[xemacs-hg @ 2001-09-20 06:28:42 by ben] The great integral types renaming. The purpose of this is to rationalize the names used for various integral types, so that they match their intended uses and follow consist conventions, and eliminate types that were not semantically different from each other. The conventions are: -- All integral types that measure quantities of anything are signed. Some people disagree vociferously with this, but their arguments are mostly theoretical, and are vastly outweighed by the practical headaches of mixing signed and unsigned values, and more importantly by the far increased likelihood of inadvertent bugs: Because of the broken "viral" nature of unsigned quantities in C (operations involving mixed signed/unsigned are done unsigned, when exactly the opposite is nearly always wanted), even a single error in declaring a quantity unsigned that should be signed, or even the even more subtle error of comparing signed and unsigned values and forgetting the necessary cast, can be catastrophic, as comparisons will yield wrong results. -Wsign-compare is turned on specifically to catch this, but this tends to result in a great number of warnings when mixing signed and unsigned, and the casts are annoying. More has been written on this elsewhere. -- All such quantity types just mentioned boil down to EMACS_INT, which is 32 bits on 32-bit machines and 64 bits on 64-bit machines. This is guaranteed to be the same size as Lisp objects of type `int', and (as far as I can tell) of size_t (unsigned!) and ssize_t. The only type below that is not an EMACS_INT is Hashcode, which is an unsigned value of the same size as EMACS_INT. -- Type names should be relatively short (no more than 10 characters or so), with the first letter capitalized and no underscores if they can at all be avoided. -- "count" == a zero-based measurement of some quantity. Includes sizes, offsets, and indexes. -- "bpos" == a one-based measurement of a position in a buffer. "Charbpos" and "Bytebpos" count text in the buffer, rather than bytes in memory; thus Bytebpos does not directly correspond to the memory representation. Use "Membpos" for this. -- "Char" refers to internal-format characters, not to the C type "char", which is really a byte. -- For the actual name changes, see the script below. I ran the following script to do the conversion. (NOTE: This script is idempotent. You can safely run it multiple times and it will not screw up previous results -- in fact, it will do nothing if nothing has changed. Thus, it can be run repeatedly as necessary to handle patches coming in from old workspaces, or old branches.) There are two tags, just before and just after the change: `pre-integral-type-rename' and `post-integral-type-rename'. When merging code from the main trunk into a branch, the best thing to do is first merge up to `pre-integral-type-rename', then apply the script and associated changes, then merge from `post-integral-type-change' to the present. (Alternatively, just do the merging in one operation; but you may then have a lot of conflicts needing to be resolved by hand.) Script `fixtypes.sh' follows: ----------------------------------- cut ------------------------------------ files="*.[ch] s/*.h m/*.h config.h.in ../configure.in Makefile.in.in ../lib-src/*.[ch] ../lwlib/*.[ch]" gr Memory_Count Bytecount $files gr Lstream_Data_Count Bytecount $files gr Element_Count Elemcount $files gr Hash_Code Hashcode $files gr extcount bytecount $files gr bufpos charbpos $files gr bytind bytebpos $files gr memind membpos $files gr bufbyte intbyte $files gr Extcount Bytecount $files gr Bufpos Charbpos $files gr Bytind Bytebpos $files gr Memind Membpos $files gr Bufbyte Intbyte $files gr EXTCOUNT BYTECOUNT $files gr BUFPOS CHARBPOS $files gr BYTIND BYTEBPOS $files gr MEMIND MEMBPOS $files gr BUFBYTE INTBYTE $files gr MEMORY_COUNT BYTECOUNT $files gr LSTREAM_DATA_COUNT BYTECOUNT $files gr ELEMENT_COUNT ELEMCOUNT $files gr HASH_CODE HASHCODE $files ----------------------------------- cut ------------------------------------ `fixtypes.sh' is a Bourne-shell script; it uses 'gr': ----------------------------------- cut ------------------------------------ #!/bin/sh # Usage is like this: # gr FROM TO FILES ... # globally replace FROM with TO in FILES. FROM and TO are regular expressions. # backup files are stored in the `backup' directory. from="$1" to="$2" shift 2 echo ${1+"$@"} | xargs global-replace "s/$from/$to/g" ----------------------------------- cut ------------------------------------ `gr' in turn uses a Perl script to do its real work, `global-replace', which follows: ----------------------------------- cut ------------------------------------ : #-*- Perl -*- ### global-modify --- modify the contents of a file by a Perl expression ## Copyright (C) 1999 Martin Buchholz. ## Copyright (C) 2001 Ben Wing. ## Authors: Martin Buchholz <martin@xemacs.org>, Ben Wing <ben@xemacs.org> ## Maintainer: Ben Wing <ben@xemacs.org> ## Current Version: 1.0, May 5, 2001 # This program 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 your option) # any later version. # # This program 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 XEmacs; see the file COPYING. If not, write to the Free # Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA # 02111-1307, USA. eval 'exec perl -w -S $0 ${1+"$@"}' if 0; use strict; use FileHandle; use Carp; use Getopt::Long; use File::Basename; (my $myName = $0) =~ s@.*/@@; my $usage=" Usage: $myName [--help] [--backup-dir=DIR] [--line-mode] [--hunk-mode] PERLEXPR FILE ... Globally modify a file, either line by line or in one big hunk. Typical usage is like this: [with GNU print, GNU xargs: guaranteed to handle spaces, quotes, etc. in file names] find . -name '*.[ch]' -print0 | xargs -0 $0 's/\bCONST\b/const/g'\n [with non-GNU print, xargs] find . -name '*.