freebsd-skq/contrib/cvs/lib/ChangeLog
1999-03-18 09:21:42 +00:00

688 lines
24 KiB
Plaintext

Tue Mar 24 16:08:00 1998 Ian Lance Taylor <ian@cygnus.com>
* Makefile.in (CFLAGS): Set to @CFLAGS@, not -g.
1998-02-20 Jim Kingdon <kingdon@harvey.cyclic.com>
* regex.c: Partial merge with version from emacs 20.2. Brings
over some trivial changes (whitespace and so on) (most such
changes I didn't bother with, for this time). Don't cast to int
before comparing old_regend[r] to regstart[r] (this is the point
of bothering; the old code was broken for 64 bit machines.
Reported by Paul Vixie).
Tue Feb 17 18:33:26 1998 Ian Lance Taylor <ian@cygnus.com>
* memmove.c: New file, resurrecting the old one.
* Makefile.in (SOURCES): Add memmove.c.
1998-02-03 Tim Pierce <twp@skepsis.com>
* system.h (CVS_LSTAT): New macro.
Sat Feb 7 17:33:39 1998 Ian Lance Taylor <ian@cygnus.com>
* getline.h (getstr): Declare.
13 Jan 1998 Jim Kingdon
* fncase.c: Include config.h before system.h.
* system.h: Just include string.h unconditionally. We already
include it unconditionally elsewhere.
Tue Jan 13 16:51:59 1998 Ian Lance Taylor <ian@cygnus.com>
* fncase.c: New file, taken from windows-NT/filesubr.c.
* system.h: If __CYGWIN32__ or WIN32 are defined, define
FOLD_FN_CHAR, FILENAMES_CASE_INSENSITIVE, and ISDIRSEP, and
declare fncmp and fnfold. Taken from windows-NT/config.h.
* Makefile.in (SOURCES): Add fncase.c.
Sat Jan 10 10:51:26 1998 Jim Kingdon <kingdon@harvey.cyclic.com>
* getline.c (getstr): Make sure to set errno when appropriate. I
didn't test the error case for the new code but inspection shows
the old code was rather broken.
Sat Nov 29 22:03:39 1997 Jim Kingdon <kingdon@harvey.cyclic.com>
getwd and getcwd were a big big mess. Although Jim's fix might
indeed be fixing a typo, the code is so tangled that I would guess
it probably breaks some system. So clean this up:
* xgetwd.c: Always assume we have getcwd (we had been anyway,
before Jim's change).
* getwd.c: Removed.
* Makefile.in: Remove getwd.c
* system.h: Remove declarations of getwd and getcwd. Move getcwd
declaration to the !HAVE_UNISTD_H section.
1997-11-29 Jim Meyering <meyering@na-net.ornl.gov>
* xgetwd.c: Fix typo s/ifndef/ifdef/ in test of HAVE_GETWD.
Wed Nov 26 10:12:33 1997 Jim Kingdon <kingdon@harvey.cyclic.com>
* system.h: Always use "rb" and "wb". Check for O_BINARY with an
#ifdef, not the error-prone LINES_CRLF_TERMINATED.
Thu Sep 25 10:57:39 1997 Jim Kingdon <kingdon@harvey.cyclic.com>
* getdate.y (get_date): If gmtime returns NULL, try to cope.
* getdate.c: Regenerated using byacc.
* getdate.y: Remove comment about sending email concerning this file
to Rich Salz.
* getdate.c: Regenerated using byacc.
Wed Sep 24 10:35:38 1997 Jim Kingdon <kingdon@harvey.cyclic.com>
* Makefile.in (OBJECTS): Add regex.o.
Wed Sep 17 16:37:17 1997 Jim Kingdon <kingdon@harvey.cyclic.com>
* getdate.y (ToSeconds): For am or pm, a hour of "12" really means 0.
* getdate.c: Regenerated using byacc (not bison per comment).
