view lisp/bytecomp-runtime.el @ 665:fdefd0186b75

[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.
author ben
date Thu, 20 Sep 2001 06:31:11 +0000
parents 7039e6323819
children 943eaba38521
line wrap: on
line source

;;; bytecomp-runtime.el --- byte-compiler support for inlining

;; Copyright (C) 1992, 1997 Free Software Foundation, Inc.

;; Author: Jamie Zawinski <jwz@jwz.org>
;; Author: Hallvard Furuseth <hbf@ulrik.uio.no>
;; Maintainer: XEmacs Development Team
;; Keywords: internal, dumped

;; This file is part of XEmacs.

;; XEmacs 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.

;; XEmacs 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.

;;; Synched up with: FSF 19.30.

;;; Commentary:

;; This file is dumped with XEmacs.

;; The code in this file should always be loaded, because it defines things 
;; like "defsubst" which should work interpreted as well.  The code in 
;; bytecomp.el and byte-optimize.el can be loaded as needed.

;; interface to selectively inlining functions.
;; This only happens when source-code optimization is turned on.

;;; Code:

;; Redefined in byte-optimize.el.
;; This is not documented--it's not clear that we should promote it.
(fset 'inline 'progn)
(put 'inline 'lisp-indent-hook 0)


;;; Interface to inline functions.

;; FSF comments the next two out, but I see no reason to do so. --ben
(defmacro proclaim-inline (&rest fns)
  "Cause the named functions to be open-coded when called from compiled code.
They will only be compiled open-coded when `byte-optimize' is true."
  (cons 'eval-and-compile
	(apply
	 'nconc
	 (mapcar
	  #'(lambda (x)
	      `((or (memq (get ',x 'byte-optimizer)
			  '(nil byte-compile-inline-expand))
		    (error
		     "%s already has a byte-optimizer, can't make it inline"
		     ',x))
		(put ',x 'byte-optimizer 'byte-compile-inline-expand)))
	  fns))))


(defmacro proclaim-notinline (&rest fns)
  "Cause the named functions to no longer be open-coded."
  (cons 'eval-and-compile
	(apply
	 'nconc
	 (mapcar
	  #'(lambda (x)
	      `((if (eq (get ',x 'byte-optimizer)
			'byte-compile-inline-expand)
		    (put ',x 'byte-optimizer nil))))
	  fns))))

;; This has a special byte-hunk-handler in bytecomp.el.
(defmacro defsubst (name arglist &rest body)
  "Define an inline function.  The syntax is just like that of `defun'."
  (or (memq (get name 'byte-optimizer)
	    '(nil byte-compile-inline-expand))
      (error "`%s' is a primitive" name))
  (list 'prog1
	(cons 'defun (cons name (cons arglist body)))
	(list 'proclaim-inline name)))
; Instead of the above line, FSF has this:
;	(list 'eval-and-compile
;	      (list 'put (list 'quote name)
;		    ''byte-optimizer ''byte-compile-inline-expand))))

(defun make-obsolete (fn new)
  "Make the byte-compiler warn that FUNCTION is obsolete.
The warning will say that NEW should be used instead.
If NEW is a string, that is the `use instead' message."
  (interactive "aMake function obsolete: \nxObsoletion replacement: ")
  (let ((handler (get fn 'byte-compile)))
    (if (eq 'byte-compile-obsolete handler)
	(setcar (get fn 'byte-obsolete-info) new)
      (put fn 'byte-obsolete-info (cons new handler))
      (put fn 'byte-compile 'byte-compile-obsolete)))
  fn)

(defun make-obsolete-variable (var new)
  "Make the byte-compiler warn that VARIABLE is obsolete,
and NEW should be used instead.  If NEW is a string, then that is the
`use instead' message."
  (interactive
   (list
    (let ((str (completing-read "Make variable obsolete: " obarray 'boundp t)))
      (if (equal str "") (error ""))
      (intern str))
    (car (read-from-string (read-string "Obsoletion replacement: ")))))
  (put var 'byte-obsolete-variable new)
  var)