[ch]' -print | xargs $0 's/\bCONST\b/const/g'\n The file is read in, either line by line (with --line-mode specified) or in one big hunk (with --hunk-mode specified; it's the default), and the Perl expression is then evalled with \$_ set to the line or hunk of text, including the terminating newline if there is one. It should destructively modify the value there, storing the changed result in \$_. Files in which any modifications are made are backed up to the directory specified using --backup-dir, or to `backup' by default. To disable this, use --backup-dir= with no argument. Hunk mode is the default because it is MUCH MUCH faster than line-by-line. Use line-by-line only when it matters, e.g. you want to do a replacement only once per line (the default without the `g' argument). Conversely, when using hunk mode, *ALWAYS* use `g'; otherwise, you will only make one replacement in the entire file! "; my %options = (); $Getopt::Long::ignorecase = 0; &GetOptions ( \%options, 'help', 'backup-dir=s', 'line-mode', 'hunk-mode', ); die $usage if $options{"help"} or @ARGV <= 1; my $code = shift; die $usage if grep (-d || ! -w, @ARGV); sub SafeOpen { open ((my $fh = new FileHandle), $_[0]); confess "Can't open $_[0]: $!" if ! defined $fh; return $fh; } sub SafeClose { close $_[0] or confess "Can't close $_[0]: $!"; } sub FileContents { my $fh = SafeOpen ("< $_[0]"); my $olddollarslash = $/; local $/ = undef; my $contents = <$fh>; $/ = $olddollarslash; return $contents; } sub WriteStringToFile { my $fh = SafeOpen ("> $_[0]"); binmode $fh; print $fh $_[1] or confess "$_[0]: $!\n"; SafeClose $fh; } foreach my $file (@ARGV) { my $changed_p = 0; my $new_contents = ""; if ($options{"line-mode"}) { my $fh = SafeOpen $file; while (<$fh>) { my $save_line = $_; eval $code; $changed_p = 1 if $save_line ne $_; $new_contents .= $_; } } else { my $orig_contents = $_ = FileContents $file; eval $code; if ($_ ne $orig_contents) { $changed_p = 1; $new_contents = $_; } } if ($changed_p) { my $backdir = $options{"backup-dir"}; $backdir = "backup" if !defined ($backdir); if ($backdir) { my ($name, $path, $suffix) = fileparse ($file, ""); my $backfulldir = $path . $backdir; my $backfile = "$backfulldir/$name"; mkdir $backfulldir, 0755 unless -d $backfulldir; print "modifying $file (original saved in $backfile)\n"; rename $file, $backfile; } WriteStringToFile ($file, $new_contents); } } ----------------------------------- cut ------------------------------------ In addition to those programs, I needed to fix up a few other things, particularly relating to the duplicate definitions of types, now that some types merged with others. Specifically: 1. in lisp.h, removed duplicate declarations of Bytecount. The changed code should now look like this: (In each code snippet below, the first and last lines are the same as the original, as are all lines outside of those lines. That allows you to locate the section to be replaced, and replace the stuff in that section, verifying that there isn't anything new added that would need to be kept.) --------------------------------- snip ------------------------------------- /* Counts of bytes or chars */ typedef EMACS_INT Bytecount; typedef EMACS_INT Charcount; /* Counts of elements */ typedef EMACS_INT Elemcount; /* Hash codes */ typedef unsigned long Hashcode; /* ------------------------ dynamic arrays ------------------- */ --------------------------------- snip ------------------------------------- 2. in lstream.h, removed duplicate declaration of Bytecount. Rewrote the comment about this type. The changed code should now look like this: --------------------------------- snip ------------------------------------- #endif /* The have been some arguments over the what the type should be that specifies a count of bytes in a data block to be written out or read in, using Lstream_read(), Lstream_write(), and related functions. Originally it was long, which worked fine; Martin "corrected" these to size_t and ssize_t on the grounds that this is theoretically cleaner and is in keeping with the C standards. Unfortunately, this practice is horribly error-prone due to design flaws in the way that mixed signed/unsigned arithmetic happens. In fact, by doing this change, Martin introduced a subtle but fatal error that caused the operation of sending large mail messages to the SMTP server under Windows to fail. By putting all values back to be signed, avoiding any signed/unsigned mixing, the bug immediately went away. The type then in use was Lstream_Data_Count, so that it be reverted cleanly if a vote came to that. Now it is Bytecount. Some earlier comments about why the type must be signed: This MUST BE SIGNED, since it also is used in functions that return the number of bytes actually read to or written from in an operation, and these functions can return -1 to signal error. Note that the standard Unix read() and write() functions define the count going in as a size_t, which is UNSIGNED, and the count going out as an ssize_t, which is SIGNED. This is a horrible design flaw. Not only is it highly likely to lead to logic errors when a -1 gets interpreted as a large positive number, but operations are bound to fail in all sorts of horrible ways when a number in the upper-half of the size_t range is passed in -- this number is unrepresentable as an ssize_t, so code that checks to see how many bytes are actually written (which is mandatory if you are dealing with certain types of devices) will get completely screwed up. --ben */ typedef enum lstream_buffering --------------------------------- snip ------------------------------------- 3. in dumper.c, there are four places, all inside of switch() statements, where XD_BYTECOUNT appears twice as a case tag. In each case, the two case blocks contain identical code, and you should *REMOVE THE SECOND* and leave the first.

