[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.

/* Fundamental definitions for XEmacs Lisp interpreter -- non-union objects.
   Copyright (C) 1985, 1986, 1987, 1992, 1993 Free Software Foundation, Inc.

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.  Split out from lisp.h. */
/* This file has diverged greatly from FSF Emacs.  Syncing is no
   longer desirable or possible */

/*
 Format of a non-union-type Lisp Object

             3         2         1         0
       bit  10987654321098765432109876543210
            --------------------------------
            VVVVVVVVVVVVVVVVVVVVVVVVVVVVVVTT

   Integers are treated specially, and look like this:

             3         2         1         0
       bit  10987654321098765432109876543210
            --------------------------------
            VVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVT

 For integral Lisp types, i.e. integers and characters, the value
 bits are the Lisp object.  Some people call such Lisp_Objects "immediate".

 The object is obtained by masking off the type bits.
     Bit 1 is used as a value bit by splitting the Lisp integer type
 into two subtypes, Lisp_Type_Int_Even and Lisp_Type_Int_Odd.
 By this trickery we get 31 bits for integers instead of 30.

 For non-integral types, the value bits of a Lisp_Object contain
 a pointer to a structure containing the object.  The pointer is
 obtained by masking off the type and mark bits.

     All pointer-based types are coalesced under a single type called
 Lisp_Type_Record.  The type bits for this type are required by the
 implementation to be 00, just like the least significant bits of
 word-aligned struct pointers on 32-bit hardware.  This requires that
 all structs implementing Lisp_Objects have an alignment of at least 4
 bytes.  Because of this, Lisp_Object pointers don't have to be masked
 and are full-sized.

 There are no mark bits in the Lisp_Object itself (there used to be).

 Integers and characters don't need to be marked.  All other types are
 lrecord-based, which means they get marked by setting the mark bit in
 the struct lrecord_header.

 Here is a brief description of the following macros:

 XTYPE     The type bits of a Lisp_Object
 XPNTRVAL  The value bits of a Lisp_Object storing a pointer
 XCHARVAL  The value bits of a Lisp_Object storing a Emchar
 XREALINT  The value bits of a Lisp_Object storing an integer, signed
 XUINT     The value bits of a Lisp_Object storing an integer, unsigned
 INTP      Non-zero if this Lisp_Object is an integer
 Qzero     Lisp Integer 0
 EQ        Non-zero if two Lisp_Objects are identical, not merely equal. */


typedef EMACS_INT Lisp_Object;

#define Lisp_Type_Int_Bit (Lisp_Type_Int_Even & Lisp_Type_Int_Odd)
#define wrap_pointer_1(ptr) ((Lisp_Object) (ptr))
#define make_int(x) ((Lisp_Object) (((x) << INT_GCBITS) | Lisp_Type_Int_Bit))
#define make_char(x) ((Lisp_Object) (((x) << GCBITS) | Lisp_Type_Char))
#define VALMASK (((1UL << VALBITS) - 1UL) << GCTYPEBITS)
#define XTYPE(x) ((enum Lisp_Type) (((EMACS_UINT)(x)) & ~VALMASK))
#define XPNTRVAL(x) (x) /* This depends on Lisp_Type_Record == 0 */
#define XCHARVAL(x) ((x) >> GCBITS)
#define XREALINT(x) ((x) >> INT_GCBITS)
#define XUINT(x) ((EMACS_UINT)(x) >> INT_GCBITS)
#define INTP(x) ((EMACS_UINT)(x) & Lisp_Type_Int_Bit)
#define INT_PLUS(x,y)  ((x)+(y)-Lisp_Type_Int_Bit)
#define INT_MINUS(x,y) ((x)-(y)+Lisp_Type_Int_Bit)
#define INT_PLUS1(x)   INT_PLUS  (x, make_int (1))
#define INT_MINUS1(x)  INT_MINUS (x, make_int (1))

#define Qzero make_int (0)
#define Qnull_pointer ((Lisp_Object) 0)
#define EQ(x,y) ((x) == (y))
#define XSETINT(var, value) ((void) ((var) = make_int (value)))
#define XSETCHAR(var, value) ((void) ((var) = make_char (value)))
#define XSETOBJ(var, value) ((void) ((var) = wrap_pointer_1 (value)))

/* Convert between a (void *) and a Lisp_Object, as when the
   Lisp_Object is passed to a toolkit callback function */
#define VOID_TO_LISP(larg,varg) ((void) ((larg) = ((Lisp_Object) (varg))))
#define CVOID_TO_LISP VOID_TO_LISP
#define LISP_TO_VOID(larg) ((void *) (larg))
#define LISP_TO_CVOID(larg) ((const void *) (larg))

/* Convert a Lisp_Object into something that can't be used as an
   lvalue.  Useful for type-checking. */
#define NON_LVALUE(larg) ((larg) + 0)