef\oprm#1 {{\rm\char`\(}#1 \bf \let(=\opnested
+% If we encounter &foo, then turn on ()-hacking afterwards
+\newif\ifampseen
+\def\amprm#1 {\ampseentrue{\bf\&#1 }}
+\def\parenfont{%
+\ifampseen
+% At the first level, print parens in roman,
+% otherwise use the default font.
+\ifnum \parencount=1 \rm \fi
+\else
+% The \sf parens (in \boldbrax) actually are a little bolder than
+% the contained text.  This is especially needed for [ and ] .
+\sf
+\fi
+}
+\def\infirstlevel#1{%
+\ifampseen
+\ifnum\parencount=1
+#1%
+\fi
+\fi
+}
+\def\bfafterword#1 {#1 \bf}
+\def\opnr{%
 \global\advance\parencount by 1
-}
+{\parenfont(}%
-%
+\infirstlevel \bfafterword
-% This is the definition of ( when already inside a level of parens.
+}
-\gdef\opnested{\char`\(\global\advance\parencount by 1 }
+\def\clnr{%
-%
+{\parenfont)}%
-\gdef\clrm{% Print a paren in roman if it is taking us back to depth of 0.
+\infirstlevel \sl
-% also in that case restore the outer-level definition of (.
+\global\advance\parencount by -1
-\ifnum \parencount=1 {\rm \char `\)}\sl \let(=\oprm \else \char `\) \fi
+}
-\global\advance \parencount by -1 }
-% If we encounter &foo, then turn on ()-hacking afterwards
+\newcount\brackcount
-\gdef\amprm#1 {{\rm\&#1}\let(=\oprm \let)=\clrm\ }
+\def\lbrb{%
-%
+\global\advance\brackcount by 1
-\gdef\normalparens{\boldbrax\let&=\ampnr}
+{\bf[}%
-} % End of definition inside \activeparens
+}
-%% These parens (in \boldbrax) actually are a little bolder than the
+\def\rbrb{%
-%% contained text.  This is especially needed for [ and ]
+{\bf]}%
-\def\opnr{{\sf\char`\(}\global\advance\parencount by 1 }
+\global\advance\brackcount by -1
-\def\clnr{{\sf\char`\)}\global\advance\parencount by -1 }
+}
-\let\ampnr = \&
-\def\lbrb{{\bf\char`\[}}
+\def\checkparencounts{%
-\def\rbrb{{\bf\char`\]}}
+\ifnum\parencount=0 \else \badparencount \fi
+\ifnum\brackcount=0 \else \badbrackcount \fi
-% Active &'s sneak into the index arguments, so make sure it's defined.
+}
-{
+% these should not use \errmessage; the glibc manual, at least, actually
-\catcode`& = 13
+% has such constructs (when documenting function pointers).
-\global\let& = \ampnr
+\def\badparencount{%
-}
+\message{Warning: unbalanced parentheses in @def...}%
+\global\parencount=0
-% First, defname, which formats the header line itself.
+}
-% #1 should be the function name.
+\def\badbrackcount{%
-% #2 should be the type of definition, such as "Function".
+\message{Warning: unbalanced square brackets in @def...}%
+\global\brackcount=0
-\def\defname #1#2{%
+}
-% Get the values of \leftskip and \rightskip as they were
-% outside the @def...
-\dimen2=\leftskip
-\advance\dimen2 by -\defbodyindent
-\noindent
-\setbox0=\hbox{\hskip \deflastargmargin{\rm #2}\hskip \deftypemargin}%
-\dimen0=\hsize \advance \dimen0 by -\wd0 % compute size for first line
-\dimen1=\hsize \advance \dimen1 by -\defargsindent %size for continuations
-\parshape 2 0in \dimen0 \defargsindent \dimen1
-% Now output arg 2 ("Function" or some such)
-% ending at \deftypemargin from the right margin,
-% but stuck inside a box of width 0 so it does not interfere with linebreaking
-{% Adjust \hsize to exclude the ambient margins,
-% so that \rightline will obey them.
-\advance \hsize by -\dimen2
-\rlap{\rightline{{\rm #2}\hskip -1.25pc }}}%
-% Make all lines underfull and no complaints:
-\tolerance=10000 \hbadness=10000
-\advance\leftskip by -\defbodyindent
-\exdentamount=\defbodyindent
-{\df #1}\enskip        % Generate function name
-}
-% Actually process the body of a definition
-% #1 should be the terminating control sequence, such as \Edefun.
-% #2 should be the "another name" control sequence, such as \defunx.
-% #3 should be the control sequence that actually processes the header,
-%    such as \defunheader.
-\def\defparsebody #1#2#3{\begingroup\inENV% Environment for definitionbody
-\medbreak %
-% Define the end token that this defining construct specifies
-% so that it will exit this group.
-\def#1{\endgraf\endgroup\medbreak}%
-\def#2{\begingroup\obeylines\activeparens\spacesplit#3}%
-\parindent=0in
-\advance\leftskip by \defbodyindent
-\exdentamount=\defbodyindent
-\begingroup %
-\catcode 61=\active % 61 is `='
-\obeylines\activeparens\spacesplit#3}
-% #1 is the \E... control sequence to end the definition (which we define).
-% #2 is the \...x control sequence for consecutive fns (which we define).
-% #3 is the control sequence to call to resume processing.
-% #4, delimited by the space, is the class name.
-%
-\def\defmethparsebody#1#2#3#4 {\begingroup\inENV %
-\medbreak %
-% Define the end token that this defining construct specifies
-% so that it will exit this group.
-\def#1{\endgraf\endgroup\medbreak}%
-\def#2##1 {\begingroup\obeylines\activeparens\spacesplit{#3{##1}}}%
-\parindent=0in
-\advance\leftskip by \defbodyindent
-\exdentamount=\defbodyindent
-\begingroup\obeylines\activeparens\spacesplit{#3{#4}}}
-% Used for @deftypemethod and @deftypeivar.
-% #1 is the \E... control sequence to end the definition (which we define).
-% #2 is the \...x control sequence for consecutive fns (which we define).
-% #3 is the control sequence to call to resume processing.
-% #4, delimited by a space, is the class name.
-% #5 is the method's return type.
-%
-\def\deftypemethparsebody#1#2#3#4 #5 {\begingroup\inENV
-\medbreak
-\def#1{\endgraf\endgroup\medbreak}%
-\def#2##1 ##2 {\begingroup\obeylines\activeparens\spacesplit{#3{##1}{##2}}}%
-\parindent=0in
-\advance\leftskip by \defbodyindent
-\exdentamount=\defbodyindent
-\begingroup\obeylines\activeparens\spacesplit{#3{#4}{#5}}}
-% Used for @deftypeop.  The change from \deftypemethparsebody is an
-% extra argument at the beginning which is the `category', instead of it
-% being the hardwired string `Method' or `Instance Variable'.  We have
-% to account for this both in the \...x definition and in parsing the
-% input at hand.  Thus also need a control sequence (passed as #5) for
-% the \E... definition to assign the category name to.
-%
-\def\deftypeopparsebody#1#2#3#4#5 #6 {\begingroup\inENV
-\medbreak
-\def#1{\endgraf\endgroup\medbreak}%
-\def#2##1 ##2 ##3 {%
-\def#4{##1}%
-\begingroup\obeylines\activeparens\spacesplit{#3{##2}{##3}}}%
-\parindent=0in
-\advance\leftskip by \defbodyindent
-\exdentamount=\defbodyindent
-\begingroup\obeylines\activeparens\spacesplit{#3{#5}{#6}}}
-\def\defopparsebody #1#2#3#4#5 {\begingroup\inENV %
-\medbreak %
-% Define the end token that this defining construct specifies
-% so that it will exit this group.
-\def#1{\endgraf\endgroup\medbreak}%
-\def#2##1 ##2 {\def#4{##1}%
-\begingroup\obeylines\activeparens\spacesplit{#3{##2}}}%
-\parindent=0in
-\advance\leftskip by \defbodyindent
-\exdentamount=\defbodyindent
-\begingroup\obeylines\activeparens\spacesplit{#3{#5}}}
-% These parsing functions are similar to the preceding ones
-% except that they do not make parens into active characters.
-% These are used for "variables" since they have no arguments.
-\def\defvarparsebody #1#2#3{\begingroup\inENV% Environment for definitionbody
-\medbreak %
-% Define the end token that this defining construct specifies
-% so that it will exit this group.
-\def#1{\endgraf\endgroup\medbreak}%
-\def#2{\begingroup\obeylines\spacesplit#3}%
-\parindent=0in
-\advance\leftskip by \defbodyindent
-\exdentamount=\defbodyindent
-\begingroup %
-\catcode 61=\active %
-\obeylines\spacesplit#3}
-% This is used for \def{tp,vr}parsebody.  It could probably be used for
-% some of the others, too, with some judicious conditionals.
-%
-\def\parsebodycommon#1#2#3{%
-\begingroup\inENV %
-\medbreak %
-% Define the end token that this defining construct specifies
-% so that it will exit this group.
-\def#1{\endgraf\endgroup\medbreak}%
-\def#2##1 {\begingroup\obeylines\spacesplit{#3{##1}}}%
-\parindent=0in
-\advance\leftskip by \defbodyindent
-\exdentamount=\defbodyindent
-\begingroup\obeylines
-}
-\def\defvrparsebody#1#2#3#4 {%
-\parsebodycommon{#1}{#2}{#3}%
-\spacesplit{#3{#4}}%
-}
-% This loses on `@deftp {Data Type} {struct termios}' -- it thinks the
-% type is just `struct', because we lose the braces in `{struct
-% termios}' when \spacesplit reads its undelimited argument.  Sigh.
-% \let\deftpparsebody=\defvrparsebody
-%
-% So, to get around this, we put \empty in with the type name.  That
-% way, TeX won't find exactly `{...}' as an undelimited argument, and
-% won't strip off the braces.
-%
-\def\deftpparsebody #1#2#3#4 {%
-\parsebodycommon{#1}{#2}{#3}%
-\spacesplit{\parsetpheaderline{#3{#4}}}\empty
-}
-% Fine, but then we have to eventually remove the \empty *and* the
-% braces (if any).  That's what this does.
-%
-\def\removeemptybraces\empty#1\relax{#1}
-% After \spacesplit has done its work, this is called -- #1 is the final
-% thing to call, #2 the type name (which starts with \empty), and #3
-% (which might be empty) the arguments.
-%
-\def\parsetpheaderline#1#2#3{%
-#1{\removeemptybraces#2\relax}{#3}%
-}%
-\def\defopvarparsebody #1#2#3#4#5 {\begingroup\inENV %
-\medbreak %
-% Define the end token that this defining construct specifies
-% so that it will exit this group.
-\def#1{\endgraf\endgroup\medbreak}%
-\def#2##1 ##2 {\def#4{##1}%
-\begingroup\obeylines\spacesplit{#3{##2}}}%
-\parindent=0in
-\advance\leftskip by \defbodyindent
-\exdentamount=\defbodyindent
-\begingroup\obeylines\spacesplit{#3{#5}}}
-% Split up #2 at the first space token.
-% call #1 with two arguments:
-%  the first is all of #2 before the space token,
-%  the second is all of #2 after that space token.
-% If #2 contains no space token, all of it is passed as the first arg
-% and the second is passed as empty.
-{\obeylines
-\gdef\spacesplit#1#2^^M{\endgroup\spacesplitfoo{#1}#2 \relax\spacesplitfoo}%
-\long\gdef\spacesplitfoo#1#2 #3#4\spacesplitfoo{%
-\ifx\relax #3%
-#1{#2}{}\else #1{#2}{#3#4}\fi}}
-% So much for the things common to all kinds of definitions.
-% Define @defun.
-% First, define the processing that is wanted for arguments of \defun
-% Use this to expand the args and terminate the paragraph they make up
-\def\defunargs#1{\functionparens \sl
-% Expand, preventing hyphenation at `-' chars.
-% Note that groups don't affect changes in \hyphenchar.
-% Set the font temporarily and use \font in case \setfont made \tensl a macro.
-{\tensl\hyphenchar\font=0}%
-#1%
-{\tensl\hyphenchar\font=45}%
-\ifnum\parencount=0 \else \errmessage{Unbalanced parentheses in @def}\fi%
-\interlinepenalty=10000
-\advance\rightskip by 0pt plus 1fil
-\endgraf\nobreak\vskip -\parskip\nobreak
-}
-\def\deftypefunargs #1{%
-% Expand, preventing hyphenation at `-' chars.
-% Note that groups don't affect changes in \hyphenchar.
-% Use \boldbraxnoamp, not \functionparens, so that & is not special.
-\boldbraxnoamp
-\tclose{#1}% avoid \code because of side effects on active chars
-\interlinepenalty=10000
-\advance\rightskip by 0pt plus 1fil
-\endgraf\nobreak\vskip -\parskip\nobreak
-}
-% Do complete processing of one @defun or @defunx line already parsed.
-% @deffn Command forward-char nchars
-\def\deffn{\defmethparsebody\Edeffn\deffnx\deffnheader}
-\def\deffnheader #1#2#3{\doind {fn}{\code{#2}}%
-\begingroup\defname {#2}{#1}\defunargs{#3}\endgroup %
-\catcode 61=\other % Turn off change made in \defparsebody
-}
-% @defun == @deffn Function
-\def\defun{\defparsebody\Edefun\defunx\defunheader}
-\def\defunheader #1#2{\doind {fn}{\code{#1}}% Make entry in function index
-\begingroup\defname {#1}{\putwordDeffunc}%
-\defunargs {#2}\endgroup %
-\catcode 61=\other % Turn off change made in \defparsebody
-}
-% @deftypefun int foobar (int @var{foo}, float @var{bar})
-\def\deftypefun{\defparsebody\Edeftypefun\deftypefunx\deftypefunheader}
-% #1 is the data type.  #2 is the name and args.
-\def\deftypefunheader #1#2{\deftypefunheaderx{#1}#2 \relax}
-% #1 is the data type, #2 the name, #3 the args.
-\def\deftypefunheaderx #1#2 #3\relax{%
-\doind {fn}{\code{#2}}% Make entry in function index
-\begingroup\defname {\defheaderxcond#1\relax$$$#2}{\putwordDeftypefun}%
-\deftypefunargs {#3}\endgroup %
-\catcode 61=\other % Turn off change made in \defparsebody
-}
-% @deftypefn {Library Function} int foobar (int @var{foo}, float @var{bar})
-\def\deftypefn{\defmethparsebody\Edeftypefn\deftypefnx\deftypefnheader}
-% \defheaderxcond#1\relax$$$
-% puts #1 in @code, followed by a space, but does nothing if #1 is null.
-\def\defheaderxcond#1#2$$${\ifx#1\relax\else\code{#1#2} \fi}
-% #1 is the classification.  #2 is the data type.  #3 is the name and args.
-\def\deftypefnheader #1#2#3{\deftypefnheaderx{#1}{#2}#3 \relax}
-% #1 is the classification, #2 the data type, #3 the name, #4 the args.
-\def\deftypefnheaderx #1#2#3 #4\relax{%
-\doind {fn}{\code{#3}}% Make entry in function index
-\begingroup
-\normalparens % notably, turn off `&' magic, which prevents
-%               at least some C++ text from working
-\defname {\defheaderxcond#2\relax$$$#3}{#1}%
-\deftypefunargs {#4}\endgroup %
-\catcode 61=\other % Turn off change made in \defparsebody
-}
-% @defmac == @deffn Macro
-\def\defmac{\defparsebody\Edefmac\defmacx\defmacheader}
-\def\defmacheader #1#2{\doind {fn}{\code{#1}}% Make entry in function index
-\begingroup\defname {#1}{\putwordDefmac}%
-\defunargs {#2}\endgroup %
-\catcode 61=\other % Turn off change made in \defparsebody
-}
-% @defspec == @deffn Special Form
-\def\defspec{\defparsebody\Edefspec\defspecx\defspecheader}
-\def\defspecheader #1#2{\doind {fn}{\code{#1}}% Make entry in function index
-\begingroup\defname {#1}{\putwordDefspec}%
-\defunargs {#2}\endgroup %
-\catcode 61=\other % Turn off change made in \defparsebody
-}
-% @defop CATEGORY CLASS OPERATION ARG...
-%
-\def\defop #1 {\def\defoptype{#1}%
-\defopparsebody\Edefop\defopx\defopheader\defoptype}
-%
-\def\defopheader#1#2#3{%
-\dosubind {fn}{\code{#2}}{\putwordon\ #1}% Make entry in function index
-\begingroup\defname {#2}{\defoptype\ \putwordon\ #1}%
-\defunargs {#3}\endgroup %
-}
-% @deftypeop CATEGORY CLASS TYPE OPERATION ARG...