This file describes what you must or might want to do to termcap entries
to make terminals work properly and efficiently with Emacs.  Information
on likely problems with specific types of terminals appears at the end
of the file.

*** What you want in a terminal ***

Vital
1. Easy to compute suitable padding for.
2. Never ever sends ^S/^Q unless you type them, at least in one mode.

Nice for speed
1. Supports insert/delete of multiple lines in one command.
2. Same for multiple characters, though doing them one by
one is usually fast enough except on emulators running on
machines with bitmap screens.

Nice for usability
1. Considerably more than 24 lines.
2. Meta key (shift-like key that controls the 0200 bit
in every character you type).

*** New termcap strings ***

Emacs supports certain termcap strings that are not described in the
4.2 manual but appear to be standard in system V.  The one exception
is `cS', which I invented.

`AL'    insert several lines.  Takes one parameter, the number of
        lines to be inserted.  You specify how to send this parameter
	using a %-construct, just like the cursor positions in the `cm'
	string.

`DL'    delete several lines.  One parameter.

`IC'    insert several characters.  One parameter.

`DC'    delete several characters.  One parameter.

`rp'    repeat a character.  Takes two parameters, the character
        to be repeated and the number of times to repeat it.
	Most likely you will use `%.' for sending the character
	to be repeated.  Emacs interprets a padding spec with a *
	as giving the amount of padding per repetition.

	WARNING: Many terminals have a command to repeat the
	*last character output* N times.  This means that the character
	will appear N+1 times in a row when the command argument is N.
	However, the `rp' string's parameter is the total number of
	times wanted, not one less.  Therefore, such repeat commands
	may be used in an `rp' string only if you use Emacs's special
	termcap operator `%a-c\001' to subtract 1 from the repeat count
	before substituting it into the string.  It is probably safe
	to use this even though the Unix termcap does not accept it
	because programs other than Emacs probably won't look for `rp'
	anyway.
	