Tue Sep 9 20:51:45 1997 Jim Kingdon <kingdon@harvey.cyclic.com>
* build_lib.com: Add vasprintf.c and vasprintf.obj.
* build_lib.com: Remove strippath.obj from library/create command.
Sun Sep 7 17:35:27 1997 Jim Kingdon <kingdon@harvey.cyclic.com>
* system.h: Replace comment referring to ChangeLog with a
comment based on the ChangeLog entries.
* strdup.c: Removed, per change to ../configure.in
* Makefile.in (SOURCES): Remove strdup.c.
Mon Jun 16 18:59:50 1997 Jim Kingdon <kingdon@harvey.cyclic.com>
* system.h: Add CVS_FNMATCH.
Sun Jun 8 23:41:11 1997 Jim Kingdon <kingdon@harvey.cyclic.com>
* system.h (mkfifo): Remove; not used anywhere.
Thu Mar 6 17:14:49 1997 Jim Kingdon <kingdon@harvey.cyclic.com>
* regex.c: Partial merge with version from emacs 19.34. I brought
over most trivial changes (whitespace and so on). Most of the
changes to portability cruft I did not bring over, on the theory
of sticking to the devil that we know. I did bring over the
change to undef MAX and MIN (this is a better solution to a
problem we had been handling a different way). There were a
variety of changes I probably could/should have brought over, but
elected not to try to understand them and whether they would cause
trouble (printchar -> putchar, changes to output format in
print_partial_compiled_pattern, internationalization,
FREE_STACK_RETURN and friends which would appear to be fixing
memory leaks in error cases, RE_TRANSLATE_TYPE, and others). I
did merge the changes (union fail_stack_elt, PUSH_FAILURE_POINTER,
etc.) to use a union for the failure stack rather than playing
games with pointers and integers (that was my reason for
bothering; the code had been broken on the Alpha).
Mon Feb 10 18:52:18 1997 Ullrich von Bassewitz <uz@musoftware.com>
* md5.c: Make the parameter to getu32 const since the function will
only read the values and this will avoid compiler warnings in other
places.
Mon Feb 10 18:29:04 1997 Ullrich von Bassewitz <uz@musoftware.com>
* vasprintf.c: Added a #define for systems where a va_list is
defined as an array, not as a pointer.
Mon Feb 10 09:31:38 1997 Ken Raeburn <raeburn@cygnus.com>
* md5.c (MD5STEP): Truncate to 32 bits before shifting right.
Thu Jan 30 11:35:26 1997 Jim Kingdon <kingdon@harvey.cyclic.com>
* regex.h: Don't prototype re_comp and re_exec.
Tue Jan 28 17:45:46 1997 Jim Kingdon <kingdon@harvey.cyclic.com>
* md5.c, md5.h: Changes so these work without having an integer
type which is exactly 32 bits. Modeled after changes by Tatu Ylonen
<ylo@cs.hut.fi> as part of SSH but rewritten.
Wed Jan 8 14:50:47 1997 Jim Kingdon <kingdon@harvey.cyclic.com>
* Makefile.in, getopt.h, sighandle.c, system.h: Remove CVSid; we
decided to get rid of these some time ago.
Thu Jan 2 13:30:56 1997 Jim Kingdon <kingdon@harvey.cyclic.com>
* Makefile.in, argmatch.c, fnmatch.c, fnmatch.h, getline.c,
getopt.c, getopt.h, getopt1.c, getwd.c, hostname.c, mkdir.c,
regex.c, regex.h, rename.c, sighandle.c, strdup.c, strerror.c,
stripslash.c, system.h, vasprintf.c, wait.h, xgetwd.c, yesno.c:
Remove "675" paragraph; see ../ChangeLog for rationale.
Sun Nov 24 13:34:25 1996 Jim Kingdon <kingdon@harvey.cyclic.com>
* getdate.y (Convert): Change last acceptable year from 1999 to
2038.