-%
-\def\deftypeop #1 {\def\deftypeopcategory{#1}%
-\deftypeopparsebody\Edeftypeop\deftypeopx\deftypeopheader
-\deftypeopcategory}
-%
-% #1 is the class name, #2 the data type, #3 the operation name, #4 the args.
-\def\deftypeopheader#1#2#3#4{%
-\dosubind{fn}{\code{#3}}{\putwordon\ \code{#1}}% entry in function index
-\begingroup
-\defname{\defheaderxcond#2\relax$$$#3}
-{\deftypeopcategory\ \putwordon\ \code{#1}}%
-\deftypefunargs{#4}%
-\endgroup
-}
-% @deftypemethod CLASS TYPE METHOD ARG...
-%
-\def\deftypemethod{%
-\deftypemethparsebody\Edeftypemethod\deftypemethodx\deftypemethodheader}
-%
-% #1 is the class name, #2 the data type, #3 the method name, #4 the args.
-\def\deftypemethodheader#1#2#3#4{%
-\dosubind{fn}{\code{#3}}{\putwordon\ \code{#1}}% entry in function index
-\begingroup
-\defname{\defheaderxcond#2\relax$$$#3}{\putwordMethodon\ \code{#1}}%
-\deftypefunargs{#4}%
-\endgroup
-}
-% @deftypeivar CLASS TYPE VARNAME
-%
-\def\deftypeivar{%
-\deftypemethparsebody\Edeftypeivar\deftypeivarx\deftypeivarheader}
-%
-% #1 is the class name, #2 the data type, #3 the variable name.
-\def\deftypeivarheader#1#2#3{%
-\dosubind{vr}{\code{#3}}{\putwordof\ \code{#1}}% entry in variable index
-\begingroup
-\defname{#3}{\putwordInstanceVariableof\ \code{#1}}%
-\defvarargs{#3}%
-\endgroup
-}
-% @defmethod == @defop Method
-%
-\def\defmethod{\defmethparsebody\Edefmethod\defmethodx\defmethodheader}
-%
-% #1 is the class name, #2 the method name, #3 the args.
-\def\defmethodheader#1#2#3{%
-\dosubind{fn}{\code{#2}}{\putwordon\ \code{#1}}% entry in function index
-\begingroup
-\defname{#2}{\putwordMethodon\ \code{#1}}%
-\defunargs{#3}%
-\endgroup
-}
-% @defcv {Class Option} foo-class foo-flag
-\def\defcv #1 {\def\defcvtype{#1}%
-\defopvarparsebody\Edefcv\defcvx\defcvarheader\defcvtype}
-\def\defcvarheader #1#2#3{%
-\dosubind {vr}{\code{#2}}{\putwordof\ #1}% Make entry in var index
-\begingroup\defname {#2}{\defcvtype\ \putwordof\ #1}%
-\defvarargs {#3}\endgroup %
-}
-% @defivar CLASS VARNAME == @defcv {Instance Variable} CLASS VARNAME
-%
-\def\defivar{\defvrparsebody\Edefivar\defivarx\defivarheader}
-%
-\def\defivarheader#1#2#3{%
-\dosubind {vr}{\code{#2}}{\putwordof\ #1}% entry in var index
-\begingroup
-\defname{#2}{\putwordInstanceVariableof\ #1}%
-\defvarargs{#3}%
-\endgroup
-}
-% @defvar
-% First, define the processing that is wanted for arguments of @defvar.
-% This is actually simple: just print them in roman.
-% This must expand the args and terminate the paragraph they make up
-\def\defvarargs #1{\normalparens #1%
-\interlinepenalty=10000
-\endgraf\nobreak\vskip -\parskip\nobreak}
-% @defvr Counter foo-count
-\def\defvr{\defvrparsebody\Edefvr\defvrx\defvrheader}
-\def\defvrheader #1#2#3{\doind {vr}{\code{#2}}%
-\begingroup\defname {#2}{#1}\defvarargs{#3}\endgroup}
-% @defvar == @defvr Variable
-\def\defvar{\defvarparsebody\Edefvar\defvarx\defvarheader}
-\def\defvarheader #1#2{\doind {vr}{\code{#1}}% Make entry in var index
-\begingroup\defname {#1}{\putwordDefvar}%
-\defvarargs {#2}\endgroup %
-}
-% @defopt == @defvr {User Option}
-\def\defopt{\defvarparsebody\Edefopt\defoptx\defoptheader}
-\def\defoptheader #1#2{\doind {vr}{\code{#1}}% Make entry in var index
-\begingroup\defname {#1}{\putwordDefopt}%
-\defvarargs {#2}\endgroup %
-}
-% @deftypevar int foobar
-\def\deftypevar{\defvarparsebody\Edeftypevar\deftypevarx\deftypevarheader}
-% #1 is the data type.  #2 is the name, perhaps followed by text that
-% is actually part of the data type, which should not be put into the index.
-\def\deftypevarheader #1#2{%
-\dovarind#2 \relax% Make entry in variables index
-\begingroup\defname {\defheaderxcond#1\relax$$$#2}{\putwordDeftypevar}%
-\interlinepenalty=10000
-\endgraf\nobreak\vskip -\parskip\nobreak
-\endgroup}
-\def\dovarind#1 #2\relax{\doind{vr}{\code{#1}}}
-% @deftypevr {Global Flag} int enable
-\def\deftypevr{\defvrparsebody\Edeftypevr\deftypevrx\deftypevrheader}
-\def\deftypevrheader #1#2#3{\dovarind#3 \relax%
-\begingroup\defname {\defheaderxcond#2\relax$$$#3}{#1}
-\interlinepenalty=10000
-\endgraf\nobreak\vskip -\parskip\nobreak
-\endgroup}
-% Now define @deftp
-% Args are printed in bold, a slight difference from @defvar.
-\def\deftpargs #1{\bf \defvarargs{#1}}
-% @deftp Class window height width ...
-\def\deftp{\deftpparsebody\Edeftp\deftpx\deftpheader}
-\def\deftpheader #1#2#3{\doind {tp}{\code{#2}}%
-\begingroup\defname {#2}{#1}\deftpargs{#3}\endgroup}
-% These definitions are used if you use @defunx (etc.)
-% anywhere other than immediately after a @defun or @defunx.
-%
-\def\defcvx#1 {\errmessage{@defcvx in invalid context}}
-\def\deffnx#1 {\errmessage{@deffnx in invalid context}}
-\def\defivarx#1 {\errmessage{@defivarx in invalid context}}
-\def\defmacx#1 {\errmessage{@defmacx in invalid context}}
-\def\defmethodx#1 {\errmessage{@defmethodx in invalid context}}
-\def\defoptx #1 {\errmessage{@defoptx in invalid context}}
-\def\defopx#1 {\errmessage{@defopx in invalid context}}
-\def\defspecx#1 {\errmessage{@defspecx in invalid context}}
-\def\deftpx#1 {\errmessage{@deftpx in invalid context}}
-\def\deftypefnx#1 {\errmessage{@deftypefnx in invalid context}}
-\def\deftypefunx#1 {\errmessage{@deftypefunx in invalid context}}
-\def\deftypeivarx#1 {\errmessage{@deftypeivarx in invalid context}}
-\def\deftypemethodx#1 {\errmessage{@deftypemethodx in invalid context}}
-\def\deftypeopx#1 {\errmessage{@deftypeopx in invalid context}}
-\def\deftypevarx#1 {\errmessage{@deftypevarx in invalid context}}
-\def\deftypevrx#1 {\errmessage{@deftypevrx in invalid context}}
-\def\defunx#1 {\errmessage{@defunx in invalid context}}
-\def\defvarx#1 {\errmessage{@defvarx in invalid context}}
-\def\defvrx#1 {\errmessage{@defvrx in invalid context}}
 \message{macros,}
 % @macro.
 % To do this right we need a feature of e-TeX, \scantokens,
 % which we arrange to emulate with a temporary file in ordinary TeX.
 \ifx\eTeXversion\undefined
 \newwrite\macscribble
-\def\scanmacro#1{%
+\def\scantokens#1{%
-\begingroup \newlinechar`\^^M
+\toks0={#1}%
-% Undo catcode changes of \startcontents and \doprintindex
+\immediate\openout\macscribble=\jobname.tmp
-\catcode`\@=0 \catcode`\\=12 \escapechar=`\@
+\immediate\write\macscribble{\the\toks0}%
-% Append \endinput to make sure that TeX does not see the ending newline.
+\immediate\closeout\macscribble
-\toks0={#1\endinput}%
+\input \jobname.tmp
-\immediate\openout\macscribble=\jobname.tmp
+}
-\immediate\write\macscribble{\the\toks0}%
+\fi
-\immediate\closeout\macscribble
-\let\xeatspaces\eatspaces
-\input \jobname.tmp
-\endgroup
-}
-\else
 \def\scanmacro#1{%
-\begingroup \newlinechar`\^^M
+\begingroup
-% Undo catcode changes of \startcontents and \doprintindex
+\newlinechar`\^^M
-\catcode`\@=0 \catcode`\\=12 \escapechar=`\@
+\let\xeatspaces\eatspaces
-\let\xeatspaces\eatspaces\scantokens{#1\endinput}\endgroup}
+% Undo catcode changes of \startcontents and \doprintindex
-\fi
+% When called from @insertcopying or (short)caption, we need active
+% backslash to get it printed correctly.  Previously, we had
+% \catcode`\\=\other instead.  We'll see whether a problem appears
+% with macro expansion.				--kasal, 19aug04
+\catcode`\@=0 \catcode`\\=\active \escapechar=`\@
+% ... and \example
+\spaceisspace
+%
+% Append \endinput to make sure that TeX does not see the ending newline.
+% I've verified that it is necessary both for e-TeX and for ordinary TeX
+%							--kasal, 29nov03
+\scantokens{#1\endinput}%
+\endgroup
+}
+\def\scanexp#1{%
+\edef\temp{\noexpand\scanmacro{#1}}%
+\temp
+}
 \newcount\paramno   % Count of parameters
 \newtoks\macname    % Macro name
 \newif\ifrecursive  % Is it recursive?
-\def\macrolist{}    % List of all defined macros in the form
-% \do\macro1\do\macro2...
+% List of all defined macros in the form
+%    \definedummyword\macro1\definedummyword\macro2...
+% Currently is also contains all @aliases; the list can be split
+% if there is a need.
+\def\macrolist{}
+% Add the macro to \macrolist
+\def\addtomacrolist#1{\expandafter \addtomacrolistxxx \csname#1\endcsname}
+\def\addtomacrolistxxx#1{%
+\toks0 = \expandafter{\macrolist\definedummyword#1}%
+\xdef\macrolist{\the\toks0}%
+}
 % Utility routines.
-% Thisdoes \let #1 = #2, except with \csnames.
+% This does \let #1 = #2, with \csnames; that is,
+%   \let \csname#1\endcsname = \csname#2\endcsname
+% (except of course we have to play expansion games).
+%
 \def\cslet#1#2{%
-\expandafter\expandafter
+\expandafter\let
-\expandafter\let
+\csname#1\expandafter\endcsname
-\expandafter\expandafter
+\csname#2\endcsname
-\csname#1\endcsname
+}
-\csname#2\endcsname}
 % Trim leading and trailing spaces off a string.
 % Concepts from aro-bend problem 15 (see CTAN).
 {\catcode`\@=11
 \gdef\eatspaces #1{\expandafter\trim@\expandafter{#1 }}
 \def\unbrace#1{#1}
 \unbrace{\gdef\trim@@@ #1 } #2@{#1}
 }
 % Trim a single trailing ^^M off a string.
-{\catcode`\^^M=12\catcode`\Q=3%
+{\catcode`\^^M=\other \catcode`\Q=3%
 \gdef\eatcr #1{\eatcra #1Q^^MQ}%
 \gdef\eatcra#1^^MQ{\eatcrb#1Q}%
 \gdef\eatcrb#1Q#2Q{#1}%
 }
 % Macro bodies are absorbed as an argument in a context where
 % all characters are catcode 10, 11 or 12, except \ which is active
 % (as in normal texinfo). It is necessary to change the definition of \.
+% Non-ASCII encodings make 8-bit characters active, so un-activate
+% them to avoid their expansion.  Must do this non-globally, to
+% confine the change to the current group.
 % It's necessary to have hard CRs when the macro is executed. This is
 % done by  making ^^M (\endlinechar) catcode 12 when reading the macro
 % body, and then making it the \newlinechar in \scanmacro.
+\def\scanctxt{%
+\catcode`\"=\other
+\catcode`\+=\other
+\catcode`\<=\other
+\catcode`\>=\other
+\catcode`\@=\other
+\catcode`\^=\other
+\catcode`\_=\other
+\catcode`\|=\other
+\catcode`\~=\other
+\ifx\declaredencoding\ascii \else \setnonasciicharscatcodenonglobal\other \fi
+}
+\def\scanargctxt{%
+\scanctxt
+\catcode`\\=\other
+\catcode`\^^M=\other
+}
 \def\macrobodyctxt{%
-\catcode`\~=12
+\scanctxt
-\catcode`\^=12
+\catcode`\{=\other
-\catcode`\_=12
+\catcode`\}=\other
-\catcode`\|=12
+\catcode`\^^M=\other
-\catcode`\<=12
+\usembodybackslash
-\catcode`\>=12
+}
-\catcode`\+=12
-\catcode`\{=12
-\catcode`\}=12
-\catcode`\@=12
-\catcode`\^^M=12
-\usembodybackslash}
 \def\macroargctxt{%
-\catcode`\~=12
+\scanctxt
-\catcode`\^=12
+\catcode`\\=\other
-\catcode`\_=12
+}
-\catcode`\|=12
-\catcode`\<=12
-\catcode`\>=12
-\catcode`\+=12
-\catcode`\@=12
-\catcode`\\=12}
 % \mbodybackslash is the definition of \ in @macro bodies.
 % It maps \foo\ => \csname macarg.foo\endcsname => #N
 % where N is the macro parameter number.
 % We define \csname macarg.\endcsname to be \realbackslash, so
 \fi
 \if1\csname ismacro.\the\macname\endcsname
 \message{Warning: redefining \the\macname}%
 \else
 \expandafter\ifx\csname \the\macname\endcsname \relax
-\else \errmessage{The name \the\macname\space is reserved}\fi
+\else \errmessage{Macro name \the\macname\space already defined}\fi
 \global\cslet{macsave.\the\macname}{\the\macname}%
 \global\expandafter\let\csname ismacro.\the\macname\endcsname=1%
-% Add the macroname to \macrolist
+\addtomacrolist{\the\macname}%
-\toks0 = \expandafter{\macrolist\do}%
-\xdef\macrolist{\the\toks0
-\expandafter\noexpand\csname\the\macname\endcsname}%
 \fi
 \begingroup \macrobodyctxt
 \ifrecursive \expandafter\parsermacbody
 \else \expandafter\parsemacbody
 \fi}
-\def\unmacro{\parsearg\unmacroxxx}
+\parseargdef\unmacro{%
-\def\unmacroxxx#1{%
 \if1\csname ismacro.#1\endcsname
 \global\cslet{#1}{macsave.#1}%
 \global\expandafter\let \csname ismacro.#1\endcsname=0%
-% Remove the macro name from \macrolist
+% Remove the macro name from \macrolist:
 \begingroup
-\edef\tempa{\expandafter\noexpand\csname#1\endcsname}%
+\expandafter\let\csname#1\endcsname \relax
-\def\do##1{%
+\let\definedummyword\unmacrodo
-\def\tempb{##1}%
+\xdef\macrolist{\macrolist}%
-\ifx\tempa\tempb
-% remove this
-\else
-\toks0 = \expandafter{\newmacrolist\do}%
-\edef\newmacrolist{\the\toks0\expandafter\noexpand\tempa}%
-\fi}%
-\def\newmacrolist{}%
-% Execute macro list to define \newmacrolist
-\macrolist
-\global\let\macrolist\newmacrolist
 \endgroup
 \else
 \errmessage{Macro #1 not defined}%
+\fi
+}
+% Called by \do from \dounmacro on each macro.  The idea is to omit any
+% macro definitions that have been changed to \relax.