`cs'    set scroll region.  Takes two parameters, the vertical
	positions of the first line to include in the scroll region
	and the last line to include in the scroll region.
	Both parameters are origin-zero.  The effect of this
	should be to cause a following insert-line or delete-line
	not to move lines below the bottom of the scroll region.

	This is not the same convention that Emacs version 16 used.
	That is because I was led astray by unclear documentation
	of the meaning of %i in termcap strings.  Since the termcap
	documentation for `cs' is also unclear, I had to deduce the
	correct parameter conventions from what would make the VT-100's
	`cs' string work properly.  From an incorrect assumption about
	%i, I reached an incorrect conclusion about `cs', but the result
	worked correctly on the VT100 and ANSII terminals.  In Emacs
	version 17, both `cs' and %i work correctly.

	The version 16 convention was to pass, for the second parameter,
	the line number of the first line beyond the end of the
	scroll region.

`cS'    set scroll region.  Differs from `cs' in taking parameters
	differently.  There are four parameters:
	1. Total number of lines on the screen.
	2. Number of lines above desired scroll region.
	3. Number of lines below (outside of) desired scroll region.
	4. Total number of lines on the screen, like #1.
	This is because an Ambassador needs the parameters like this.

`cr', `do', `le'
	Emacs will not attempt to use ^M, ^J or ^H for cursor motion
	unless these capabilities are present and say to use those
	characters.

`km'    Says the terminal has a Meta key.

Defining these strings is important for getting maximum performance
from your terminal.

Make sure that the `ti' string sets all modes needed for editing
in Emacs.  For example, if your terminal has a mode that controls
wrap at the end of the line, you must decide whether to specify
the `am' flag in the termcap entry; whichever you decide, the `ti'
string should contain commands to set the mode that way.
(Emacs also sends the `vs' string after the `ti' string.
You can put the mode-setting commands in either one of them.)

*** Specific Terminal Types ***

Watch out for termcap entries for Ann Arbor Ambassadors that
give too little padding for clear-screen.  7.2 msec per line is right.
These are the strings whose padding you probably should change:
    :al=1*\E[L:dl=1*\E[M:cd=7.2*\E[J:cl=7.2*\E[H\E[J:
I have sometimes seen `\E[2J' at the front of the `ti' string;
this is a clear-screen, very slow, and it can cause you to get
Control-s sent by the terminal at startup.  I recommend removing
the `\E[2J' from the `ti' string.
The `ti' or `vs' strings also usually need stuff added to them, such as
    \E[>33;52;54h\E[>30;37;38;39l
You might want to add the following to the `te' or `ve' strings:
    \E[>52l\E[>37h
The following additional capabilities will improve performance:
    :AL=1*\E[%dL:DL=1*\E[%dM:IC=4\E[%d@:DC=4\E[%dP:rp=1*%.\E[%a-c\001%db:
If you find that the Meta key does not work, make sure that
    :km:
is present in the termcap entry.

Watch out for termcap entries for VT100's that fail to specify
the `sf' string, or that omit the padding needed for the `sf' and `sr'
strings (2msec per line affected).  What you need is
    :sf=2*^J:sr=2*\EM:cs=\E[%i%d;%dr:

The Concept-100 and Concept-108 have many modes that `ti' strings
often fail to initialize.  If you have problems on one of these
terminals, that is probably the place to fix them.  These terminals
can support an `rp' string.

Watch out on HP terminals for problems with standout disappearing on
part of the mode line.  These problems are due to the absence of
:sg#0: which some HP terminals need.

The vi55 is said to require `ip=2'.

The Sun console should have these capabilities for good performance.
	   :AL=\E[%dL:DL=\E[%dM:IC=\E[%d@:DC=\E[%dP:

The vt220 needs to be set to vt220 mode, 7 bit, space parity
in order to work fully with TERM=vt220.

If you are using a LAT terminal concentrator, you need to issue these
commands to turn off flow control:

    set port flow control disable
    define port flow control disable

On System V, in the terminfo database, various terminals may have
the `xt' flag that should not have it.  `xt' should be present only
for the Teleray 1061 or equivalent terminal.