;; By overwhelming demand, we separate out truly obsolete symbols from
;; those that are present for GNU Emacs compatibility.
(defun make-compatible (fn new)
  "Make the byte-compiler know that FUNCTION is provided for compatibility.
The warning will say that NEW should be used instead.
If NEW is a string, that is the `use instead' message."
  (interactive "aMake function compatible: \nxCompatible replacement: ")
  (let ((handler (get fn 'byte-compile)))
    (if (eq 'byte-compile-compatible handler)
	(setcar (get fn 'byte-compatible-info) new)
      (put fn 'byte-compatible-info (cons new handler))
      (put fn 'byte-compile 'byte-compile-compatible)))
  fn)

(defun make-compatible-variable (var new)
  "Make the byte-compiler know that VARIABLE is provided for compatibility.
and NEW should be used instead.  If NEW is a string, then that is the
`use instead' message."
  (interactive
   (list
    (let ((str (completing-read "Make variable compatible: "
				obarray 'boundp t)))
      (if (equal str "") (error ""))
      (intern str))
    (car (read-from-string (read-string "Compatible replacement: ")))))
  (put var 'byte-compatible-variable new)
  var)

(put 'dont-compile 'lisp-indent-hook 0)
(defmacro dont-compile (&rest body)
  "Like `progn', but the body always runs interpreted (not compiled).
If you think you need this, you're probably making a mistake somewhere."
  (list 'eval (list 'quote (if (cdr body) (cons 'progn body) (car body)))))


;;; interface to evaluating things at compile time and/or load time
;;; these macro must come after any uses of them in this file, as their
;;; definition in the file overrides the magic definitions on the
;;; byte-compile-macro-environment.

(put 'eval-when-compile 'lisp-indent-hook 0)
(defmacro eval-when-compile (&rest body)
  "Like `progn', but evaluates the body at compile time.
The result of the body appears to the compiler as a quoted constant."
  ;; Not necessary because we have it in b-c-initial-macro-environment
  ;; (list 'quote (eval (cons 'progn body)))
  (cons 'progn body))

(put 'eval-and-compile 'lisp-indent-hook 0)
(defmacro eval-and-compile (&rest body)
  "Like `progn', but evaluates the body at compile time and at load time."
  ;; Remember, it's magic.
  (cons 'progn body))

;;; From Emacs 20.
(put 'eval-when-feature 'lisp-indent-hook 1)
(defmacro eval-when-feature (feature &rest body)
  "Run the body forms when FEATURE is featurep, be it now or later.
Called (eval-when-feature (FEATURE [. FILENAME]) BODYFORMS...).
If (featurep 'FEATURE), evals now; otherwise adds an elt to
`after-load-alist' (which see), using FEATURE as filename if FILENAME is nil."
  (let ((file (or (cdr feature) (symbol-name (car feature)))))
    `(let ((bodythunk #'(lambda () ,@body)))
       (if (featurep ',(car feature))
	   (funcall bodythunk)
	 (setq after-load-alist (cons '(,file . (list 'lambda '() bodythunk))
				      after-load-alist))))))



;;; Functions to cleanly eliminate warnings about undefined functions
;;; or variables when the code knows what it's doing.  These macros DO
;;; NOT rely on any byte-compiler changes, and thus can be copied into
;;; a package and used within it.

;; NOTE: As a result of the above requirement, the macros rely on
;; "tricks" to get the warnings suppressed.  A cleaner way, of course,
;; would be to extend the byte compiler to provide a proper interface.

;; #### Should we require an unquoted symbol rather than a quoted one,
;; as we currently do?  The quoting gets no generality, as `eval' is
;; called at compile time.  But most functions and macros want quoted
;; arguments, and I find it extremely confusing to deal with cases
;; such as `throw' requiring a quoted argument but `block' an unquoted
;; one.