* getdate.c: Regenerated using byacc 1.9.
Tue Nov 19 17:11:17 1996 Jim Kingdon <kingdon@harvey.cyclic.com>
* Makefile.in (OBJECTS): Remove strippath.o; we don't use
strip_path anymore.
(SOURCES): Remove strippath.c.
* strippath.c: Removed.
* build_lib.com: Remove strippath.c.
Wed Oct 2 10:43:35 1996 Norbert Kiesel <nk@col.sw-ley.de>
* getdate.y: removed CVSid variable
* getdate.c: regenerated (using byacc 1.9)
Wed Sep 25 10:25:00 1996 Larry Jones <larry.jones@sdrc.com>
* vasprintf.c: Fix type clashes in calls to strtoul.
Wed Sep 11 15:55:31 1996 Jim Kingdon <kingdon@harvey.cyclic.com>
* build_lib.com: Add valloc.c.
Tue Sep 10 23:04:34 1996 Jim Kingdon <kingdon@harvey.cyclic.com>
* Makefile.in (DISTFILES): Add build_lib.com.
Fri Aug 16 16:01:57 1996 Norbert Kiesel <nk@col.sw-ley.de>
* Makefile.in (installdirs): new (empty) target
Mon Aug 12 11:03:43 1996 Jim Kingdon <kingdon@harvey.cyclic.com>
* system.h: Don't use #elif. It is said to cause problems with
one of the HP compilers on HPUX 9.01.
Sun Jul 7 23:25:46 1996 Jim Kingdon <kingdon@harvey.cyclic.com>
* memmove.c: Removed. The memove function was used by a very old
version of the CVS server for nefarious purposes and it has been
long gone.
* Makefile.in (SOURCES): Remove memmove.c.
Thu Jun 6 15:12:59 1996 Jim Kingdon <kingdon@harvey.cyclic.com>
* vasprintf.c: If STDC_HEADERS, include stdlib.h rather than
declaring its functions ourself.
Wed Jun 05 10:14:29 1996 Mike Ladwig <mike@twinpeaks.prc.com>
and Jim Kingdon <kingdon@cyclic.com>
* system.h: If ERRNO_H_MISSING is defined, don't include errno.h.
Wed Jun 05 10:14:29 1996 Mike Ladwig <mike@twinpeaks.prc.com>
* regex.c: Don't define MAX and MIN if already defined.
Sun May 12 09:40:08 1996 Jim Kingdon <kingdon@harvey.cyclic.com>
* getdate.y: Replace alloca.h include with a comment explaining
why we avoid alloca and the consequences of that.
* getdate.c: Regenerated.
Wed May 8 09:31:03 1996 Jim Kingdon <kingdon@harvey.cyclic.com>
* getdate.c: Regenerate with the version of byacc in Red Hat 3.0.3
(which I believe is byacc 1.9). byacc, unlike bison, does not
require alloca in the generated parser.
Thu Apr 25 18:26:34 1996 Jim Kingdon <kingdon@harvey.cyclic.com>
* getdate.y (get_date): Set Start from nowtime, not now->time,
which may not be set.
* getdate.c: Regenerated.
Wed Apr 10 17:55:02 1996 Jim Kingdon <kingdon@harvey.cyclic.com>
* getdate.y (get_date): Use a time_t variable rather than a field
in a struct timeb. Works around Solaris compiler bug. Sure, it
is a compiler bug, but the workaround is completely painless.
* getdate.c: Regenerated.
Fri Mar 22 11:17:05 1996 Jim Kingdon <kingdon@harvey.cyclic.com>
* system.h: If EXIT_FAILURE is not defined by stdlib.h, define it
ourself.
Thu Mar 14 16:27:53 1996 Jim Kingdon <kingdon@harvey.cyclic.com>
* system.h: Remove alloca cruft.