+%
+\def\unmacrodo#1{%
+\ifx #1\relax
+% remove this
+\else
+\noexpand\definedummyword \noexpand#1%
 \fi
 }
 % This makes use of the obscure feature that if the last token of a
 % <parameter list> is #, then the preceding argument is delimited by
 % \braceorline decides whether the next nonwhitespace character is a
 % {.  If so it reads up to the closing }, if not, it reads the whole
 % line.  Whatever was read is then fed to the next control sequence
 % as an argument (by \parsebrace or \parsearg)
-\def\braceorline#1{\let\next=#1\futurelet\nchar\braceorlinexxx}
+\def\braceorline#1{\let\macnamexxx=#1\futurelet\nchar\braceorlinexxx}
 \def\braceorlinexxx{%
 \ifx\nchar\bgroup\else
 \expandafter\parsearg
-\fi \next}
+\fi \macnamexxx}
-% We mant to disable all macros during \shipout so that they are not
-% expanded by \write.
-\def\turnoffmacros{\begingroup \def\do##1{\let\noexpand##1=\relax}%
-\edef\next{\macrolist}\expandafter\endgroup\next}
 % @alias.
 % We need some trickery to remove the optional spaces around the equal
 % sign.  Just make them active and then expand them all to nothing.
-\def\alias{\begingroup\obeyspaces\parsearg\aliasxxx}
+\def\alias{\parseargusing\obeyspaces\aliasxxx}
 \def\aliasxxx #1{\aliasyyy#1\relax}
-\def\aliasyyy #1=#2\relax{\ignoreactivespaces
+\def\aliasyyy #1=#2\relax{%
-\edef\next{\global\let\expandafter\noexpand\csname#1\endcsname=%
+{%
-\expandafter\noexpand\csname#2\endcsname}%
+\expandafter\let\obeyedspace=\empty
-\expandafter\endgroup\next}
+\addtomacrolist{#1}%
+\xdef\next{\global\let\makecsname{#1}=\makecsname{#2}}%
+}%
+\next
+}
 \message{cross references,}
-% @xref etc.
 \newwrite\auxfile
 \newif\ifhavexrefs    % True if xref values are known.
 \newif\ifwarnedxrefs  % True if we warned once that they aren't known.
 % @inforef is relatively simple.
 \def\inforef #1{\inforefzzz #1,,,,**}
 \def\inforefzzz #1,#2,#3,#4**{\putwordSee{} \putwordInfo{} \putwordfile{} \file{\ignorespaces #3{}},
 node \samp{\ignorespaces#1{}}}
-% @node's job is to define \lastnode.
+% @node's only job in TeX is to define \lastnode, which is used in
-\def\node{\ENVcheck\parsearg\nodezzz}
+% cross-references.  The @node line might or might not have commas, and
-\def\nodezzz#1{\nodexxx [#1,]}
+% might or might not have spaces before the first comma, like:
-\def\nodexxx[#1,#2]{\gdef\lastnode{#1}}
+% @node foo , bar , ...
+% We don't want such trailing spaces in the node name.
+%
+\parseargdef\node{\checkenv{}\donode #1 ,\finishnodeparse}
+%
+% also remove a trailing comma, in case of something like this:
+% @node Help-Cross,  ,  , Cross-refs
+\def\donode#1 ,#2\finishnodeparse{\dodonode #1,\finishnodeparse}
+\def\dodonode#1,#2\finishnodeparse{\gdef\lastnode{#1}}
 \let\nwnode=\node
-\let\lastnode=\relax
+\let\lastnode=\empty
-% The sectioning commands (@chapter, etc.) call these.
+% Write a cross-reference definition for the current node.  #1 is the
-\def\donoderef{%
+% type (Ynumbered, Yappendix, Ynothing).
-\ifx\lastnode\relax\else
+%
-\expandafter\expandafter\expandafter\setref{\lastnode}%
+\def\donoderef#1{%
-{Ysectionnumberandtype}%
+\ifx\lastnode\empty\else
-\global\let\lastnode=\relax
+\setref{\lastnode}{#1}%
+\global\let\lastnode=\empty
 \fi
 }
-\def\unnumbnoderef{%
-\ifx\lastnode\relax\else
+% @anchor{NAME} -- define xref target at arbitrary point.
-\expandafter\expandafter\expandafter\setref{\lastnode}{Ynothing}%
+%
-\global\let\lastnode=\relax
+\newcount\savesfregister
+%
+\def\savesf{\relax \ifhmode \savesfregister=\spacefactor \fi}
+\def\restoresf{\relax \ifhmode \spacefactor=\savesfregister \fi}
+\def\anchor#1{\savesf \setref{#1}{Ynothing}\restoresf \ignorespaces}
+% \setref{NAME}{SNT} defines a cross-reference point NAME (a node or an
+% anchor), which consists of three parts:
+% 1) NAME-title - the current sectioning name taken from \lastsection,
+%                 or the anchor name.
+% 2) NAME-snt   - section number and type, passed as the SNT arg, or
+%                 empty for anchors.
+% 3) NAME-pg    - the page number.
+%
+% This is called from \donoderef, \anchor, and \dofloat.  In the case of
+% floats, there is an additional part, which is not written here:
+% 4) NAME-lof   - the text as it should appear in a @listoffloats.
+%
+\def\setref#1#2{%
+\pdfmkdest{#1}%
+\iflinks
+{%
+\atdummies  % preserve commands, but don't expand them
+\edef\writexrdef##1##2{%
+	\write\auxfile{@xrdef{#1-% #1 of \setref, expanded by the \edef
+	  ##1}{##2}}% these are parameters of \writexrdef
+}%
+\toks0 = \expandafter{\lastsection}%
+\immediate \writexrdef{title}{\the\toks0 }%
+\immediate \writexrdef{snt}{\csname #2\endcsname}% \Ynumbered etc.
+\safewhatsit{\writexrdef{pg}{\folio}}% will be written later, during \shipout
+}%
 \fi
 }
-\def\appendixnoderef{%
-\ifx\lastnode\relax\else
-\expandafter\expandafter\expandafter\setref{\lastnode}%
-{Yappendixletterandtype}%
-\global\let\lastnode=\relax
-\fi
-}
-% @anchor{NAME} -- define xref target at arbitrary point.
-%
-\newcount\savesfregister
-\gdef\savesf{\relax \ifhmode \savesfregister=\spacefactor \fi}
-\gdef\restoresf{\relax \ifhmode \spacefactor=\savesfregister \fi}
-\gdef\anchor#1{\savesf \setref{#1}{Ynothing}\restoresf \ignorespaces}
-% \setref{NAME}{SNT} defines a cross-reference point NAME, namely
-% NAME-title, NAME-pg, and NAME-SNT.  Called from \foonoderef.  We have
-% to set \indexdummies so commands such as @code in a section title
-% aren't expanded.  It would be nicer not to expand the titles in the
-% first place, but there's so many layers that that is hard to do.
-%
-\def\setref#1#2{{%
-\indexdummies
-\pdfmkdest{#1}%
-\dosetq{#1-title}{Ytitle}%
-\dosetq{#1-pg}{Ypagenumber}%
-\dosetq{#1-snt}{#2}%
-}}
 % @xref, @pxref, and @ref generate cross-references.  For \xrefX, #1 is
 % the node name, #2 the name of the Info cross-reference, #3 the printed
 % node name, #4 the name of the Info file, #5 the name of the printed
 % manual.  All but the node name can be omitted.
 \def\xref#1{\putwordSee{} \xrefX[#1,,,,,,,]}
 \def\ref#1{\xrefX[#1,,,,,,,]}
 \def\xrefX[#1,#2,#3,#4,#5,#6]{\begingroup
 \unsepspaces
 \def\printedmanual{\ignorespaces #5}%
-\def\printednodename{\ignorespaces #3}%
+\def\printedrefname{\ignorespaces #3}%
-\setbox1=\hbox{\printedmanual}%
+\setbox1=\hbox{\printedmanual\unskip}%
-\setbox0=\hbox{\printednodename}%
+\setbox0=\hbox{\printedrefname\unskip}%
 \ifdim \wd0 = 0pt
 % No printed node name was explicitly given.
 \expandafter\ifx\csname SETxref-automatic-section-title\endcsname\relax
 % Use the node name inside the square brackets.
-\def\printednodename{\ignorespaces #1}%
+\def\printedrefname{\ignorespaces #1}%
 \else
 % Use the actual chapter/section title appear inside
 % the square brackets.  Use the real section title if we have it.
 \ifdim \wd1 > 0pt
 % It is in another manual, so we don't have it.
-\def\printednodename{\ignorespaces #1}%
+\def\printedrefname{\ignorespaces #1}%
 \else
 \ifhavexrefs
 % We know the real title if we have the xref values.
-\def\printednodename{\refx{#1-title}{}}%
+\def\printedrefname{\refx{#1-title}{}}%
 \else
 % Otherwise just copy the Info node name.
-\def\printednodename{\ignorespaces #1}%
+\def\printedrefname{\ignorespaces #1}%
 \fi%
 \fi
 \fi
 \fi
 %
-% If we use \unhbox0 and \unhbox1 to print the node names, TeX does not
+% Make link in pdf output.
-% insert empty discretionaries after hyphens, which means that it will
-% not find a line break at a hyphen in a node names.  Since some manuals
-% are best written with fairly long node names, containing hyphens, this
-% is a loss.  Therefore, we give the text of the node name again, so it
-% is as if TeX is seeing it for the first time.
 \ifpdf
-\leavevmode
+{\indexnofonts
-\getfilename{#4}%
+\turnoffactive
-\ifnum\filenamelength>0
+% This expands tokens, so do it after making catcode changes, so _
-\startlink attr{/Border [0 0 0]}%
+% etc. don't get their TeX definitions.
-goto file{\the\filename.pdf} name{#1@}%
+\getfilename{#4}%
+%
+% See comments at \activebackslashdouble.
+{\activebackslashdouble \xdef\pdfxrefdest{#1}%
+\backslashparens\pdfxrefdest}%
+%
+\leavevmode
+\startlink attr{/Border [0 0 0]}%
+\ifnum\filenamelength>0
+goto file{\the\filename.pdf} name{\pdfxrefdest}%
+\else
+goto name{\pdfmkpgn{\pdfxrefdest}}%
+\fi
+}%
+\setcolor{\linkcolor}%
+\fi
+%
+% Float references are printed completely differently: "Figure 1.2"
+% instead of "[somenode], p.3".  We distinguish them by the
+% LABEL-title being set to a magic string.
+{%
+% Have to otherify everything special to allow the \csname to
+% include an _ in the xref name, etc.
+\indexnofonts
+\turnoffactive
+\expandafter\global\expandafter\let\expandafter\Xthisreftitle
+\csname XR#1-title\endcsname
+}%
+\iffloat\Xthisreftitle
+% If the user specified the print name (third arg) to the ref,
+% print it instead of our usual "Figure 1.2".
+\ifdim\wd0 = 0pt
+\refx{#1-snt}{}%
 \else
-\startlink attr{/Border [0 0 0]}%
+\printedrefname
-goto name{#1@}%
 \fi
-\linkcolor
+%
-\fi
+% if the user also gave the printed manual name (fifth arg), append
-%
+% "in MANUALNAME".
 \ifdim \wd1 > 0pt
-\putwordsection{} ``\printednodename'' \putwordin{} \cite{\printedmanual}%
+\space \putwordin{} \cite{\printedmanual}%
+\fi
 \else
-% _ (for example) has to be the character _ for the purposes of the
+% node/anchor (non-float) references.
-% control sequence corresponding to the node, but it has to expand
+%
-% into the usual \leavevmode...\vrule stuff for purposes of
+% If we use \unhbox0 and \unhbox1 to print the node names, TeX does not
-% printing. So we \turnoffactive for the \refx-snt, back on for the
+% insert empty discretionaries after hyphens, which means that it will
-% printing, back off for the \refx-pg.
+% not find a line break at a hyphen in a node names.  Since some manuals
-{\normalturnoffactive
+% are best written with fairly long node names, containing hyphens, this
-% Only output a following space if the -snt ref is nonempty; for
+% is a loss.  Therefore, we give the text of the node name again, so it
-% @unnumbered and @anchor, it won't be.
+% is as if TeX is seeing it for the first time.
-\setbox2 = \hbox{\ignorespaces \refx{#1-snt}{}}%
+\ifdim \wd1 > 0pt
-\ifdim \wd2 > 0pt \refx{#1-snt}\space\fi
+\putwordSection{} ``\printedrefname'' \putwordin{} \cite{\printedmanual}%
-}%
+\else
-% [mynode],
+% _ (for example) has to be the character _ for the purposes of the
-[\printednodename],\space
+% control sequence corresponding to the node, but it has to expand
-% page 3
+% into the usual \leavevmode...\vrule stuff for purposes of
-\turnoffactive \putwordpage\tie\refx{#1-pg}{}%
+% printing. So we \turnoffactive for the \refx-snt, back on for the
+% printing, back off for the \refx-pg.
+{\turnoffactive
+% Only output a following space if the -snt ref is nonempty; for
+% @unnumbered and @anchor, it won't be.
+\setbox2 = \hbox{\ignorespaces \refx{#1-snt}{}}%
+\ifdim \wd2 > 0pt \refx{#1-snt}\space\fi
+}%
+% output the `[mynode]' via a macro so it can be overridden.
+\xrefprintnodename\printedrefname
+%
+% But we always want a comma and a space:
+,\space
+%
+% output the `page 3'.
+\turnoffactive \putwordpage\tie\refx{#1-pg}{}%
+\fi
 \fi
 \endlink
 \endgroup}
-% \dosetq is the interface for calls from other macros
+% This macro is called from \xrefX for the `[nodename]' part of xref
+% output.  It's a separate macro only so it can be changed more easily,
-% Use \normalturnoffactive so that punctuation chars such as underscore
+% since square brackets don't work well in some documents.  Particularly
-% and backslash work in node names.  (\turnoffactive doesn't do \.)
+% one that Bob is working on :).
-\def\dosetq#1#2{%
+%
-{\let\folio=0%
+\def\xrefprintnodename#1{[#1]}
-\normalturnoffactive
-\edef\next{\write\auxfile{\internalsetq{#1}{#2}}}%
+% Things referred to by \setref.
-\iflinks
+%
-\next
-\fi
-}%
-}
-% \internalsetq {foo}{page} expands into
-% CHARACTERS 'xrdef {foo}{...expansion of \Ypage...}
-% When the aux file is read, ' is the escape character
-\def\internalsetq #1#2{'xrdef {#1}{\csname #2\endcsname}}
-% Things to be expanded by \internalsetq
-\def\Ypagenumber{\folio}
-\def\Ytitle{\thissection}
 \def\Ynothing{}
+\def\Yomitfromtoc{}
-\def\Ysectionnumberandtype{%
+\def\Ynumbered{%
-\ifnum\secno=0 \putwordChapter\xreftie\the\chapno %
+\ifnum\secno=0
-\else \ifnum \subsecno=0 \putwordSection\xreftie\the\chapno.\the\secno %
+\putwordChapter@tie \the\chapno
-\else \ifnum \subsubsecno=0 %
+\else \ifnum\subsecno=0
-\putwordSection\xreftie\the\chapno.\the\secno.\the\subsecno %
+\putwordSection@tie \the\chapno.\the\secno
-\else %
+\else \ifnum\subsubsecno=0
-\putwordSection\xreftie\the\chapno.\the\secno.\the\subsecno.\the\subsubsecno %
+\putwordSection@tie \the\chapno.\the\secno.\the\subsecno
-\fi \fi \fi }
+\else
+\putwordSection@tie \the\chapno.\the\secno.\the\subsecno.\the\subsubsecno
-\def\Yappendixletterandtype{%
+\fi\fi\fi
-\ifnum\secno=0 \putwordAppendix\xreftie'char\the\appendixno{}%
+}
-\else \ifnum \subsecno=0 \putwordSection\xreftie'char\the\appendixno.\the\secno %
+\def\Yappendix{%
-\else \ifnum \subsubsecno=0 %
+\ifnum\secno=0
-\putwordSection\xreftie'char\the\appendixno.\the\secno.\the\subsecno %
+\putwordAppendix@tie @char\the\appendixno{}%
-\else %
+\else \ifnum\subsecno=0
-\putwordSection\xreftie'char\the\appendixno.\the\secno.\the\subsecno.\the\subsubsecno %
+\putwordSection@tie @char\the\appendixno.\the\secno
-\fi \fi \fi }
+\else \ifnum\subsubsecno=0
+\putwordSection@tie @char\the\appendixno.\the\secno.\the\subsecno
-\gdef\xreftie{'tie}
+\else
+\putwordSection@tie
-% Use TeX 3.0's \inputlineno to get the line number, for better error
+@char\the\appendixno.\the\secno.\the\subsecno.\the\subsubsecno
-% messages, but if we're using an old version of TeX, don't do anything.
+\fi\fi\fi
-%
+}
-\ifx\inputlineno\thisisundefined
-\let\linenumber = \empty % Non-3.0.