In particular, System V for the 386 often has `xt' for terminal type
AT386 or AT386-M, which is used for the console.  You should delete
this flag.  Here is how:

You can get a copy of the terminfo "source" for at386 using the
command: `infocmp at386 >at386.tic'.  Edit the file at386.tic and remove
the `xt' flag.  Then compile the new entry with: `tic at386.tic'.

It is also reported that these terminal types sometimes have the wrong
reverse-scroll string.  It should be \E[T, but sometimes is given as \E[S.

Here is what watserv1!maytag!focsys!larry recommends for these terminals:

# This copy of the terminfo description has been fixed.
# The suggestions came from a number of usenet postings.
#
# Intel AT/386 for color card with monochrome display
#
AT386-M|at386-m|386AT-M|386at-m|at/386 console,
	am, bw, eo, xon,
	cols#80, lines#25,
	acsc=``a1fxgqh0jYk?lZm@nEooppqDrrsstCu4vAwBx3yyzz{{||}}~~,
	bel=^G, blink=\E[5m, bold=\E[1m, cbt=\E[Z,
	clear=\E[2J\E[H,
	cr=\r, cub=\E[%p1%dD, cub1=\E[D, cud=\E[%p1%dB,
	cud1=\E[B, cuf=\E[%p1%dC, cuf1=\E[C,
	cup=\E[%i%p1%02d;%p2%02dH, cuu=\E[%p1%dA, cuu1=\E[A,
	dch=\E[%p1%dP, dch1=\E[P, dl=\E[%p1%dM, dl1=\E[1M,
	ech=\E[%p1%dX,ed=\E[J, el=\E[K, el1=\E[1K\E[X, flash=^G, home=\E[H,
	hpa=\E[%i%p1%dG, ich=\E[%p1%d@, ich1=\E[1@, il=\E[%p1%dL, il1=\E[1L,
	ind=\E[S, indn=\E[%p1%dS, invis=\E[9m,
	is2=\E[0;10;38m, kbs=\b, kcbt=^], kclr=\E[2J,
	kcub1=\E[D, kcud1=\E[B, kcuf1=\E[C, kcuu1=\E[A,
	kdch1=\E[P, kend=\E[Y, kf1=\EOP, kf10=\EOY, kf11=\EOZ,
	kf12=\EOA, kf2=\EOQ, kf3=\EOR, kf4=\EOS, kf5=\EOT,
	kf6=\EOU, kf7=\EOV, kf8=\EOW, kf9=\EOX, khome=\E[H,
	kich1=\E[@, knp=\E[U, kpp=\E[V, krmir=\E0, rev=\E[7m, ri=\E[T,
	rin=\E[%p1%dT, rmacs=\E[10m, rmso=\E[m, rmul=\E[m,
	sgr=\E[10m\E[0%?%p1%p3%|%t;7%;%?%p2%t;4%;%?%p4%t;5%;%?%p6%t;1%;%?%p9%t;12%;%?%p7%t;9%;m,
	sgr0=\E[0;10m, smacs=\E[12m, smso=\E[7m, smul=\E[4m,

#
# AT&T 386 color console 
#
AT386|at386|386AT|386at|at/386 console,
	colors#8, ncv#3, pairs#64,
	is2=\E[0;10;39m,
	op=\E[0m, 
	setb=\E[%?%p1%{0}%=%t40m
                %e%p1%{1}%=%t44m
                %e%p1%{2}%=%t42m
                %e%p1%{3}%=%t46m
                %e%p1%{4}%=%t41m
                %e%p1%{5}%=%t45m
                %e%p1%{6}%=%t43m
                %e%p1%{7}%=%t47m%;,
	setf=\E[%?%p1%{0}%=%t30m
                %e%p1%{1}%=%t34m
                %e%p1%{2}%=%t32m
                %e%p1%{3}%=%t36m
                %e%p1%{4}%=%t31m
                %e%p1%{5}%=%t35m
                %e%p1%{6}%=%t33m
                %e%p1%{6}%=%t33m
                %e%p1%{7}%=%t37m%;,
	use=at386-m,
#
# Color console version that supports underline but maps blue
# foreground color to cyan.
#
AT386-UL|at386-ul|386AT-UL|386at-ul|at/386 console,
	is2=\E[0;10;38m,
	use=at386,