Wed Feb 28 03:16:48 1996 Benjamin J. Lee <benjamin@cyclic.com>
* build_lib.com: Changed definition of symbol CC to search
for include files in [-.VMS] so VMS config.h can be picked
up without copying.
Tue Feb 27 21:26:34 1996 Benjamin J. Lee <benjamin@cyclic.com>
* build_lib.com: Added. DCL File to build contents of [.lib]
Tue Feb 27 21:18:38 1996 Benjamin J. Lee <benjamin@cyclic.com>
* system.h: added an existence_error macro check for EVMSERR
necessary for happiness under VMS
Thu Feb 22 22:30:04 1996 Jim Kingdon <kingdon@harvey.cyclic.com>
* Makefile.in (OBJECTS): Remove @ALLOCA@
(SOURCES): Remove alloca.c
* alloca.c: Removed.
* regex.c (REGEX_MALLOC): Define.
Thu Feb 15 14:00:00 Jim Kingdon <kingdon@cyclic.com>
* vasprintf.c: Declare abs().
Wed Feb 14 14:48:31 1996 Jim Kingdon <kingdon@harvey.cyclic.com>
* vasprintf.c (int_vasprintf): Don't cast arguments to memcpy.
* vasprintf.c, strtoul.c: Don't include ansidecl.h. Do include
config.h if HAVE_CONFIG_H (for const).
* strtoul.c: Change CONST to const.
Tue Feb 13 20:04:39 1996 Jim Kingdon <kingdon@harvey.cyclic.com>
* strtoul.c: Added (needed by vasprintf.c, and missing on SunOS4).
* Makefile.in (SOURCES): Add strtoul.c.
Mon Feb 12 10:04:46 1996 Jim Kingdon <kingdon@harvey.cyclic.com>
* vasprintf.c: Added (same contents as before).
* Makefile.in (SOURCES): Add vasprintf.c.
Thu Feb 1 14:33:17 1996 Karl Fogel <kfogel@floss.red-bean.com>
* Makefile.in (xlint): new rule; does nothing, as I'm not sure
running lint is actually advisable in here, but the top-level
Makefile thinks it can `make xlint' here.
Thu Feb 1 15:07:42 1996 Jim Kingdon <kingdon@harvey.cyclic.com>
* getopt.c: Remove rcsid.
Tue Jan 30 18:20:27 1996 Jim Kingdon <kingdon@harvey.cyclic.com>
* getline.c: Don't define NDEBUG.
(getstr): Rewrite assertions in a way which should stay clear of
signed/unsigned problems and compiler warnings thereof.
Thu Jan 25 00:14:06 1996 Jim Kingdon <kingdon@beezley.cyclic.com>
* yesno.c (yesno): fflush stdout as well as stderr.
Wed Jan 3 18:16:50 1996 Jim Kingdon <kingdon@harvey.cyclic.com>
* sighandle.c (SIG_register): Use memset not bzero.
* system.h: Remove defines for index, rindex, bcmp, and bzero.
All the calls to those functions are gone from CVS.
Tue Jan 2 13:00:00 1996 Jim Kingdon <kingdon@peary.cyclic.com>
Visual C++ lint:
* sighandle.c: Prototype SIG_handle and SIG_defaults.
Use SIG_ERR where appropriate.
Mon Dec 18 10:15:05 1995 Jim Kingdon <kingdon@harvey.cyclic.com>
* rename.c: Check ENOENT rather than existence_error. The latter
is undefined in this file, and including system.h is said to cause
(unspecified) problems.
Sun Dec 17 23:58:06 1995 Jim Kingdon <kingdon@harvey.cyclic.com>
* vasprintf.c: Removed (it is no longer used).
* Makefile.in (SOURCES): Remove vasprintf.c.
Sat Dec 16 17:18:33 1995 Jim Kingdon <kingdon@harvey.cyclic.com>
* vasprintf.c: Added.