-\else
-\def\linenumber{\the\inputlineno:\space}
-\fi
 % Define \refx{NAME}{SUFFIX} to reference a cross-reference string named NAME.
 % If its value is nonempty, SUFFIX is output afterward.
+%
 \def\refx#1#2{%
-\expandafter\ifx\csname X#1\endcsname\relax
+{%
+\indexnofonts
+\otherbackslash
+\expandafter\global\expandafter\let\expandafter\thisrefX
+\csname XR#1\endcsname
+}%
+\ifx\thisrefX\relax
 % If not defined, say something at least.
 \angleleft un\-de\-fined\angleright
 \iflinks
 \ifhavexrefs
 \message{\linenumber Undefined cross reference `#1'.}%
 \fi
 \fi
 \fi
 \else
 % It's defined, so just use it.
-\csname X#1\endcsname
+\thisrefX
 \fi
 #2% Output the suffix in any case.
 }
-% This is the macro invoked by entries in the aux file.
+% This is the macro invoked by entries in the aux file.  Usually it's
-%
+% just a \def (we prepend XR to the control sequence name to avoid
-\def\xrdef#1{\begingroup
+% collisions).  But if this is a float type, we have more work to do.
-% Reenable \ as an escape while reading the second argument.
+%
-\catcode`\\ = 0
+\def\xrdef#1#2{%
-\afterassignment\endgroup
+{% The node name might contain 8-bit characters, which in our current
-\expandafter\gdef\csname X#1\endcsname
+% implementation are changed to commands like @'e.  Don't let these
+% mess up the control sequence name.
+\indexnofonts
+\turnoffactive
+\xdef\safexrefname{#1}%
+}%
+%
+\expandafter\gdef\csname XR\safexrefname\endcsname{#2}% remember this xref
+%
+% Was that xref control sequence that we just defined for a float?
+\expandafter\iffloat\csname XR\safexrefname\endcsname
+% it was a float, and we have the (safe) float type in \iffloattype.
+\expandafter\let\expandafter\floatlist
+\csname floatlist\iffloattype\endcsname
+%
+% Is this the first time we've seen this float type?
+\expandafter\ifx\floatlist\relax
+\toks0 = {\do}% yes, so just \do
+\else
+% had it before, so preserve previous elements in list.
+\toks0 = \expandafter{\floatlist\do}%
+\fi
+%
+% Remember this xref in the control sequence \floatlistFLOATTYPE,
+% for later use in \listoffloats.
+\expandafter\xdef\csname floatlist\iffloattype\endcsname{\the\toks0
+{\safexrefname}}%
+\fi
 }
 % Read the last existing aux file, if any.  No error if none exists.
-\def\readauxfile{\begingroup
+%
+\def\tryauxfile{%
+\openin 1 \jobname.aux
+\ifeof 1 \else
+\readdatafile{aux}%
+\global\havexrefstrue
+\fi
+\closein 1
+}
+\def\setupdatafile{%
 \catcode`\^^@=\other
 \catcode`\^^A=\other
 \catcode`\^^B=\other
 \catcode`\^^C=\other
 \catcode`\^^D=\other
 \catcode`\^^[=\other
 \catcode`\^^\=\other
 \catcode`\^^]=\other
 \catcode`\^^^=\other
 \catcode`\^^_=\other
-\catcode`\@=\other
+% It was suggested to set the catcode of ^ to 7, which would allow ^^e4 etc.
-\catcode`\^=\other
-% It was suggested to define this as 7, which would allow ^^e4 etc.
 % in xref tags, i.e., node names.  But since ^^e4 notation isn't
 % supported in the main text, it doesn't seem desirable.  Furthermore,
 % that is not enough: for node names that actually contain a ^
 % character, we would end up writing a line like this: 'xrdef {'hat
 % b-title}{'hat b} and \xrdef does a \csname...\endcsname on the first
 %
 % The other change necessary for this was to define \auxhat:
 % \def\auxhat{\def^{'hat }}% extra space so ok if followed by letter
 % and then to call \auxhat in \setq.
 %
+\catcode`\^=\other
+%
+% Special characters.  Should be turned off anyway, but...
 \catcode`\~=\other
 \catcode`\[=\other
 \catcode`\]=\other
 \catcode`\"=\other
 \catcode`\_=\other
 \catcode`\<=\other
 \catcode`\>=\other
 \catcode`\$=\other
 \catcode`\#=\other
 \catcode`\&=\other
+\catcode`\%=\other
 \catcode`+=\other % avoid \+ for paranoia even though we've turned it off
-% Make the characters 128-255 be printing characters
+%
+% This is to support \ in node names and titles, since the \
+% characters end up in a \csname.  It's easier than
+% leaving it active and making its active definition an actual \
+% character.  What I don't understand is why it works in the *value*
+% of the xrdef.  Seems like it should be a catcode12 \, and that
+% should not typeset properly.  But it works, so I'm moving on for
+% now.  --karl, 15jan04.
+\catcode`\\=\other
+%
+% Make the characters 128-255 be printing characters.
 {%
-\count 1=128
+\count1=128
 \def\loop{%
-\catcode\count 1=\other
+\catcode\count1=\other
-\advance\count 1 by 1
+\advance\count1 by 1
-\ifnum \count 1<256 \loop \fi
+\ifnum \count1<256 \loop \fi
 }%
 }%
-% The aux file uses ' as the escape (for now).
+%
-% Turn off \ as an escape so we do not lose on
+% @ is our escape character in .aux files, and we need braces.
-% entries which were dumped with control sequences in their names.
-% For example, 'xrdef {$\leq $-fun}{page ...} made by @defun ^^
-% Reference to such entries still does not work the way one would wish,
-% but at least they do not bomb out when the aux file is read in.
 \catcode`\{=1
 \catcode`\}=2
-\catcode`\%=\other
+\catcode`\@=0
-\catcode`\'=0
+}
-\catcode`\\=\other
-%
+\def\readdatafile#1{%
-\openin 1 \jobname.aux
+\begingroup
-\ifeof 1 \else
+\setupdatafile
-\closein 1
+\input\jobname.#1
-\input \jobname.aux
-\global\havexrefstrue
-\global\warnedobstrue
-\fi
-% Open the new aux file.  TeX will close it automatically at exit.
-\openout\auxfile=\jobname.aux
 \endgroup}
-% Footnotes.
+\message{insertions,}
+% including footnotes.
 \newcount \footnoteno
 % The trailing space in the following definition for supereject is
 % vital for proper filling; pages come out unaligned when you do a
 \def\supereject{\par\penalty -20000\footnoteno =0 }
 % @footnotestyle is meaningful for info output only.
 \let\footnotestyle=\comment
-\let\ptexfootnote=\footnote
 {\catcode `\@=11
 %
 % Auto-number footnotes.  Otherwise like plain.
 \gdef\footnote{%
+\let\indent=\ptexindent
+\let\noindent=\ptexnoindent
 \global\advance\footnoteno by \@ne
 \edef\thisfootno{$^{\the\footnoteno}$}%
 %
 % In case the footnote comes at the end of a sentence, preserve the
 % extra spacing after we do the footnote number.
 \let\@sf\empty
-\ifhmode\edef\@sf{\spacefactor\the\spacefactor}\/\fi
+\ifhmode\edef\@sf{\spacefactor\the\spacefactor}\ptexslash\fi
 %
 % Remove inadvertent blank space before typesetting the footnote number.
 \unskip
 \thisfootno\@sf
-\footnotezzz
+\dofootnote
 }%
 % Don't bother with the trickery in plain.tex to not require the
 % footnote text as a parameter.  Our footnotes don't need to be so general.
 %
-% Oh yes, they do; otherwise, @ifset and anything else that uses
+% Oh yes, they do; otherwise, @ifset (and anything else that uses
-% \parseargline fail inside footnotes because the tokens are fixed when
+% \parseargline) fails inside footnotes because the tokens are fixed when
 % the footnote is read.  --karl, 16nov96.
 %
-\long\gdef\footnotezzz{\insert\footins\bgroup
+\gdef\dofootnote{%
+\insert\footins\bgroup
 % We want to typeset this text as a normal paragraph, even if the
 % footnote reference occurs in (for example) a display environment.
 % So reset some parameters.
+\hsize=\pagewidth
 \interlinepenalty\interfootnotelinepenalty
 \splittopskip\ht\strutbox % top baseline for broken footnotes
 \splitmaxdepth\dp\strutbox
 \floatingpenalty\@MM
 \leftskip\z@skip
 \xspaceskip\z@skip
 \parindent\defaultparindent
 %
 \smallfonts \rm
 %
-% Hang the footnote text off the number.
+% Because we use hanging indentation in footnotes, a @noindent appears
-\hang
+% to exdent this text, so make it be a no-op.  makeinfo does not use
+% hanging indentation so @noindent can still be needed within footnote
+% text after an @example or the like (not that this is good style).
+\let\noindent = \relax
+%
+% Hang the footnote text off the number.  Use \everypar in case the
+% footnote extends for more than one paragraph.
+\everypar = {\hang}%
 \textindent{\thisfootno}%
 %
 % Don't crash into the line above the footnote text.  Since this
 % expands into a box, it must come within the paragraph, lest it
 % provide a place where TeX can split the footnote.
 \footstrut
 \futurelet\next\fo@t
 }
-\def\fo@t{\ifcat\bgroup\noexpand\next \let\next\f@@t
-\else\let\next\f@t\fi \next}
-\def\f@@t{\bgroup\aftergroup\@foot\let\next}
-\def\f@t#1{#1\@foot}
-\def\@foot{\strut\par\egroup}
 }%end \catcode `\@=11
-% Set the baselineskip to #1, and the lineskip and strut size
+% In case a @footnote appears in a vbox, save the footnote text and create
-% correspondingly.  There is no deep meaning behind these magic numbers
+% the real \insert just after the vbox finished.  Otherwise, the insertion
-% used as factors; they just match (closely enough) what Knuth defined.
+% would be lost.
-%
+% Similarly, if a @footnote appears inside an alignment, save the footnote
-\def\lineskipfactor{.08333}
+% text to a box and make the \insert when a row of the table is finished.
-\def\strutheightpercent{.70833}
+% And the same can be done for other insert classes.  --kasal, 16nov03.
-\def\strutdepthpercent {.29167}
-%
+% Replace the \insert primitive by a cheating macro.
-\def\setleading#1{%
+% Deeper inside, just make sure that the saved insertions are not spilled
-\normalbaselineskip = #1\relax
+% out prematurely.
-\normallineskip = \lineskipfactor\normalbaselineskip
+%
-\normalbaselines
+\def\startsavinginserts{%
-\setbox\strutbox =\hbox{%
+\ifx \insert\ptexinsert
-\vrule width0pt height\strutheightpercent\baselineskip
+\let\insert\saveinsert
-depth \strutdepthpercent \baselineskip
+\else
-}%
+\let\checkinserts\relax
-}
+\fi
+}
-% @| inserts a changebar to the left of the current line.  It should
-% surround any changed text.  This approach does *not* work if the
+% This \insert replacement works for both \insert\footins{foo} and
-% change spans more than two lines of output.  To handle that, we would
+% \insert\footins\bgroup foo\egroup, but it doesn't work for \insert27{foo}.
-% have adopt a much more difficult approach (putting marks into the main
+%
-% vertical list for the beginning and end of each change).
+\def\saveinsert#1{%
-%
+\edef\next{\noexpand\savetobox \makeSAVEname#1}%
-\def\|{%
+\afterassignment\next
-% \vadjust can only be used in horizontal mode.
+% swallow the left brace
-\leavevmode
+\let\temp =
-%
+}
-% Append this vertical mode material after the current line in the output.
+\def\makeSAVEname#1{\makecsname{SAVE\expandafter\gobble\string#1}}
-\vadjust{%
+\def\savetobox#1{\global\setbox#1 = \vbox\bgroup \unvbox#1}
-% We want to insert a rule with the height and depth of the current
-% leading; that is exactly what \strutbox is supposed to record.
+\def\checksaveins#1{\ifvoid#1\else \placesaveins#1\fi}
-\vskip-\baselineskip
-%
+\def\placesaveins#1{%
-% \vadjust-items are inserted at the left edge of the type.  So
+\ptexinsert \csname\expandafter\gobblesave\string#1\endcsname
-% the \llap here moves out into the left-hand margin.
+{\box#1}%
-\llap{%
+}
-%
-% For a thicker or thinner bar, change the `1pt'.
+% eat @SAVE -- beware, all of them have catcode \other:
-\vrule height\baselineskip width1pt
+{
-%
+\def\dospecials{\do S\do A\do V\do E} \uncatcodespecials  %  ;-)
-% This is the space between the bar and the text.
+\gdef\gobblesave @SAVE{}
-\hskip 12pt
+}
-}%
-}%
+% initialization:
-}
+\def\newsaveins #1{%
+\edef\next{\noexpand\newsaveinsX \makeSAVEname#1}%
-% For a final copy, take out the rectangles
+\next
-% that mark overfull boxes (in case you have decided
+}
-% that the text looks ok even though it passes the margin).
+\def\newsaveinsX #1{%
-%
+\csname newbox\endcsname #1%
-\def\finalout{\overfullrule=0pt}
+\expandafter\def\expandafter\checkinserts\expandafter{\checkinserts
+\checksaveins #1}%
+}
+% initialize:
+\let\checkinserts\empty
+\newsaveins\footins
+\newsaveins\margin
 % @image.  We use the macros from epsf.tex to support this.
 % If epsf.tex is not installed and @image is used, we complain.
 %
 % Check for and read epsf.tex up front.  If we read it only at @image
 % time, we might be inside a group, and then its definitions would get
 % undone and the next image would fail.
 \openin 1 = epsf.tex
 \ifeof 1 \else
-\closein 1
+% Do not bother showing banner with epsf.tex v2.7k (available in
-% Do not bother showing banner with post-v2.7 epsf.tex (available in
+% doc/epsf.tex and on ctan).
-% doc/epsf.tex until it shows up on ctan).
 \def\epsfannounce{\toks0 = }%
 \input epsf.tex
 \fi
+\closein 1
 %
 % We will only complain once about lack of epsf.tex.
 \newif\ifwarnednoepsf
 \newhelp\noepsfhelp{epsf.tex must be installed for images to
 work.  It is also included in the Texinfo distribution, or you can get
 \errhelp = \noepsfhelp
 \errmessage{epsf.tex not found, images will be ignored}%
 \global\warnednoepsftrue
 \fi
 \else
-\imagexxx #1,,,\finish
+\imagexxx #1,,,,,\finish
 \fi
 }
 %
 % Arguments to @image:
 % #1 is (mandatory) image filename; we tack on .eps extension.
 % #2 is (optional) width, #3 is (optional) height.
-% #4 is just the usual extra ignored arg for parsing this stuff.
+% #4 is (ignored optional) html alt text.
-\def\imagexxx#1,#2,#3,#4\finish{%
+% #5 is (ignored optional) extension.
+% #6 is just the usual extra ignored arg for parsing this stuff.
+\newif\ifimagevmode
+\def\imagexxx#1,#2,#3,#4,#5,#6\finish{\begingroup
+\catcode`\^^M = 5     % in case we're inside an example
+\normalturnoffactive  % allow _ et al. in names
+% If the image is by itself, center it.
+\ifvmode
+\imagevmodetrue
+\nobreak\medskip
+% Usually we'll have text after the image which will insert
+% \parskip glue, so insert it here too to equalize the space
+% above and below.
+\nobreak\vskip\parskip
+\nobreak
+\fi
+%
+% Leave vertical mode so that indentation from an enclosing
+% environment such as @quotation is respected.  On the other hand, if
+% it's at the top level, we don't want the normal paragraph indentation.
+\noindent
+%
+% Output the image.
 \ifpdf
-\centerline{\dopdfimage{#1}{#2}{#3}}%
+\dopdfimage{#1}{#2}{#3}%
 \else
 % \epsfbox itself resets \epsf?size at each figure.
 \setbox0 = \hbox{\ignorespaces #2}\ifdim\wd0 > 0pt \epsfxsize=#2\relax \fi
 \setbox0 = \hbox{\ignorespaces #3}\ifdim\wd0 > 0pt \epsfysize=#3\relax \fi
+\epsfbox{#1.eps}%
+\fi
+%
+\ifimagevmode \medskip \fi  % space after the standalone image
+\endgroup}
+% @float FLOATTYPE,LABEL,LOC ... @end float for displayed figures, tables,
+% etc.  We don't actually implement floating yet, we always include the
+% float "here".  But it seemed the best name for the future.
+%
+\envparseargdef\float{\eatcommaspace\eatcommaspace\dofloat#1, , ,\finish}
+% There may be a space before second and/or third parameter; delete it.
