[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.

/* Code shared between all event loops that use select() and have a
   different input descriptor for each device.
   Copyright (C) 1991, 1992, 1993, 1994, 1995 Free Software Foundation, Inc.
   Copyright (C) 1995 Board of Trustees, University of Illinois.
   Copyright (C) 1995 Sun Microsystems, Inc.
   Copyright (C) 1995, 1996 Ben Wing.

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: Not in FSF. */

/* This file has been Mule-ized. */

#include <config.h>
#include "lisp.h"

#include "console-stream.h"
#include "console-tty.h"
#include "device.h"
#include "events.h"
#include "lstream.h"
#include "process.h"

#include "sysdep.h"
#include "sysfile.h"
#include "sysproc.h"		/* select stuff */
#include "systime.h"

/* Mask of bits indicating the descriptors that we wait for input on.
   These work as follows:

   input_wait_mask == mask of all file descriptors we select() on,
                      including TTY/stream console descriptors,
		      process descriptors, and the signal event pipe.
		      Only used in event-tty.c; event-Xt.c uses
		      XtAppAddInput(), and the call to select() is down in
		      the guts of Xt.

   non_fake_input_wait_mask == same as input_wait_mask but minus the
                               signal event pipe.  Also only used in
			       event-tty.c.

   process_only_mask == only the process descriptors.

   tty_only_mask == only the TTY/stream console descriptors.
   */
SELECT_TYPE input_wait_mask, non_fake_input_wait_mask;
SELECT_TYPE process_only_mask, tty_only_mask;

/* This is used to terminate the select(), when an event came in
   through a signal (e.g. window-change or C-g on controlling TTY). */
int signal_event_pipe[2];

int signal_event_pipe_initialized;

int fake_event_occurred;

int
read_event_from_tty_or_stream_desc (Lisp_Event *event,
				    struct console *con, int fd)
{
  unsigned char ch;
  int nread;
  Lisp_Object console;

  XSETCONSOLE (console, con);

  nread = read (fd, &ch, 1);
  if (nread <= 0)
    {
      /* deleting the console might not be safe right now ... */
      enqueue_magic_eval_event (io_error_delete_console, console);
      /* but we definitely need to unselect it to avoid infinite
	 loops reading EOF's */
      Fconsole_disable_input (console);
    }
  else
    {
      character_to_event (ch, event, con, 1, 1);
      event->channel = console;
      return 1;
    }
  return 0;
}

void
signal_fake_event (void)
{
  char byte = 0;
  /* We do the write always.  Formerly I tried to "optimize" this
     by setting a flag indicating whether we're blocking and only
     doing the write in that case, but there is a race condition
     if the signal occurs after we've checked for the signal
     occurrence (which could occur in many places throughout
     an iteration of the command loop, e.g. in status_notify()),
     but before we set the blocking flag.

     This should be OK as long as write() is reentrant, which
     I'm fairly sure it is since it's a system call. */

  if (signal_event_pipe_initialized)
    /* In case a signal comes through while we're dumping */
    {
      int old_errno = errno;
      write (signal_event_pipe[1], &byte, 1);
      errno = old_errno;
    }
}

void
drain_signal_event_pipe (void)
{
  char chars[128];
  /* The input end of the pipe has been set to non-blocking. */
  while (read (signal_event_pipe[0], chars, sizeof (chars)) > 0)
    ;
}

int
event_stream_unixoid_select_console (struct console *con)
{
  int infd;

  if (CONSOLE_STREAM_P (con))
    infd = fileno (CONSOLE_STREAM_DATA (con)->in);
  else
    {
      assert (CONSOLE_TTY_P (con));
      infd = CONSOLE_TTY_DATA (con)->infd;
    }

  assert (infd >= 0);

  FD_SET (infd, &input_wait_mask);
  FD_SET (infd, &non_fake_input_wait_mask);
  FD_SET (infd, &tty_only_mask);
  return infd;
}

int
event_stream_unixoid_unselect_console (struct console *con)
{
  int infd;

  if (CONSOLE_STREAM_P (con))
    infd = fileno (CONSOLE_STREAM_DATA (con)->in);
  else
    {
      assert (CONSOLE_TTY_P (con));
      infd = CONSOLE_TTY_DATA (con)->infd;
    }

  assert (infd >= 0);

  FD_CLR (infd, &input_wait_mask);
  FD_CLR (infd, &non_fake_input_wait_mask);
  FD_CLR (infd, &tty_only_mask);
  return infd;
}

static int
get_process_infd (Lisp_Process *p)
{
  Lisp_Object instr, outstr;
  get_process_streams (p, &instr, &outstr);
  assert (!NILP (instr));
  return filedesc_stream_fd (XLSTREAM (instr));
}

int
event_stream_unixoid_select_process (Lisp_Process *proc)
{
  int infd = get_process_infd (proc);

  FD_SET (infd, &input_wait_mask);
  FD_SET (infd, &non_fake_input_wait_mask);
  FD_SET (infd, &process_only_mask);
  return infd;
}

int
event_stream_unixoid_unselect_process (Lisp_Process *proc)
{
  int infd = get_process_infd (proc);