(put 'with-boundp 'lisp-indent-function 1)
(defmacro with-boundp (symbols &rest body)
  "Evaluate BODY, but do not issue bytecomp warnings about SYMBOLS undefined.
SYMBOLS can be a symbol or a list of symbols and must be quoted.  When
compiling this file, the warning `reference to free variable SYMBOL'
will not occur.  This is a clean way to avoid such warnings.  See also
`declare-boundp' and `if-boundp'."
  (setq symbols (eval symbols))
  (unless (consp symbols)
      (setq symbols (list symbols)))
  `(progn
     (declare (special ,@symbols))
     ,@body))

(put 'if-boundp 'lisp-indent-function 2)
(defmacro if-boundp (symbol then &rest else)
  "Equivalent to (if (boundp SYMBOL) THEN ELSE) but handles bytecomp warnings.
When compiling this file, the warning `reference to free variable SYMBOL'
will not occur.  This is a clean way to avoid such warnings.  See also
`with-boundp' and `declare-boundp'."
  `(with-boundp ,symbol
     (if (boundp ,symbol) ,then ,@else)))

(defmacro declare-boundp (symbol)
  "Evaluate SYMBOL without bytecomp warnings about the symbol.
Sample usage is

  (declare-boundp gpm-minor-mode)

which is equivalent to

  (with-fboundp 'gpm-minor-mode
    gpm-minor-mode)"
  `(with-boundp ',symbol ,symbol))

(defmacro globally-declare-boundp (symbol)
  "Declare that all free uses of SYMBOL in this file are valid.
SYMBOL can also be a list of symbols.  SYMBOL must be quoted.

When compiling this file, the warning `reference to free variable
SYMBOL' will not occur regardless of where calls to SYMBOL occur in
the file.

In general, you should *NOT* use this; use `declare-boundp',
`if-boundp', or `with-boundp' to wrap individual uses, as necessary.
That way, you're more likely to remember to put in the explicit checks
for the variable's existence that are usually necessary.  However,
`globally-declare-boundp' is better in some circumstances, such as
when writing an ELisp package that makes integral use of
optionally-compiled-in functionality (typically, an interface onto a
system library) and checks for the existence of the functionality at
some entry point to the package.  See `globally-declare-fboundp' for
more information."
  (setq symbol (eval symbol))
  (if (not (consp symbol))
      (setq symbol (list symbol)))
  `(progn
     ;; (defvar FOO) has no side effects.
     ,@(mapcar #'(lambda (sym) `(defvar ,sym)) symbol)))

(defun byte-compile-with-fboundp (form)
  (byte-compile-form (cons 'progn (cdr (cdr form))))
  ;; Unfortunately, byte-compile-unresolved-functions is used not only
  ;; for unresolved-function warnings, but also in connection with the
  ;; following warnings:

  ;; "defsubst %s was used before it was defined"
  ;; "%s being defined to take %s%s, but was previously called with %s"

  ;; By hacking byte-compile-unresolved-functions like this, we
  ;; effectively disable these warnings.  But code should not be using
  ;; `with-fboundp' with a function defined later on in the same
  ;; file, so this is not a big deal.

  (let ((symbols (eval (car (cdr form)))))
    (unless (consp symbols)
      (setq symbols (list symbols)))
    (setq symbols (mapcar #'(lambda (sym) (cons sym nil)) symbols))
    (setq byte-compile-unresolved-functions
	  (set-difference byte-compile-unresolved-functions symbols
			  :key #'car))
    ))

;; EEEEEEEEVIL hack.  We need to create our own byte-compilation
;; method so that the proper variables are bound while compilation
;; takes place (which is when the warnings get noticed and batched
;; up).  What we really want to do is make `with-fboundp' a macro
;; that simply `progn's its BODY; but GOD DAMN IT, macros can't have
;; their own byte-compilation methods!  So we make `with-fboundp' a
;; macro calling `with-fboundp-1', which is cleverly aliased to
;; progn.  This way we can put a byte-compilation method on
;; `with-fboundp-1', and when interpreting, progn will duly skip
;; the first, quoted argument, i.e. the symbol name. (We could make
;; `with-fboundp-1' a regular function, but then we'd have to thunk
;; BODY and eval it at runtime.  We could probably just do this using
;; (apply 'progn BODY), but the existing method is more obviously
;; guaranteed to work.)
;;
;; In defense, cl-macs.el does a very similar thing with
;; `cl-block-wrapper'.