+\def\eatcommaspace#1, {#1,}
+% #1 is the optional FLOATTYPE, the text label for this float, typically
+% "Figure", "Table", "Example", etc.  Can't contain commas.  If omitted,
+% this float will not be numbered and cannot be referred to.
+%
+% #2 is the optional xref label.  Also must be present for the float to
+% be referable.
+%
+% #3 is the optional positioning argument; for now, it is ignored.  It
+% will somehow specify the positions allowed to float to (here, top, bottom).
+%
+% We keep a separate counter for each FLOATTYPE, which we reset at each
+% chapter-level command.
+\let\resetallfloatnos=\empty
+%
+\def\dofloat#1,#2,#3,#4\finish{%
+\let\thiscaption=\empty
+\let\thisshortcaption=\empty
+%
+% don't lose footnotes inside @float.
+%
+% BEWARE: when the floats start float, we have to issue warning whenever an
+% insert appears inside a float which could possibly float. --kasal, 26may04
+%
+\startsavinginserts
+%
+% We can't be used inside a paragraph.
+\par
+%
+\vtop\bgroup
+\def\floattype{#1}%
+\def\floatlabel{#2}%
+\def\floatloc{#3}% we do nothing with this yet.
+%
+\ifx\floattype\empty
+\let\safefloattype=\empty
+\else
+{%
+% the floattype might have accents or other special characters,
+% but we need to use it in a control sequence name.
+\indexnofonts
+\turnoffactive
+\xdef\safefloattype{\floattype}%
+}%
+\fi
+%
+% If label is given but no type, we handle that as the empty type.
+\ifx\floatlabel\empty \else
+% We want each FLOATTYPE to be numbered separately (Figure 1,
+% Table 1, Figure 2, ...).  (And if no label, no number.)
+%
+\expandafter\getfloatno\csname\safefloattype floatno\endcsname
+\global\advance\floatno by 1
+%
+{%
+% This magic value for \lastsection is output by \setref as the
+% XREFLABEL-title value.  \xrefX uses it to distinguish float
+% labels (which have a completely different output format) from
+% node and anchor labels.  And \xrdef uses it to construct the
+% lists of floats.
+%
+\edef\lastsection{\floatmagic=\safefloattype}%
+\setref{\floatlabel}{Yfloat}%
+}%
+\fi
+%
+% start with \parskip glue, I guess.
+\vskip\parskip
+%
+% Don't suppress indentation if a float happens to start a section.
+\restorefirstparagraphindent
+}
+% we have these possibilities:
+% @float Foo,lbl & @caption{Cap}: Foo 1.1: Cap
+% @float Foo,lbl & no caption:    Foo 1.1
+% @float Foo & @caption{Cap}:     Foo: Cap
+% @float Foo & no caption:        Foo
+% @float ,lbl & Caption{Cap}:     1.1: Cap
+% @float ,lbl & no caption:       1.1
+% @float & @caption{Cap}:         Cap
+% @float & no caption:
+%
+\def\Efloat{%
+\let\floatident = \empty
+%
+% In all cases, if we have a float type, it comes first.
+\ifx\floattype\empty \else \def\floatident{\floattype}\fi
+%
+% If we have an xref label, the number comes next.
+\ifx\floatlabel\empty \else
+\ifx\floattype\empty \else % if also had float type, need tie first.
+\appendtomacro\floatident{\tie}%
+\fi
+% the number.
+\appendtomacro\floatident{\chaplevelprefix\the\floatno}%
+\fi
+%
+% Start the printed caption with what we've constructed in
+% \floatident, but keep it separate; we need \floatident again.
+\let\captionline = \floatident
+%
+\ifx\thiscaption\empty \else
+\ifx\floatident\empty \else
+	\appendtomacro\captionline{: }% had ident, so need a colon between
+\fi
+%
+% caption text.
+\appendtomacro\captionline{\scanexp\thiscaption}%
+\fi
+%
+% If we have anything to print, print it, with space before.
+% Eventually this needs to become an \insert.
+\ifx\captionline\empty \else
+\vskip.5\parskip
+\captionline
+%
+% Space below caption.
+\vskip\parskip
+\fi
+%
+% If have an xref label, write the list of floats info.  Do this
+% after the caption, to avoid chance of it being a breakpoint.
+\ifx\floatlabel\empty \else
+% Write the text that goes in the lof to the aux file as
+% \floatlabel-lof.  Besides \floatident, we include the short
+% caption if specified, else the full caption if specified, else nothing.
+{%
+\atdummies
+%
+% since we read the caption text in the macro world, where ^^M
+% is turned into a normal character, we have to scan it back, so
+% we don't write the literal three characters "^^M" into the aux file.
+	\scanexp{%
+	  \xdef\noexpand\gtemp{%
+	    \ifx\thisshortcaption\empty
+	      \thiscaption
+	    \else
+	      \thisshortcaption
+	    \fi
+	  }%
+	}%
+\immediate\write\auxfile{@xrdef{\floatlabel-lof}{\floatident
+	  \ifx\gtemp\empty \else : \gtemp \fi}}%
+}%
+\fi
+\egroup  % end of \vtop
+%
+% place the captured inserts
+%
+% BEWARE: when the floats start floating, we have to issue warning
+% whenever an insert appears inside a float which could possibly
+% float. --kasal, 26may04
+%
+\checkinserts
+}
+% Append the tokens #2 to the definition of macro #1, not expanding either.
+%
+\def\appendtomacro#1#2{%
+\expandafter\def\expandafter#1\expandafter{#1#2}%
+}
+% @caption, @shortcaption
+%
+\def\caption{\docaption\thiscaption}
+\def\shortcaption{\docaption\thisshortcaption}
+\def\docaption{\checkenv\float \bgroup\scanargctxt\defcaption}
+\def\defcaption#1#2{\egroup \def#1{#2}}
+% The parameter is the control sequence identifying the counter we are
+% going to use.  Create it if it doesn't exist and assign it to \floatno.
+\def\getfloatno#1{%
+\ifx#1\relax
+% Haven't seen this figure type before.
+\csname newcount\endcsname #1%
+%
+% Remember to reset this floatno at the next chap.
+\expandafter\gdef\expandafter\resetallfloatnos
+\expandafter{\resetallfloatnos #1=0 }%
+\fi
+\let\floatno#1%
+}
+% \setref calls this to get the XREFLABEL-snt value.  We want an @xref
+% to the FLOATLABEL to expand to "Figure 3.1".  We call \setref when we
+% first read the @float command.
+%
+\def\Yfloat{\floattype@tie \chaplevelprefix\the\floatno}%
+% Magic string used for the XREFLABEL-title value, so \xrefX can
+% distinguish floats from other xref types.
+\def\floatmagic{!!float!!}
+% #1 is the control sequence we are passed; we expand into a conditional
+% which is true if #1 represents a float ref.  That is, the magic
+% \lastsection value which we \setref above.
+%
+\def\iffloat#1{\expandafter\doiffloat#1==\finish}
+%
+% #1 is (maybe) the \floatmagic string.  If so, #2 will be the
+% (safe) float type for this float.  We set \iffloattype to #2.
+%
+\def\doiffloat#1=#2=#3\finish{%
+\def\temp{#1}%
+\def\iffloattype{#2}%
+\ifx\temp\floatmagic
+}
+% @listoffloats FLOATTYPE - print a list of floats like a table of contents.
+%
+\parseargdef\listoffloats{%
+\def\floattype{#1}% floattype
+{%
+% the floattype might have accents or other special characters,
+% but we need to use it in a control sequence name.
+\indexnofonts
+\turnoffactive
+\xdef\safefloattype{\floattype}%
+}%
+%
+% \xrdef saves the floats as a \do-list in \floatlistSAFEFLOATTYPE.
+\expandafter\ifx\csname floatlist\safefloattype\endcsname \relax
+\ifhavexrefs
+% if the user said @listoffloats foo but never @float foo.
+\message{\linenumber No `\safefloattype' floats to list.}%
+\fi
+\else
 \begingroup
-\catcode`\^^M = 5 % in case we're inside an example
+\leftskip=\tocindent  % indent these entries like a toc
-% If the image is by itself, center it.
+\let\do=\listoffloatsdo
-\ifvmode
+\csname floatlist\safefloattype\endcsname
-\nobreak\bigskip
-% Usually we'll have text after the image which will insert
-% \parskip glue, so insert it here too to equalize the space
-% above and below.
-\nobreak\vskip\parskip
-\nobreak
-\centerline{\epsfbox{#1.eps}}%
-\bigbreak
-\else
-% In the middle of a paragraph, no extra space.
-\epsfbox{#1.eps}%
-\fi
 \endgroup
 \fi
 }
+% This is called on each entry in a list of floats.  We're passed the
+% xref label, in the form LABEL-title, which is how we save it in the
+% aux file.  We strip off the -title and look up \XRLABEL-lof, which
+% has the text we're supposed to typeset here.
+%
+% Figures without xref labels will not be included in the list (since
+% they won't appear in the aux file).
+%
+\def\listoffloatsdo#1{\listoffloatsdoentry#1\finish}
+\def\listoffloatsdoentry#1-title\finish{{%
+% Can't fully expand XR#1-lof because it can contain anything.  Just
+% pass the control sequence.  On the other hand, XR#1-pg is just the
+% page number, and we want to fully expand that so we can get a link
+% in pdf output.
+\toksA = \expandafter{\csname XR#1-lof\endcsname}%
+%
+% use the same \entry macro we use to generate the TOC and index.
+\edef\writeentry{\noexpand\entry{\the\toksA}{\csname XR#1-pg\endcsname}}%
+\writeentry
+}}
 \message{localization,}
-% and i18n.
 % @documentlanguage is usually given very early, just after
 % @setfilename.  If done too late, it may not override everything
-% properly.  Single argument is the language abbreviation.
+% properly.  Single argument is the language (de) or locale (de_DE)
-% It would be nice if we could set up a hyphenation file here.
+% abbreviation.  It would be nice if we could set up a hyphenation file.
 %
-\def\documentlanguage{\parsearg\dodocumentlanguage}
+{
-\def\dodocumentlanguage#1{%
+\catcode`\_ = \active
+\globaldefs=1
+\parseargdef\documentlanguage{\begingroup
+\let_=\normalunderscore  % normal _ character for filenames
 \tex % read txi-??.tex file in plain TeX.
-% Read the file if it exists.
+% Read the file by the name they passed if it exists.
+\openin 1 txi-#1.tex
+\ifeof 1
+\documentlanguagetrywithoutunderscore{#1_\finish}%
+\else
+\input txi-#1.tex
+\fi
+\closein 1
+\endgroup
+\endgroup}
+}
+%
+% If they passed de_DE, and txi-de_DE.tex doesn't exist,
+% try txi-de.tex.
+%
+\def\documentlanguagetrywithoutunderscore#1_#2\finish{%
 \openin 1 txi-#1.tex
-\ifeof1
+\ifeof 1
 \errhelp = \nolanghelp
 \errmessage{Cannot read language file txi-#1.tex}%
-\let\temp = \relax
 \else
-\def\temp{\input txi-#1.tex }%
+\input txi-#1.tex
 \fi
-\temp
+\closein 1
-\endgroup
+}
-}
+%
 \newhelp\nolanghelp{The given language definition file cannot be found or
 is empty.  Maybe you need to install it?  In the current directory
 should work if nowhere else does.}
+% Set the catcode of characters 128 through 255 to the specified number.
-% @documentencoding should change something in TeX eventually, most
+%
-% likely, but for now just recognize it.
+\def\setnonasciicharscatcode#1{%
-\let\documentencoding = \comment
+\count255=128
+\loop\ifnum\count255<256
+\global\catcode\count255=#1\relax
-% Page size parameters.
+\advance\count255 by 1
-%
+\repeat
+}
+\def\setnonasciicharscatcodenonglobal#1{%
+\count255=128
+\loop\ifnum\count255<256
+\catcode\count255=#1\relax
+\advance\count255 by 1
+\repeat
+}
+% @documentencoding sets the definition of non-ASCII characters
+% according to the specified encoding.
+%
+\parseargdef\documentencoding{%
+% Encoding being declared for the document.
+\def\declaredencoding{\csname #1.enc\endcsname}%
+%
+% Supported encodings: names converted to tokens in order to be able
+% to compare them with \ifx.
+\def\ascii{\csname US-ASCII.enc\endcsname}%
+\def\latnine{\csname ISO-8859-15.enc\endcsname}%
+\def\latone{\csname ISO-8859-1.enc\endcsname}%
+\def\lattwo{\csname ISO-8859-2.enc\endcsname}%
+\def\utfeight{\csname UTF-8.enc\endcsname}%
+%
+\ifx \declaredencoding \ascii
+\asciichardefs
+%
+\else \ifx \declaredencoding \lattwo
+\setnonasciicharscatcode\active
+\lattwochardefs
+%
+\else \ifx \declaredencoding \latone
+\setnonasciicharscatcode\active
+\latonechardefs
+%
+\else \ifx \declaredencoding \latnine
+\setnonasciicharscatcode\active
+\latninechardefs
+%
+\else \ifx \declaredencoding \utfeight
+\setnonasciicharscatcode\active
+\utfeightchardefs
+%
+\else
+\message{Unknown document encoding #1, ignoring.}%
+%
+\fi % utfeight
+\fi % latnine
+\fi % latone
+\fi % lattwo
+\fi % ascii
+}
+% A message to be logged when using a character that isn't available
+% the default font encoding (OT1).
+%
+\def\missingcharmsg#1{\message{Character missing in OT1 encoding: #1.}}
+% Take account of \c (plain) vs. \, (Texinfo) difference.
+\def\cedilla#1{\ifx\c\ptexc\c{#1}\else\,{#1}\fi}
+% First, make active non-ASCII characters in order for them to be
+% correctly categorized when TeX reads the replacement text of
+% macros containing the character definitions.
+\setnonasciicharscatcode\active
+%
+% Latin1 (ISO-8859-1) character definitions.
+\def\latonechardefs{%
+\gdef^^a0{~}
+\gdef^^a1{\exclamdown}
+\gdef^^a2{\missingcharmsg{CENT SIGN}}
+\gdef^^a3{{\pounds}}
+\gdef^^a4{\missingcharmsg{CURRENCY SIGN}}
+\gdef^^a5{\missingcharmsg{YEN SIGN}}
+\gdef^^a6{\missingcharmsg{BROKEN BAR}}
+\gdef^^a7{\S}
+\gdef^^a8{\"{}}
+\gdef^^a9{\copyright}
+\gdef^^aa{\ordf}
+\gdef^^ab{\missingcharmsg{LEFT-POINTING DOUBLE ANGLE QUOTATION MARK}}
+\gdef^^ac{$\lnot$}
+\gdef^^ad{\-}
+\gdef^^ae{\registeredsymbol}
+\gdef^^af{\={}}
+%
+\gdef^^b0{\textdegree}
+\gdef^^b1{$\pm$}
+\gdef^^b2{$^2$}
+\gdef^^b3{$^3$}
+\gdef^^b4{\'{}}
+\gdef^^b5{$\mu$}
+\gdef^^b6{\P}
+%
+\gdef^^b7{$^.$}
+\gdef^^b8{\cedilla\ }
+\gdef^^b9{$^1$}
+\gdef^^ba{\ordm}
+%
+\gdef^^bb{\missingcharmsg{RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK}}
+\gdef^^bc{$1\over4$}
+\gdef^^bd{$1\over2$}
+\gdef^^be{$3\over4$}
+\gdef^^bf{\questiondown}
+%
+\gdef^^c0{\`A}
+\gdef^^c1{\'A}
+\gdef^^c2{\^A}
+\gdef^^c3{\~A}
+\gdef^^c4{\"A}
+\gdef^^c5{\ringaccent A}
+\gdef^^c6{\AE}
+\gdef^^c7{\cedilla C}
+\gdef^^c8{\`E}
+\gdef^^c9{\'E}
+\gdef^^ca{\^E}
+\gdef^^cb{\"E}
+\gdef^^cc{\`I}
+\gdef^^cd{\'I}
+\gdef^^ce{\^I}
+\gdef^^cf{\"I}
+%
+\gdef^^d0{\missingcharmsg{LATIN CAPITAL LETTER ETH}}
+\gdef^^d1{\~N}
+\gdef^^d2{\`O}
+\gdef^^d3{\'O}
+\gdef^^d4{\^O}
+\gdef^^d5{\~O}
+\gdef^^d6{\"O}
+\gdef^^d7{$\times$}
+\gdef^^d8{\O}
+\gdef^^d9{\`U}
+\gdef^^da{\'U}
+\gdef^^db{\^U}
+\gdef^^dc{\"U}
+\gdef^^dd{\'Y}
+\gdef^^de{\missingcharmsg{LATIN CAPITAL LETTER THORN}}
+\gdef^^df{\ss}
+%
+\gdef^^e0{\`a}
+\gdef^^e1{\'a}
+\gdef^^e2{\^a}
+\gdef^^e3{\~a}
+\gdef^^e4{\"a}
+\gdef^^e5{\ringaccent a}
+\gdef^^e6{\ae}
+\gdef^^e7{\cedilla c}
+\gdef^^e8{\`e}
+\gdef^^e9{\'e}
+\gdef^^ea{\^e}
+\gdef^^eb{\"e}
+\gdef^^ec{\`{\dotless i}}
+\gdef^^ed{\'{\dotless i}}
+\gdef^^ee{\^{\dotless i}}
+\gdef^^ef{\"{\dotless i}}
+%
+\gdef^^f0{\missingcharmsg{LATIN SMALL LETTER ETH}}
+\gdef^^f1{\~n}
+\gdef^^f2{\`o}
+\gdef^^f3{\'o}
+\gdef^^f4{\^o}
+\gdef^^f5{\~o}
+\gdef^^f6{\"o}
+\gdef^^f7{$\div$}
+\gdef^^f8{\o}
+\gdef^^f9{\`u}
+\gdef^^fa{\'u}
+\gdef^^fb{\^u}
+\gdef^^fc{\"u}
+\gdef^^fd{\'y}
+\gdef^^fe{\missingcharmsg{LATIN SMALL LETTER THORN}}
+\gdef^^ff{\"y}
+}
+% Latin9 (ISO-8859-15) encoding character definitions.