  FD_CLR (infd, &input_wait_mask);
  FD_CLR (infd, &non_fake_input_wait_mask);
  FD_CLR (infd, &process_only_mask);
  return infd;
}

int
poll_fds_for_input (SELECT_TYPE mask)
{
  EMACS_TIME sometime;
  EMACS_SELECT_TIME select_time;
  SELECT_TYPE temp_mask;
  int retval;

  while (1)
    {
      EMACS_SET_SECS_USECS (sometime, 0, 0);
      EMACS_TIME_TO_SELECT_TIME (sometime, select_time);
      temp_mask = mask;
      /* To effect a poll, tell select() to block for zero seconds. */
      retval = select (MAXDESC, &temp_mask, 0, 0, &select_time);
      if (retval >= 0)
	return retval;
      if (errno != EINTR)
	{
	  /* Something went seriously wrong; don't abort since maybe
	     the TTY just died at the wrong time. */
	  stderr_out ("xemacs: select failed: errno = %d\n", errno);
	  return 0;
	}
      /* else, we got interrupted by a signal, so try again. */
    }

  RETURN_NOT_REACHED(0) /* not reached */
}

/****************************************************************************/
/*     Unixoid (file descriptors based) process I/O streams routines        */
/****************************************************************************/

USID
event_stream_unixoid_create_stream_pair (void* inhandle, void* outhandle,
					 Lisp_Object* instream,
					 Lisp_Object* outstream,
					 int flags)
{
  int infd, outfd;
  /* Decode inhandle and outhandle. Their meaning depends on
     the process implementation being used. */
#if defined (HAVE_WIN32_PROCESSES)
  /* We're passed in Windows handles. Open new fds for them */
  if ((HANDLE)inhandle != INVALID_HANDLE_VALUE)
    {
      infd = open_osfhandle ((HANDLE)inhandle, 0);
      if (infd < 0)
	return USID_ERROR;
    }
  else
    infd = -1;

  if ((HANDLE)outhandle != INVALID_HANDLE_VALUE)
    {
      outfd = open_osfhandle ((HANDLE)outhandle, 0);
      if (outfd < 0)
	{
	  if (infd >= 0)
	    close (infd);
	  return USID_ERROR;
	}
    }
  else
    outfd = -1;

  flags = 0;
#elif defined (HAVE_UNIX_PROCESSES)
  /* We are passed plain old file descs */
  infd  = (int)inhandle;
  outfd = (int)outhandle;
#else
# error Which processes do you have?
#endif

  *instream = (infd >= 0
	       ? make_filedesc_input_stream (infd, 0, -1, 0)
	       : Qnil);

  *outstream = (outfd >= 0
		? make_filedesc_output_stream (outfd, 0, -1, LSTR_BLOCKED_OK)
		: Qnil);

#if defined(HAVE_UNIX_PROCESSES)
  /* FLAGS is process->pty_flag for UNIX_PROCESSES */
  if ((flags & STREAM_PTY_FLUSHING) && outfd >= 0)
    {
      Intbyte eof_char = get_eof_char (outfd);
      int pty_max_bytes = get_pty_max_bytes (outfd);
      filedesc_stream_set_pty_flushing (XLSTREAM(*outstream), pty_max_bytes, eof_char);
    }
#endif

  return FD_TO_USID (infd);
}

USID
event_stream_unixoid_delete_stream_pair (Lisp_Object instream,
					 Lisp_Object outstream)
{
  int in = (NILP(instream) ? -1
	    : filedesc_stream_fd (XLSTREAM (instream)));
  int out = (NILP(outstream) ? -1
	     : filedesc_stream_fd (XLSTREAM (outstream)));

  if (in >= 0)
    close (in);
  if (out != in && out >= 0)
    close (out);

  return FD_TO_USID (in);
}


void
init_event_unixoid (void)
{
  /* Do this first; the init_event_*_late() functions
     pay attention to it. */
  if (pipe (signal_event_pipe) < 0)
    {
      perror ("XEmacs: can't open pipe");
      exit (-1);
    }
  signal_event_pipe_initialized = 1;

  /* Set it non-blocking so we can drain its output. */
  set_descriptor_non_blocking (signal_event_pipe[0]);

  /* Also set the write descriptor non-blocking so we don't
     hang in case a long time passes between times when
     we drain the pipe. */
  set_descriptor_non_blocking (signal_event_pipe[1]);

  /* WARNING: In order for the signal-event pipe to work correctly
     and not cause lockups, the following need to be followed:

     1) event_pending_p() must ignore input on the signal-event pipe.
     2) As soon as next_event() notices input on the signal-event
     pipe, it must drain it. */
  FD_ZERO (&input_wait_mask);
  FD_ZERO (&non_fake_input_wait_mask);
  FD_ZERO (&process_only_mask);
  FD_ZERO (&tty_only_mask);

  FD_SET (signal_event_pipe[0], &input_wait_mask);
}