(put 'with-fboundp-1 'byte-compile 'byte-compile-with-fboundp)
(defalias 'with-fboundp-1 'progn)

(put 'with-fboundp 'lisp-indent-function 1)
(defmacro with-fboundp (symbol &rest body)
  "Evaluate BODY, but do not issue bytecomp warnings about SYMBOL.
SYMBOL must be quoted.  When compiling this file, the warning `the
function SYMBOL is not known to be defined' will not occur.  This is a
clean way to avoid such warnings.  See also `declare-fboundp',
`if-fboundp', and `globally-declare-fboundp'."
  `(with-fboundp-1 ,symbol ,@body))

(put 'if-fboundp 'lisp-indent-function 2)
(defmacro if-fboundp (symbol then &rest else)
  "Equivalent to (if (fboundp SYMBOL) THEN ELSE) but handles bytecomp warnings.
When compiling this file, the warning `the function SYMBOL is not
known to be defined' will not occur.  This is a clean way to avoid
such warnings.  See also `declare-fboundp', `with-fboundp', and
`globally-declare-fboundp'."
  `(with-fboundp ,symbol
     (if (fboundp ,symbol) ,then ,@else)))

(defmacro declare-fboundp (form)
  "Execute FORM (a function call) without bytecomp warnings about the call.
Sample usage is

  (declare-fboundp (x-keysym-on-keyboard-sans-modifiers-p 'backspace))

which is equivalent to

  (with-fboundp 'x-keysym-on-keyboard-sans-modifiers-p
    (x-keysym-on-keyboard-sans-modifiers-p 'backspace))"
  `(with-fboundp ',(car form) ,form))

(defmacro globally-declare-fboundp (symbol)
  "Declare that all calls to function SYMBOL in this file are valid.
SYMBOL can also be a list of symbols.  SYMBOL must be quoted.

When compiling this file, the warning `the function SYMBOL is not
known to be defined' will not occur regardless of where calls to
SYMBOL occur in the file.

In general, you should *NOT* use this; use `declare-fboundp',
`if-fboundp', or `with-fboundp' to wrap individual uses, as necessary.
That way, you're more likely to remember to put in the explicit checks
for the function's existence that are usually necessary.  However,
`globally-declare-fboundp' is better in some circumstances, such as
when writing an ELisp package that makes integral use of
optionally-compiled-in functionality (typically, an interface onto a
system library) and checks for the existence of the functionality at
some entry point to the package.  The file `ldap.el' is a good
example: It provides a layer on top of the optional LDAP ELisp
primitives, makes calls to them throughout its code, and verifies the
presence of LDAP support at load time.  Putting calls to
`declare-fboundp' throughout the code would be a major annoyance."
  (when (cl-compiling-file)
    (setq symbol (eval symbol))
    (if (not (consp symbol))
	(setq symbol (list symbol)))
    ;; Another hack.  This works because the autoload environment is
    ;; currently used ONLY to suppress warnings, and the actual
    ;; autoload definition is not used. (NOTE: With this definition,
    ;; we will get spurious "multiple autoloads for %s" warnings if we
    ;; have an autoload later in the file for any functions in SYMBOL.
    ;; This is not something that code should ever do, though.)
    (setq byte-compile-autoload-environment
	  (append (mapcar #'(lambda (sym) (cons sym nil)) symbol)
		  byte-compile-autoload-environment)))
  nil)

(defun byte-compile-with-byte-compiler-warnings-suppressed (form)
  (let ((byte-compile-warnings byte-compile-warnings)
	(types (car (cdr form))))
    (unless (consp types)
      (setq types (list types)))
    (if (eq byte-compile-warnings t)
	(setq byte-compile-warnings byte-compile-default-warnings))
    (setq byte-compile-warnings (set-difference byte-compile-warnings types))
    (byte-compile-form (cons 'progn (cdr (cdr form))))))

;; Same hack here as with `with-fboundp'.
(put 'with-byte-compiler-warnings-suppressed-1 'byte-compile
     'byte-compile-with-byte-compiler-warnings-suppressed)
(defalias 'with-byte-compiler-warnings-suppressed-1 'progn)

(put 'with-byte-compiler-warnings-suppressed 'lisp-indent-function 1)
(defmacro with-byte-compiler-warnings-suppressed (type &rest body)
  "Evaluate BODY, but do not issue bytecomp warnings TYPE.
TYPE should be one of `redefine', `callargs', `subr-callargs',
`free-vars', `unresolved', `unused-vars', `obsolete', or `pedantic',
or a list of one or more of these symbols. (See `byte-compile-warnings'.)
TYPE must be quoted.