+\def\latninechardefs{%
+% Encoding is almost identical to Latin1.
+\latonechardefs
+%
+\gdef^^a4{\euro}
+\gdef^^a6{\v S}
+\gdef^^a8{\v s}
+\gdef^^b4{\v Z}
+\gdef^^b8{\v z}
+\gdef^^bc{\OE}
+\gdef^^bd{\oe}
+\gdef^^be{\"Y}
+}
+% Latin2 (ISO-8859-2) character definitions.
+\def\lattwochardefs{%
+\gdef^^a0{~}
+\gdef^^a1{\missingcharmsg{LATIN CAPITAL LETTER A WITH OGONEK}}
+\gdef^^a2{\u{}}
+\gdef^^a3{\L}
+\gdef^^a4{\missingcharmsg{CURRENCY SIGN}}
+\gdef^^a5{\v L}
+\gdef^^a6{\'S}
+\gdef^^a7{\S}
+\gdef^^a8{\"{}}
+\gdef^^a9{\v S}
+\gdef^^aa{\cedilla S}
+\gdef^^ab{\v T}
+\gdef^^ac{\'Z}
+\gdef^^ad{\-}
+\gdef^^ae{\v Z}
+\gdef^^af{\dotaccent Z}
+%
+\gdef^^b0{\textdegree}
+\gdef^^b1{\missingcharmsg{LATIN SMALL LETTER A WITH OGONEK}}
+\gdef^^b2{\missingcharmsg{OGONEK}}
+\gdef^^b3{\l}
+\gdef^^b4{\'{}}
+\gdef^^b5{\v l}
+\gdef^^b6{\'s}
+\gdef^^b7{\v{}}
+\gdef^^b8{\cedilla\ }
+\gdef^^b9{\v s}
+\gdef^^ba{\cedilla s}
+\gdef^^bb{\v t}
+\gdef^^bc{\'z}
+\gdef^^bd{\H{}}
+\gdef^^be{\v z}
+\gdef^^bf{\dotaccent z}
+%
+\gdef^^c0{\'R}
+\gdef^^c1{\'A}
+\gdef^^c2{\^A}
+\gdef^^c3{\u A}
+\gdef^^c4{\"A}
+\gdef^^c5{\'L}
+\gdef^^c6{\'C}
+\gdef^^c7{\cedilla C}
+\gdef^^c8{\v C}
+\gdef^^c9{\'E}
+\gdef^^ca{\missingcharmsg{LATIN CAPITAL LETTER E WITH OGONEK}}
+\gdef^^cb{\"E}
+\gdef^^cc{\v E}
+\gdef^^cd{\'I}
+\gdef^^ce{\^I}
+\gdef^^cf{\v D}
+%
+\gdef^^d0{\missingcharmsg{LATIN CAPITAL LETTER D WITH STROKE}}
+\gdef^^d1{\'N}
+\gdef^^d2{\v N}
+\gdef^^d3{\'O}
+\gdef^^d4{\^O}
+\gdef^^d5{\H O}
+\gdef^^d6{\"O}
+\gdef^^d7{$\times$}
+\gdef^^d8{\v R}
+\gdef^^d9{\ringaccent U}
+\gdef^^da{\'U}
+\gdef^^db{\H U}
+\gdef^^dc{\"U}
+\gdef^^dd{\'Y}
+\gdef^^de{\cedilla T}
+\gdef^^df{\ss}
+%
+\gdef^^e0{\'r}
+\gdef^^e1{\'a}
+\gdef^^e2{\^a}
+\gdef^^e3{\u a}
+\gdef^^e4{\"a}
+\gdef^^e5{\'l}
+\gdef^^e6{\'c}
+\gdef^^e7{\cedilla c}
+\gdef^^e8{\v c}
+\gdef^^e9{\'e}
+\gdef^^ea{\missingcharmsg{LATIN SMALL LETTER E WITH OGONEK}}
+\gdef^^eb{\"e}
+\gdef^^ec{\v e}
+\gdef^^ed{\'\i}
+\gdef^^ee{\^\i}
+\gdef^^ef{\v d}
+%
+\gdef^^f0{\missingcharmsg{LATIN SMALL LETTER D WITH STROKE}}
+\gdef^^f1{\'n}
+\gdef^^f2{\v n}
+\gdef^^f3{\'o}
+\gdef^^f4{\^o}
+\gdef^^f5{\H o}
+\gdef^^f6{\"o}
+\gdef^^f7{$\div$}
+\gdef^^f8{\v r}
+\gdef^^f9{\ringaccent u}
+\gdef^^fa{\'u}
+\gdef^^fb{\H u}
+\gdef^^fc{\"u}
+\gdef^^fd{\'y}
+\gdef^^fe{\cedilla t}
+\gdef^^ff{\dotaccent{}}
+}
+% UTF-8 character definitions.
+%
+% This code to support UTF-8 is based on LaTeX's utf8.def, with some
+% changes for Texinfo conventions.  It is included here under the GPL by
+% permission from Frank Mittelbach and the LaTeX team.
+%
+\newcount\countUTFx
+\newcount\countUTFy
+\newcount\countUTFz
+\gdef\UTFviiiTwoOctets#1#2{\expandafter
+\UTFviiiDefined\csname u8:#1\string #2\endcsname}
+%
+\gdef\UTFviiiThreeOctets#1#2#3{\expandafter
+\UTFviiiDefined\csname u8:#1\string #2\string #3\endcsname}
+%
+\gdef\UTFviiiFourOctets#1#2#3#4{\expandafter
+\UTFviiiDefined\csname u8:#1\string #2\string #3\string #4\endcsname}
+\gdef\UTFviiiDefined#1{%
+\ifx #1\relax
+\message{\linenumber Unicode char \string #1 not defined for Texinfo}%
+\else
+\expandafter #1%
+\fi
+}
+\begingroup
+\catcode`\~13
+\catcode`\"12
+\def\UTFviiiLoop{%
+\global\catcode\countUTFx\active
+\uccode`\~\countUTFx
+\uppercase\expandafter{\UTFviiiTmp}%
+\advance\countUTFx by 1
+\ifnum\countUTFx < \countUTFy
+\expandafter\UTFviiiLoop
+\fi}
+\countUTFx = "C2
+\countUTFy = "E0
+\def\UTFviiiTmp{%
+\xdef~{\noexpand\UTFviiiTwoOctets\string~}}
+\UTFviiiLoop
+\countUTFx = "E0
+\countUTFy = "F0
+\def\UTFviiiTmp{%
+\xdef~{\noexpand\UTFviiiThreeOctets\string~}}
+\UTFviiiLoop
+\countUTFx = "F0
+\countUTFy = "F4
+\def\UTFviiiTmp{%
+\xdef~{\noexpand\UTFviiiFourOctets\string~}}
+\UTFviiiLoop
+\endgroup
+\begingroup
+\catcode`\"=12
+\catcode`\<=12
+\catcode`\.=12
+\catcode`\,=12
+\catcode`\;=12
+\catcode`\!=12
+\catcode`\~=13
+\gdef\DeclareUnicodeCharacter#1#2{%
+\countUTFz = "#1\relax
+\wlog{\space\space defining Unicode char U+#1 (decimal \the\countUTFz)}%
+\begingroup
+\parseXMLCharref
+\def\UTFviiiTwoOctets##1##2{%
+\csname u8:##1\string ##2\endcsname}%
+\def\UTFviiiThreeOctets##1##2##3{%
+\csname u8:##1\string ##2\string ##3\endcsname}%
+\def\UTFviiiFourOctets##1##2##3##4{%
+\csname u8:##1\string ##2\string ##3\string ##4\endcsname}%
+\expandafter\expandafter\expandafter\expandafter
+\expandafter\expandafter\expandafter
+\gdef\UTFviiiTmp{#2}%
+\endgroup}
+\gdef\parseXMLCharref{%
+\ifnum\countUTFz < "A0\relax
+\errhelp = \EMsimple
+\errmessage{Cannot define Unicode char value < 00A0}%
+\else\ifnum\countUTFz < "800\relax
+\parseUTFviiiA,%
+\parseUTFviiiB C\UTFviiiTwoOctets.,%
+\else\ifnum\countUTFz < "10000\relax
+\parseUTFviiiA;%
+\parseUTFviiiA,%
+\parseUTFviiiB E\UTFviiiThreeOctets.{,;}%
+\else
+\parseUTFviiiA;%
+\parseUTFviiiA,%
+\parseUTFviiiA!%
+\parseUTFviiiB F\UTFviiiFourOctets.{!,;}%
+\fi\fi\fi
+}
+\gdef\parseUTFviiiA#1{%
+\countUTFx = \countUTFz
+\divide\countUTFz by 64
+\countUTFy = \countUTFz
+\multiply\countUTFz by 64
+\advance\countUTFx by -\countUTFz
+\advance\countUTFx by 128
+\uccode `#1\countUTFx
+\countUTFz = \countUTFy}
+\gdef\parseUTFviiiB#1#2#3#4{%
+\advance\countUTFz by "#10\relax
+\uccode `#3\countUTFz
+\uppercase{\gdef\UTFviiiTmp{#2#3#4}}}
+\endgroup
+\def\utfeightchardefs{%
+\DeclareUnicodeCharacter{00A0}{\tie}
+\DeclareUnicodeCharacter{00A1}{\exclamdown}
+\DeclareUnicodeCharacter{00A3}{\pounds}
+\DeclareUnicodeCharacter{00A8}{\"{ }}
+\DeclareUnicodeCharacter{00A9}{\copyright}
+\DeclareUnicodeCharacter{00AA}{\ordf}
+\DeclareUnicodeCharacter{00AB}{\guillemetleft}
+\DeclareUnicodeCharacter{00AD}{\-}
+\DeclareUnicodeCharacter{00AE}{\registeredsymbol}
+\DeclareUnicodeCharacter{00AF}{\={ }}
+\DeclareUnicodeCharacter{00B0}{\ringaccent{ }}
+\DeclareUnicodeCharacter{00B4}{\'{ }}
+\DeclareUnicodeCharacter{00B8}{\cedilla{ }}
+\DeclareUnicodeCharacter{00BA}{\ordm}
+\DeclareUnicodeCharacter{00BB}{\guillemetright}
+\DeclareUnicodeCharacter{00BF}{\questiondown}
+\DeclareUnicodeCharacter{00C0}{\`A}
+\DeclareUnicodeCharacter{00C1}{\'A}
+\DeclareUnicodeCharacter{00C2}{\^A}
+\DeclareUnicodeCharacter{00C3}{\~A}
+\DeclareUnicodeCharacter{00C4}{\"A}
+\DeclareUnicodeCharacter{00C5}{\AA}
+\DeclareUnicodeCharacter{00C6}{\AE}
+\DeclareUnicodeCharacter{00C7}{\cedilla{C}}
+\DeclareUnicodeCharacter{00C8}{\`E}
+\DeclareUnicodeCharacter{00C9}{\'E}
+\DeclareUnicodeCharacter{00CA}{\^E}
+\DeclareUnicodeCharacter{00CB}{\"E}
+\DeclareUnicodeCharacter{00CC}{\`I}
+\DeclareUnicodeCharacter{00CD}{\'I}
+\DeclareUnicodeCharacter{00CE}{\^I}
+\DeclareUnicodeCharacter{00CF}{\"I}
+\DeclareUnicodeCharacter{00D1}{\~N}
+\DeclareUnicodeCharacter{00D2}{\`O}
+\DeclareUnicodeCharacter{00D3}{\'O}
+\DeclareUnicodeCharacter{00D4}{\^O}
+\DeclareUnicodeCharacter{00D5}{\~O}
+\DeclareUnicodeCharacter{00D6}{\"O}
+\DeclareUnicodeCharacter{00D8}{\O}
+\DeclareUnicodeCharacter{00D9}{\`U}
+\DeclareUnicodeCharacter{00DA}{\'U}
+\DeclareUnicodeCharacter{00DB}{\^U}
+\DeclareUnicodeCharacter{00DC}{\"U}
+\DeclareUnicodeCharacter{00DD}{\'Y}
+\DeclareUnicodeCharacter{00DF}{\ss}
+\DeclareUnicodeCharacter{00E0}{\`a}
+\DeclareUnicodeCharacter{00E1}{\'a}
+\DeclareUnicodeCharacter{00E2}{\^a}
+\DeclareUnicodeCharacter{00E3}{\~a}
+\DeclareUnicodeCharacter{00E4}{\"a}
+\DeclareUnicodeCharacter{00E5}{\aa}
+\DeclareUnicodeCharacter{00E6}{\ae}
+\DeclareUnicodeCharacter{00E7}{\cedilla{c}}
+\DeclareUnicodeCharacter{00E8}{\`e}
+\DeclareUnicodeCharacter{00E9}{\'e}
+\DeclareUnicodeCharacter{00EA}{\^e}
+\DeclareUnicodeCharacter{00EB}{\"e}
+\DeclareUnicodeCharacter{00EC}{\`{\dotless{i}}}
+\DeclareUnicodeCharacter{00ED}{\'{\dotless{i}}}
+\DeclareUnicodeCharacter{00EE}{\^{\dotless{i}}}
+\DeclareUnicodeCharacter{00EF}{\"{\dotless{i}}}
+\DeclareUnicodeCharacter{00F1}{\~n}
+\DeclareUnicodeCharacter{00F2}{\`o}
+\DeclareUnicodeCharacter{00F3}{\'o}
+\DeclareUnicodeCharacter{00F4}{\^o}
+\DeclareUnicodeCharacter{00F5}{\~o}
+\DeclareUnicodeCharacter{00F6}{\"o}
+\DeclareUnicodeCharacter{00F8}{\o}
+\DeclareUnicodeCharacter{00F9}{\`u}
+\DeclareUnicodeCharacter{00FA}{\'u}
+\DeclareUnicodeCharacter{00FB}{\^u}
+\DeclareUnicodeCharacter{00FC}{\"u}
+\DeclareUnicodeCharacter{00FD}{\'y}
+\DeclareUnicodeCharacter{00FF}{\"y}
+\DeclareUnicodeCharacter{0100}{\=A}
+\DeclareUnicodeCharacter{0101}{\=a}
+\DeclareUnicodeCharacter{0102}{\u{A}}
+\DeclareUnicodeCharacter{0103}{\u{a}}
+\DeclareUnicodeCharacter{0106}{\'C}
+\DeclareUnicodeCharacter{0107}{\'c}
+\DeclareUnicodeCharacter{0108}{\^C}
+\DeclareUnicodeCharacter{0109}{\^c}
+\DeclareUnicodeCharacter{010A}{\dotaccent{C}}
+\DeclareUnicodeCharacter{010B}{\dotaccent{c}}
+\DeclareUnicodeCharacter{010C}{\v{C}}
+\DeclareUnicodeCharacter{010D}{\v{c}}
+\DeclareUnicodeCharacter{010E}{\v{D}}
+\DeclareUnicodeCharacter{0112}{\=E}
+\DeclareUnicodeCharacter{0113}{\=e}
+\DeclareUnicodeCharacter{0114}{\u{E}}
+\DeclareUnicodeCharacter{0115}{\u{e}}
+\DeclareUnicodeCharacter{0116}{\dotaccent{E}}
+\DeclareUnicodeCharacter{0117}{\dotaccent{e}}
+\DeclareUnicodeCharacter{011A}{\v{E}}
+\DeclareUnicodeCharacter{011B}{\v{e}}
+\DeclareUnicodeCharacter{011C}{\^G}
+\DeclareUnicodeCharacter{011D}{\^g}
+\DeclareUnicodeCharacter{011E}{\u{G}}
+\DeclareUnicodeCharacter{011F}{\u{g}}
+\DeclareUnicodeCharacter{0120}{\dotaccent{G}}
+\DeclareUnicodeCharacter{0121}{\dotaccent{g}}
+\DeclareUnicodeCharacter{0124}{\^H}
+\DeclareUnicodeCharacter{0125}{\^h}
+\DeclareUnicodeCharacter{0128}{\~I}
+\DeclareUnicodeCharacter{0129}{\~{\dotless{i}}}
+\DeclareUnicodeCharacter{012A}{\=I}
+\DeclareUnicodeCharacter{012B}{\={\dotless{i}}}
+\DeclareUnicodeCharacter{012C}{\u{I}}
+\DeclareUnicodeCharacter{012D}{\u{\dotless{i}}}
+\DeclareUnicodeCharacter{0130}{\dotaccent{I}}
+\DeclareUnicodeCharacter{0131}{\dotless{i}}
+\DeclareUnicodeCharacter{0132}{IJ}
+\DeclareUnicodeCharacter{0133}{ij}
+\DeclareUnicodeCharacter{0134}{\^J}
+\DeclareUnicodeCharacter{0135}{\^{\dotless{j}}}
+\DeclareUnicodeCharacter{0139}{\'L}
+\DeclareUnicodeCharacter{013A}{\'l}
+\DeclareUnicodeCharacter{0141}{\L}
+\DeclareUnicodeCharacter{0142}{\l}
+\DeclareUnicodeCharacter{0143}{\'N}
+\DeclareUnicodeCharacter{0144}{\'n}
+\DeclareUnicodeCharacter{0147}{\v{N}}
+\DeclareUnicodeCharacter{0148}{\v{n}}
+\DeclareUnicodeCharacter{014C}{\=O}
+\DeclareUnicodeCharacter{014D}{\=o}
+\DeclareUnicodeCharacter{014E}{\u{O}}
+\DeclareUnicodeCharacter{014F}{\u{o}}
+\DeclareUnicodeCharacter{0150}{\H{O}}
+\DeclareUnicodeCharacter{0151}{\H{o}}
+\DeclareUnicodeCharacter{0152}{\OE}
+\DeclareUnicodeCharacter{0153}{\oe}
+\DeclareUnicodeCharacter{0154}{\'R}
+\DeclareUnicodeCharacter{0155}{\'r}
+\DeclareUnicodeCharacter{0158}{\v{R}}
+\DeclareUnicodeCharacter{0159}{\v{r}}
+\DeclareUnicodeCharacter{015A}{\'S}
+\DeclareUnicodeCharacter{015B}{\'s}
+\DeclareUnicodeCharacter{015C}{\^S}
+\DeclareUnicodeCharacter{015D}{\^s}
+\DeclareUnicodeCharacter{015E}{\cedilla{S}}
+\DeclareUnicodeCharacter{015F}{\cedilla{s}}
+\DeclareUnicodeCharacter{0160}{\v{S}}
+\DeclareUnicodeCharacter{0161}{\v{s}}
+\DeclareUnicodeCharacter{0162}{\cedilla{t}}
+\DeclareUnicodeCharacter{0163}{\cedilla{T}}
+\DeclareUnicodeCharacter{0164}{\v{T}}
+\DeclareUnicodeCharacter{0168}{\~U}
+\DeclareUnicodeCharacter{0169}{\~u}
+\DeclareUnicodeCharacter{016A}{\=U}
+\DeclareUnicodeCharacter{016B}{\=u}
+\DeclareUnicodeCharacter{016C}{\u{U}}
+\DeclareUnicodeCharacter{016D}{\u{u}}
+\DeclareUnicodeCharacter{016E}{\ringaccent{U}}
+\DeclareUnicodeCharacter{016F}{\ringaccent{u}}
+\DeclareUnicodeCharacter{0170}{\H{U}}
+\DeclareUnicodeCharacter{0171}{\H{u}}
+\DeclareUnicodeCharacter{0174}{\^W}
+\DeclareUnicodeCharacter{0175}{\^w}
+\DeclareUnicodeCharacter{0176}{\^Y}
+\DeclareUnicodeCharacter{0177}{\^y}
+\DeclareUnicodeCharacter{0178}{\"Y}
+\DeclareUnicodeCharacter{0179}{\'Z}
+\DeclareUnicodeCharacter{017A}{\'z}
+\DeclareUnicodeCharacter{017B}{\dotaccent{Z}}
+\DeclareUnicodeCharacter{017C}{\dotaccent{z}}
+\DeclareUnicodeCharacter{017D}{\v{Z}}
+\DeclareUnicodeCharacter{017E}{\v{z}}
+\DeclareUnicodeCharacter{01C4}{D\v{Z}}
+\DeclareUnicodeCharacter{01C5}{D\v{z}}
+\DeclareUnicodeCharacter{01C6}{d\v{z}}
+\DeclareUnicodeCharacter{01C7}{LJ}
+\DeclareUnicodeCharacter{01C8}{Lj}
+\DeclareUnicodeCharacter{01C9}{lj}
+\DeclareUnicodeCharacter{01CA}{NJ}
+\DeclareUnicodeCharacter{01CB}{Nj}
+\DeclareUnicodeCharacter{01CC}{nj}
+\DeclareUnicodeCharacter{01CD}{\v{A}}
+\DeclareUnicodeCharacter{01CE}{\v{a}}
+\DeclareUnicodeCharacter{01CF}{\v{I}}
+\DeclareUnicodeCharacter{01D0}{\v{\dotless{i}}}
+\DeclareUnicodeCharacter{01D1}{\v{O}}
+\DeclareUnicodeCharacter{01D2}{\v{o}}
+\DeclareUnicodeCharacter{01D3}{\v{U}}
+\DeclareUnicodeCharacter{01D4}{\v{u}}
+\DeclareUnicodeCharacter{01E2}{\={\AE}}
+\DeclareUnicodeCharacter{01E3}{\={\ae}}
+\DeclareUnicodeCharacter{01E6}{\v{G}}
+\DeclareUnicodeCharacter{01E7}{\v{g}}
+\DeclareUnicodeCharacter{01E8}{\v{K}}
+\DeclareUnicodeCharacter{01E9}{\v{k}}
+\DeclareUnicodeCharacter{01F0}{\v{\dotless{j}}}
+\DeclareUnicodeCharacter{01F1}{DZ}
+\DeclareUnicodeCharacter{01F2}{Dz}
+\DeclareUnicodeCharacter{01F3}{dz}
+\DeclareUnicodeCharacter{01F4}{\'G}
+\DeclareUnicodeCharacter{01F5}{\'g}
+\DeclareUnicodeCharacter{01F8}{\`N}
+\DeclareUnicodeCharacter{01F9}{\`n}
+\DeclareUnicodeCharacter{01FC}{\'{\AE}}
+\DeclareUnicodeCharacter{01FD}{\'{\ae}}
+\DeclareUnicodeCharacter{01FE}{\'{\O}}
+\DeclareUnicodeCharacter{01FF}{\'{\o}}
+\DeclareUnicodeCharacter{021E}{\v{H}}
+\DeclareUnicodeCharacter{021F}{\v{h}}
+\DeclareUnicodeCharacter{0226}{\dotaccent{A}}
+\DeclareUnicodeCharacter{0227}{\dotaccent{a}}
+\DeclareUnicodeCharacter{0228}{\cedilla{E}}
+\DeclareUnicodeCharacter{0229}{\cedilla{e}}
+\DeclareUnicodeCharacter{022E}{\dotaccent{O}}
+\DeclareUnicodeCharacter{022F}{\dotaccent{o}}
+\DeclareUnicodeCharacter{0232}{\=Y}
+\DeclareUnicodeCharacter{0233}{\=y}
+\DeclareUnicodeCharacter{0237}{\dotless{j}}
+\DeclareUnicodeCharacter{1E02}{\dotaccent{B}}
+\DeclareUnicodeCharacter{1E03}{\dotaccent{b}}
+\DeclareUnicodeCharacter{1E04}{\udotaccent{B}}
+\DeclareUnicodeCharacter{1E05}{\udotaccent{b}}
+\DeclareUnicodeCharacter{1E06}{\ubaraccent{B}}
+\DeclareUnicodeCharacter{1E07}{\ubaraccent{b}}
+\DeclareUnicodeCharacter{1E0A}{\dotaccent{D}}
+\DeclareUnicodeCharacter{1E0B}{\dotaccent{d}}
+\DeclareUnicodeCharacter{1E0C}{\udotaccent{D}}
+\DeclareUnicodeCharacter{1E0D}{\udotaccent{d}}
+\DeclareUnicodeCharacter{1E0E}{\ubaraccent{D}}
+\DeclareUnicodeCharacter{1E0F}{\ubaraccent{d}}
+\DeclareUnicodeCharacter{1E1E}{\dotaccent{F}}
+\DeclareUnicodeCharacter{1E1F}{\dotaccent{f}}
+\DeclareUnicodeCharacter{1E20}{\=G}
+\DeclareUnicodeCharacter{1E21}{\=g}
+\DeclareUnicodeCharacter{1E22}{\dotaccent{H}}
+\DeclareUnicodeCharacter{1E23}{\dotaccent{h}}
+\DeclareUnicodeCharacter{1E24}{\udotaccent{H}}
+\DeclareUnicodeCharacter{1E25}{\udotaccent{h}}
+\DeclareUnicodeCharacter{1E26}{\"H}
+\DeclareUnicodeCharacter{1E27}{\"h}
+\DeclareUnicodeCharacter{1E30}{\'K}
+\DeclareUnicodeCharacter{1E31}{\'k}
+\DeclareUnicodeCharacter{1E32}{\udotaccent{K}}
+\DeclareUnicodeCharacter{1E33}{\udotaccent{k}}
+\DeclareUnicodeCharacter{1E34}{\ubaraccent{K}}
+\DeclareUnicodeCharacter{1E35}{\ubaraccent{k}}
+\DeclareUnicodeCharacter{1E36}{\udotaccent{L}}
+\DeclareUnicodeCharacter{1E37}{\udotaccent{l}}
+\DeclareUnicodeCharacter{1E3A}{\ubaraccent{L}}
+\DeclareUnicodeCharacter{1E3B}{\ubaraccent{l}}
+\DeclareUnicodeCharacter{1E3E}{\'M}
+\DeclareUnicodeCharacter{1E3F}{\'m}
+\DeclareUnicodeCharacter{1E40}{\dotaccent{M}}
+\DeclareUnicodeCharacter{1E41}{\dotaccent{m}}
+\DeclareUnicodeCharacter{1E42}{\udotaccent{M}}
+\DeclareUnicodeCharacter{1E43}{\udotaccent{m}}
+\DeclareUnicodeCharacter{1E44}{\dotaccent{N}}
+\DeclareUnicodeCharacter{1E45}{\dotaccent{n}}
+\DeclareUnicodeCharacter{1E46}{\udotaccent{N}}
+\DeclareUnicodeCharacter{1E47}{\udotaccent{n}}
+\DeclareUnicodeCharacter{1E48}{\ubaraccent{N}}
+\DeclareUnicodeCharacter{1E49}{\ubaraccent{n}}
+\DeclareUnicodeCharacter{1E54}{\'P}
+\DeclareUnicodeCharacter{1E55}{\'p}
+\DeclareUnicodeCharacter{1E56}{\dotaccent{P}}
+\DeclareUnicodeCharacter{1E57}{\dotaccent{p}}
+\DeclareUnicodeCharacter{1E58}{\dotaccent{R}}
+\DeclareUnicodeCharacter{1E59}{\dotaccent{r}}
+\DeclareUnicodeCharacter{1E5A}{\udotaccent{R}}
+\DeclareUnicodeCharacter{1E5B}{\udotaccent{r}}
+\DeclareUnicodeCharacter{1E5E}{\ubaraccent{R}}
+\DeclareUnicodeCharacter{1E5F}{\ubaraccent{r}}
+\DeclareUnicodeCharacter{1E60}{\dotaccent{S}}
+\DeclareUnicodeCharacter{1E61}{\dotaccent{s}}
+\DeclareUnicodeCharacter{1E62}{\udotaccent{S}}
+\DeclareUnicodeCharacter{1E63}{\udotaccent{s}}
+\DeclareUnicodeCharacter{1E6A}{\dotaccent{T}}
+\DeclareUnicodeCharacter{1E6B}{\dotaccent{t}}
+\DeclareUnicodeCharacter{1E6C}{\udotaccent{T}}
+\DeclareUnicodeCharacter{1E6D}{\udotaccent{t}}
+\DeclareUnicodeCharacter{1E6E}{\ubaraccent{T}}
+\DeclareUnicodeCharacter{1E6F}{\ubaraccent{t}}
+\DeclareUnicodeCharacter{1E7C}{\~V}
+\DeclareUnicodeCharacter{1E7D}{\~v}
+\DeclareUnicodeCharacter{1E7E}{\udotaccent{V}}
+\DeclareUnicodeCharacter{1E7F}{\udotaccent{v}}
+\DeclareUnicodeCharacter{1E80}{\`W}
+\DeclareUnicodeCharacter{1E81}{\`w}
+\DeclareUnicodeCharacter{1E82}{\'W}
+\DeclareUnicodeCharacter{1E83}{\'w}
+\DeclareUnicodeCharacter{1E84}{\"W}
+\DeclareUnicodeCharacter{1E85}{\"w}
+\DeclareUnicodeCharacter{1E86}{\dotaccent{W}}
+\DeclareUnicodeCharacter{1E87}{\dotaccent{w}}
+\DeclareUnicodeCharacter{1E88}{\udotaccent{W}}
+\DeclareUnicodeCharacter{1E89}{\udotaccent{w}}
+\DeclareUnicodeCharacter{1E8A}{\dotaccent{X}}
+\DeclareUnicodeCharacter{1E8B}{\dotaccent{x}}
+\DeclareUnicodeCharacter{1E8C}{\"X}
+\DeclareUnicodeCharacter{1E8D}{\"x}
+\DeclareUnicodeCharacter{1E8E}{\dotaccent{Y}}
+\DeclareUnicodeCharacter{1E8F}{\dotaccent{y}}
+\DeclareUnicodeCharacter{1E90}{\^Z}
+\DeclareUnicodeCharacter{1E91}{\^z}
+\DeclareUnicodeCharacter{1E92}{\udotaccent{Z}}
+\DeclareUnicodeCharacter{1E93}{\udotaccent{z}}
+\DeclareUnicodeCharacter{1E94}{\ubaraccent{Z}}
+\DeclareUnicodeCharacter{1E95}{\ubaraccent{z}}
+\DeclareUnicodeCharacter{1E96}{\ubaraccent{h}}
+\DeclareUnicodeCharacter{1E97}{\"t}
+\DeclareUnicodeCharacter{1E98}{\ringaccent{w}}
+\DeclareUnicodeCharacter{1E99}{\ringaccent{y}}
+\DeclareUnicodeCharacter{1EA0}{\udotaccent{A}}
+\DeclareUnicodeCharacter{1EA1}{\udotaccent{a}}
+\DeclareUnicodeCharacter{1EB8}{\udotaccent{E}}
+\DeclareUnicodeCharacter{1EB9}{\udotaccent{e}}
+\DeclareUnicodeCharacter{1EBC}{\~E}
+\DeclareUnicodeCharacter{1EBD}{\~e}
+\DeclareUnicodeCharacter{1ECA}{\udotaccent{I}}
+\DeclareUnicodeCharacter{1ECB}{\udotaccent{i}}
+\DeclareUnicodeCharacter{1ECC}{\udotaccent{O}}
+\DeclareUnicodeCharacter{1ECD}{\udotaccent{o}}
+\DeclareUnicodeCharacter{1EE4}{\udotaccent{U}}
+\DeclareUnicodeCharacter{1EE5}{\udotaccent{u}}
+\DeclareUnicodeCharacter{1EF2}{\`Y}
+\DeclareUnicodeCharacter{1EF3}{\`y}
+\DeclareUnicodeCharacter{1EF4}{\udotaccent{Y}}
+\DeclareUnicodeCharacter{1EF8}{\~Y}
+\DeclareUnicodeCharacter{1EF9}{\~y}
+\DeclareUnicodeCharacter{2013}{--}
+\DeclareUnicodeCharacter{2014}{---}
+\DeclareUnicodeCharacter{2018}{\quoteleft}
+\DeclareUnicodeCharacter{2019}{\quoteright}
+\DeclareUnicodeCharacter{201A}{\quotesinglbase}
+\DeclareUnicodeCharacter{201C}{\quotedblleft}
+\DeclareUnicodeCharacter{201D}{\quotedblright}
+\DeclareUnicodeCharacter{201E}{\quotedblbase}
+\DeclareUnicodeCharacter{2022}{\bullet}
+\DeclareUnicodeCharacter{2026}{\dots}
+\DeclareUnicodeCharacter{2039}{\guilsinglleft}
+\DeclareUnicodeCharacter{203A}{\guilsinglright}
+\DeclareUnicodeCharacter{20AC}{\euro}
+\DeclareUnicodeCharacter{2192}{\expansion}
+\DeclareUnicodeCharacter{21D2}{\result}
+\DeclareUnicodeCharacter{2212}{\minus}
+\DeclareUnicodeCharacter{2217}{\point}
+\DeclareUnicodeCharacter{2261}{\equiv}
+}% end of \utfeightchardefs
+% US-ASCII character definitions.