NOTE: You should *NOT* under normal circumstances be using this!
There are better ways of avoiding most of these warnings.  In particular:

-- use (declare (special ...)) if you are making use of
   dynamically-scoped variables.
-- use `with-fboundp', `declare-fboundp', `if-fboundp', or
   `globally-declare-fboundp' to avoid warnings about undefined
   functions when you know the function actually exists.
-- use `with-boundp', `declare-boundp', or `if-boundp' to avoid
   warnings about undefined variables when you know the variable
   actually exists.
-- use `with-obsolete-variable' or `with-obsolete-function' if you
   are purposely using such a variable or function."
  `(with-byte-compiler-warnings-suppressed-1 ,type ,@body))

;; #### These should be more clever.  You could (e.g.) try fletting
;; `byte-compile-obsolete' or temporarily removing the obsolete info
;; from the symbol and putting it back with an unwind-protect. (Or
;; better, modify the byte-compiler to provide a proper solution, and
;; fix these macros to use it if available, or fall back on the way
;; below.  Remember, these definitions need to work with an unchanged
;; byte compiler so that they can be copied and used in packages.)

(put 'with-obsolete-variable 'lisp-indent-function 1)
(defmacro with-obsolete-variable (symbol &rest body)
  "Evaluate BODY but do not warn about usage of obsolete variable SYMBOL.
SYMBOL must be quoted.  See also `with-obsolete-function'."
  `(with-byte-compiler-warnings-suppressed 'obsolete ,@body))

(put 'with-obsolete-function 'lisp-indent-function 1)
(defmacro with-obsolete-function (symbol &rest body)
  "Evaluate BODY but do not warn about usage of obsolete function SYMBOL.
SYMBOL must be quoted.  See also `with-obsolete-variable'."
  `(with-byte-compiler-warnings-suppressed 'obsolete ,@body))


;;; Interface to file-local byte-compiler parameters.
;;; Redefined in bytecomp.el.

;;; The great RMS speaketh:
;;;
;;; I nuked this because it's not a good idea for users to think of
;;; using it.  These options are a matter of installation preference,
;;; and have nothing to do with particular source files; it's a
;;; mistake to suggest to users that they should associate these with
;;; particular source files.  There is hardly any reason to change
;;; these parameters, anyway.  --rms.
;;;
;;; But I'll leave this stuff alone. --ben

(put 'byte-compiler-options 'lisp-indent-hook 0)
(defmacro byte-compiler-options (&rest args)
  "Set some compilation-parameters for this file.  
This will affect only the file in which it appears; this does nothing when
evaluated, or when loaded from a .el file.

Each argument to this macro must be a list of a key and a value.

  Keys:		  Values:		Corresponding variable:

  verbose	  t, nil		byte-compile-verbose
  optimize	  t, nil, source, byte	byte-optimize
  warnings	  list of warnings	byte-compile-warnings
  file-format	  emacs19, emacs20	byte-compile-emacs19-compatibility

The value specified with the `warnings' option must be a list, containing
some subset of the following flags:

  free-vars	references to variables not in the current lexical scope.
  unused-vars	references to non-global variables bound but not referenced.
  unresolved	calls to unknown functions.
  callargs	lambda calls with args that don't match the definition.
  subr-callargs	calls to subrs with args that don't match the definition.
  redefine	function cell redefined from a macro to a lambda or vice
		versa, or redefined to take a different number of arguments.
  obsolete	use of an obsolete function or variable.
  pedantic	warn of use of compatible symbols.

If the first element if the list is `+' or `-' then the specified elements 
are added to or removed from the current set of warnings, instead of the
entire set of warnings being overwritten.

For example, something like this might appear at the top of a source file:

    (byte-compiler-options
      (optimize t)
      (warnings (- callargs))		; Don't warn about arglist mismatch
      (warnings (+ unused-vars))	; Do warn about unused bindings
      (file-format emacs19))"
  nil)

;;; bytecomp-runtime.el ends here