+\def\asciichardefs{% nothing need be done
+\relax
+}
+% Make non-ASCII characters printable again for compatibility with
+% existing Texinfo documents that may use them, even without declaring a
+% document encoding.
+%
+\setnonasciicharscatcode \other
+\message{formatting,}
 \newdimen\defaultparindent \defaultparindent = 15pt
 \chapheadingskip = 15pt plus 4pt minus 2pt
 \secheadingskip = 12pt plus 3pt minus 2pt
 \subsecheadingskip = 9pt plus 2pt minus 2pt
 \vbadness = 10000
 % Don't be so finicky about underfull hboxes, either.
 \hbadness = 2000
-% Following George Bush, just get rid of widows and orphans.
+% Following George Bush, get rid of widows and orphans.
 \widowpenalty=10000
 \clubpenalty=10000
 % Use TeX 3.0's \emergencystretch to help line breaking, but if we're
 % using an old version of TeX, don't do anything.  We want the amount of
 \else
 \emergencystretch = .15\hsize
 \fi
 }
-% Parameters in order: 1) textheight; 2) textwidth; 3) voffset;
+% Parameters in order: 1) textheight; 2) textwidth;
-% 4) hoffset; 5) binding offset; 6) topskip.  Then whoever calls us can
+% 3) voffset; 4) hoffset; 5) binding offset; 6) topskip;
-% set \parskip and call \setleading for \baselineskip.
+% 7) physical page height; 8) physical page width.
 %
-\def\internalpagesizes#1#2#3#4#5#6{%
+% We also call \setleading{\textleading}, so the caller should define
+% \textleading.  The caller should also set \parskip.
+%
+\def\internalpagesizes#1#2#3#4#5#6#7#8{%
 \voffset = #3\relax
 \topskip = #6\relax
 \splittopskip = \topskip
 %
 \vsize = #1\relax
 \pagewidth = \hsize
 %
 \normaloffset = #4\relax
 \bindingoffset = #5\relax
 %
+\ifpdf
+\pdfpageheight #7\relax
+\pdfpagewidth #8\relax
+% if we don't reset these, they will remain at "1 true in" of
+% whatever layout pdftex was dumped with.
+\pdfhorigin = 1 true in
+\pdfvorigin = 1 true in
+\fi
+%
+\setleading{\textleading}
+%
 \parindent = \defaultparindent
 \setemergencystretch
 }
 % @letterpaper (the default).
 \def\letterpaper{{\globaldefs = 1
 \parskip = 3pt plus 2pt minus 1pt
-\setleading{13.2pt}%
+\textleading = 13.2pt
 %
 % If page is nothing but text, make it come out even.
-\internalpagesizes{46\baselineskip}{6in}{\voffset}{.25in}{\bindingoffset}{36pt}%
+\internalpagesizes{607.2pt}{6in}% that's 46 lines
+{\voffset}{.25in}%
+{\bindingoffset}{36pt}%
+{11in}{8.5in}%
 }}
-% Use @smallbook to reset parameters for 7x9.5 (or so) format.
+% Use @smallbook to reset parameters for 7x9.25 trim size.
 \def\smallbook{{\globaldefs = 1
 \parskip = 2pt plus 1pt
-\setleading{12pt}%
+\textleading = 12pt
 %
-\internalpagesizes{7.5in}{5.in}{\voffset}{.25in}{\bindingoffset}{16pt}%
+\internalpagesizes{7.5in}{5in}%
+{-.2in}{0in}%
+{\bindingoffset}{16pt}%
+{9.25in}{7in}%
 %
 \lispnarrowing = 0.3in
 \tolerance = 700
 \hfuzz = 1pt
 \contentsrightmargin = 0pt
-\deftypemargin = 0pt
 \defbodyindent = .5cm
-%
+}}
-\let\smalldisplay = \smalldisplayx
-\let\smallexample = \smalllispx
+% Use @smallerbook to reset parameters for 6x9 trim size.
-\let\smallformat = \smallformatx
+% (Just testing, parameters still in flux.)
-\let\smalllisp = \smalllispx
+\def\smallerbook{{\globaldefs = 1
+\parskip = 1.5pt plus 1pt
+\textleading = 12pt
+%
+\internalpagesizes{7.4in}{4.8in}%
+{-.2in}{-.4in}%
+{0pt}{14pt}%
+{9in}{6in}%
+%
+\lispnarrowing = 0.25in
+\tolerance = 700
+\hfuzz = 1pt
+\contentsrightmargin = 0pt
+\defbodyindent = .4cm
 }}
 % Use @afourpaper to print on European A4 paper.
 \def\afourpaper{{\globaldefs = 1
-\setleading{12pt}%
 \parskip = 3pt plus 2pt minus 1pt
-%
+\textleading = 13.2pt
-\internalpagesizes{53\baselineskip}{160mm}{\voffset}{4mm}{\bindingoffset}{44pt}%
+%
+% Double-side printing via postscript on Laserjet 4050
+% prints double-sided nicely when \bindingoffset=10mm and \hoffset=-6mm.
+% To change the settings for a different printer or situation, adjust
+% \normaloffset until the front-side and back-side texts align.  Then
+% do the same for \bindingoffset.  You can set these for testing in
+% your texinfo source file like this:
+% @tex
+% \global\normaloffset = -6mm
+% \global\bindingoffset = 10mm
+% @end tex
+\internalpagesizes{673.2pt}{160mm}% that's 51 lines
+{\voffset}{\hoffset}%
+{\bindingoffset}{44pt}%
+{297mm}{210mm}%
 %
 \tolerance = 700
 \hfuzz = 1pt
+\contentsrightmargin = 0pt
+\defbodyindent = 5mm
 }}
-% A specific text layout, 24x15cm overall, intended for A4 paper.  Top margin
+% Use @afivepaper to print on European A5 paper.
-% 29mm, hence bottom margin 28mm, nominal side margin 3cm.
+% From romildo@urano.iceb.ufop.br, 2 July 2000.
+% He also recommends making @example and @lisp be small.
+\def\afivepaper{{\globaldefs = 1
+\parskip = 2pt plus 1pt minus 0.1pt
+\textleading = 12.5pt
+%
+\internalpagesizes{160mm}{120mm}%
+{\voffset}{\hoffset}%
+{\bindingoffset}{8pt}%
+{210mm}{148mm}%
+%
+\lispnarrowing = 0.2in
+\tolerance = 800
+\hfuzz = 1.2pt
+\contentsrightmargin = 0pt
+\defbodyindent = 2mm
+\tableindent = 12mm
+}}
+% A specific text layout, 24x15cm overall, intended for A4 paper.
 \def\afourlatex{{\globaldefs = 1
-\setleading{13.6pt}%
-%
 \afourpaper
-\internalpagesizes{237mm}{150mm}{3.6mm}{3.6mm}{3mm}{7mm}%
+\internalpagesizes{237mm}{150mm}%
-%
+{\voffset}{4.6mm}%
+{\bindingoffset}{7mm}%
+{297mm}{210mm}%
+%
+% Must explicitly reset to 0 because we call \afourpaper.
 \globaldefs = 0
 }}
-% Use @afourwide to print on European A4 paper in wide format.
+% Use @afourwide to print on A4 paper in landscape format.
-\def\afourwide{%
+\def\afourwide{{\globaldefs = 1
 \afourpaper
-\internalpagesizes{9.5in}{6.5in}{\hoffset}{\normaloffset}{\bindingoffset}{7mm}%
+\internalpagesizes{241mm}{165mm}%
-%
+{\voffset}{-2.95mm}%
+{\bindingoffset}{7mm}%
+{297mm}{210mm}%
 \globaldefs = 0
-}
+}}
 % @pagesizes TEXTHEIGHT[,TEXTWIDTH]
 % Perhaps we should allow setting the margins, \topskip, \parskip,
 % and/or leading, also. Or perhaps we should compute them somehow.
 %
-\def\pagesizes{\parsearg\pagesizesxxx}
+\parseargdef\pagesizes{\pagesizesyyy #1,,\finish}
-\def\pagesizesxxx#1{\pagesizesyyy #1,,\finish}
 \def\pagesizesyyy#1,#2,#3\finish{{%
 \setbox0 = \hbox{\ignorespaces #2}\ifdim\wd0 > 0pt \hsize=#2\relax \fi
 \globaldefs = 1
 %
 \parskip = 3pt plus 2pt minus 1pt
-\setleading{13.2pt}%
+\setleading{\textleading}%
 %
-\internalpagesizes{#1}{\hsize}{\voffset}{\normaloffset}{\bindingoffset}{44pt}%
+\dimen0 = #1\relax
+\advance\dimen0 by \voffset
+%
+\dimen2 = \hsize
+\advance\dimen2 by \normaloffset
+%
+\internalpagesizes{#1}{\hsize}%
+{\voffset}{\normaloffset}%
+{\bindingoffset}{44pt}%
+{\dimen0}{\dimen2}%
 }}
 % Set default to letter.
 %
 \letterpaper
 \def\normalunderscore{_}
 \def\normalverticalbar{|}
 \def\normalless{<}
 \def\normalgreater{>}
 \def\normalplus{+}
-\def\normaldollar{$}
+\def\normaldollar{$}%$ font-lock fix
-% This macro is used to make a character print one way in ttfont
+% This macro is used to make a character print one way in \tt
-% where it can probably just be output, and another way in other fonts,
+% (where it can probably be output as-is), and another way in other fonts,
 % where something hairier probably needs to be done.
 %
 % #1 is what to print if we are indeed using \tt; #2 is what to print
 % otherwise.  Since all the Computer Modern typewriter fonts have zero
 % interword stretch (and shrink), and it is reasonable to expect all
 \catcode`\^=\active
 \def^{{\tt \hat}}
 \catcode`\_=\active
 \def_{\ifusingtt\normalunderscore\_}
+\let\realunder=_
 % Subroutine for the previous macro.
-\def\_{\leavevmode \kern.06em \vbox{\hrule width.3em height.1ex}}
+\def\_{\leavevmode \kern.07em \vbox{\hrule width.3em height.1ex}\kern .07em }
 \catcode`\|=\active
 \def|{{\tt\char124}}
 \chardef \less=`\<
 \catcode`\<=\active
 \catcode`\>=\active
 \def>{{\tt \gtr}}
 \catcode`\+=\active
 \def+{{\tt \char 43}}
 \catcode`\$=\active
-\def${\ifusingit{{\sl\$}}\normaldollar}
+\def${\ifusingit{{\sl\$}}\normaldollar}%$ font-lock fix
-%\catcode 27=\active
-%\def^^[{$\diamondsuit$}
-% Set up an active definition for =, but don't enable it most of the time.
-{\catcode`\==\active
-\global\def={{\tt \char 61}}}
-\catcode`+=\active
-\catcode`\_=\active
 % If a .fmt file is being used, characters that might appear in a file
 % name cannot be active until we have parsed the command line.
 % So turn them off again, and have \everyjob (or @setfilename) turn them on.
 % \otherifyactive is called near the end of this file.
 \def\otherifyactive{\catcode`+=\other \catcode`\_=\other}
+% Used sometimes to turn off (effectively) the active characters even after
+% parsing them.
+\def\turnoffactive{%
+\normalturnoffactive
+\otherbackslash
+}
 \catcode`\@=0
-% \rawbackslashxx output one backslash character in current font
+% \backslashcurfont outputs one backslash character in current font,
-\global\chardef\rawbackslashxx=`\\
+% as in \char`\\.
-%{\catcode`\\=\other
+\global\chardef\backslashcurfont=`\\
-%@gdef@rawbackslashxx{\}}
+\global\let\rawbackslashxx=\backslashcurfont  % let existing .??s files work
-% \rawbackslash redefines \ as input to do \rawbackslashxx.
+% \realbackslash is an actual character `\' with catcode other, and
-{\catcode`\\=\active
+% \doublebackslash is two of them (for the pdf outlines).
-@gdef@rawbackslash{@let\=@rawbackslashxx }}
+{\catcode`\\=\other @gdef@realbackslash{\} @gdef@doublebackslash{\\}}
-% \normalbackslash outputs one backslash in fixed width font.
+% In texinfo, backslash is an active character; it prints the backslash
-\def\normalbackslash{{\tt\rawbackslashxx}}
+% in fixed width font.
-% \catcode 17=0   % Define control-q
 \catcode`\\=\active
+@def@normalbackslash{{@tt@backslashcurfont}}
-% Used sometimes to turn off (effectively) the active characters
+% On startup, @fixbackslash assigns:
-% even after parsing them.
+%  @let \ = @normalbackslash
-@def@turnoffactive{@let"=@normaldoublequote
-@let\=@realbackslash
+% \rawbackslash defines an active \ to do \backslashcurfont.
-@let~=@normaltilde
+% \otherbackslash defines an active \ to be a literal `\' character with
-@let^=@normalcaret
+% catcode other.
-@let_=@normalunderscore
+@gdef@rawbackslash{@let\=@backslashcurfont}
-@let|=@normalverticalbar
+@gdef@otherbackslash{@let\=@realbackslash}
-@let<=@normalless
-@let>=@normalgreater
+% Same as @turnoffactive except outputs \ as {\tt\char`\\} instead of
-@let+=@normalplus
+% the literal character `\'.
-@let$=@normaldollar}
+%
+@def@normalturnoffactive{%
-@def@normalturnoffactive{@let"=@normaldoublequote
+@let\=@normalbackslash
-@let\=@normalbackslash
+@let"=@normaldoublequote
 @let~=@normaltilde
 @let^=@normalcaret
 @let_=@normalunderscore
 @let|=@normalverticalbar
 @let<=@normalless
 @let>=@normalgreater
 @let+=@normalplus
-@let$=@normaldollar}
+@let$=@normaldollar %$ font-lock fix
+@unsepspaces
+}
 % Make _ and + \other characters, temporarily.
 % This is canceled by @fixbackslash.
 @otherifyactive
 %
 @gdef@eatinput input texinfo{@fixbackslash}
 @global@let\ = @eatinput
 % On the other hand, perhaps the file did not have a `\input texinfo'. Then
-% the first `\{ in the file would cause an error. This macro tries to fix
+% the first `\' in the file would cause an error. This macro tries to fix
 % that, assuming it is called before the first `\' could plausibly occur.
-% Also back turn on active characters that might appear in the input
+% Also turn back on active characters that might appear in the input
 % file name, in case not using a pre-dumped format.
 %
 @gdef@fixbackslash{%
 @ifx\@eatinput @let\ = @normalbackslash @fi
 @catcode`+=@active
 }
 % Say @foo, not \foo, in error messages.
 @escapechar = `@@
 % These look ok in all fonts, so just make them not special.
 @catcode`@& = @other
 @catcode`@# = @other
 @catcode`@% = @other
-@c Set initial fonts.
-@textfonts
-@rm
 @c Local variables:
 @c eval: (add-hook 'write-file-hooks 'time-stamp)
 @c page-delimiter: "^\\\\message"
 @c time-stamp-start: "def\\\\texinfoversion{"
 @c time-stamp-format: "%:y-%02m-%02d.%02H"
 @c time-stamp-end: "}"
 @c End:
+@c vim:sw=2:
+@ignore
+arch-tag: e1b36e32-c96e-4135-a41a-0b2efa2ea115
+@end ignore

Mercurial > hg > xemacs-beta

comparison man/texinfo.tex @ 5492:e82f5b7010fe