* Makefile.in (SOURCES): Add vasprintf.c
Mon Dec 4 10:54:04 1995 Jim Kingdon <kingdon@harvey.cyclic.com>
* getdate.c: Remove #line directives. I know, this is a kludge,
but Visual C++ 2.1 seems to require it (why, I have no idea. It
has no trouble with the #line directives in getdate in CVS 1.6).
Sat Nov 18 16:20:37 1995 Karl Fogel <kfogel@floss.cyclic.com>
* rename.c: same.
* mkdir.c: Use new macro `existence_error', instead of comparing
errno to ENOENT directly.
* system.h (existence_error): new macro, tries to portably ask if
errno represents a file-not-exist error.
Fri Nov 17 20:08:58 1995 Karl Fogel <kfogel@floss.cyclic.com>
* system.h (NEED_DECOY_PERMISSIONS): moved this section to where
it belongs, duh.
* getdate.c: if STDC_HEADERS, then just include <stdlib.h> instead
of declaring malloc() and realloc() to be char *.
* system.h: ifdef NEED_DECOY_PERMISSIONS, then define the S_I*
permission masks for USR, GRP, and OTH in terms of the simpler
OS/2 masks.
Wed Nov 15 15:36:03 1995 Karl Fogel <kfogel@floss.cyclic.com>
* system.h: ifdef USE_OWN_TCPIP_H, then include "tcpip.h". Only
OS/2 does this right now.
Tue Nov 14 18:44:57 1995 Greg A. Woods <woods@most.weird.com>
* getdate.c: OK, this one is from SunOS-4.1 yacc and may be more
portable -- at least it compiles silently here! ;-)
Mon Nov 13 03:53:45 1995 Karl Fogel <kfogel@floss.cyclic.com>
* fnmatch.c: conform to 80 column standard (yes, I'm a pedant).
Wed Nov 8 11:10:59 1995 Karl Fogel <kfogel@floss.cyclic.com>
* system.h (STAT_MACROS): ifdef S_IFMT, then use it as before; but
if it's not defined, then just do a single mask and assume
acceptance any of non-zero result. Norbert, I trust you'll let me
know if this is unsatisfactory. :-)
Ifdef HAVE_SYS_UTIME_H, then include <sys/utime.h>. Only OS/2
defines this right now.
Wed Nov 8 13:18:51 1995 Norbert Kiesel <nk@col.sw-ley.de>
* valloc.c: omit malloc declaration (it's already in system.h
which is included and conflicts with <stdlib.h> on some
systems).
Tue Nov 7 19:38:48 1995 Norbert Kiesel <nk@col.sw-ley.de>
* system.h (STAT_MACROS_BROKEN): undo previous change, because
else all regular files will be identified as links (the mask for
links is S_IFREG|S_IFCHR).
Mon Nov 6 19:20:56 1995 Karl Fogel <kfogel@floss.cyclic.com>
* system.h (STAT_MACROS_BROKEN): in defining the S_IF* macros,
don't fold to 1 or 0 by first masking with S_IFMT; not all
systems have that macro, and anyway it's only necessary that we
return non-zero.
Fri Oct 27 13:43:35 1995 Karl Fogel <kfogel@floss.cyclic.com>
* save-cwd.c: use __PROTO instead of __P (see below).
* getline.h (__PROTO): same as below.
* save-cwd.h (__PROTO): replaces __P. New name, so don't ask if
already defined. The conflict was that OS/2 w/ IBM C/C++ uses
`__P' for something else, in <ctype.h> of all places.
* system.h: do nothing about alloca ifdef ALLOCA_IN_STDLIB (see
../src/ChangeLog).
Tue Oct 24 13:01:25 1995 Norbert Kiesel <nk@col.sw-ley.de>
* wait.h: include sys/resource.h if available. This is needed at
least under AIX-3.2 where <sys/wait.h> doesn't include it.
Mon Oct 23 17:39:11 1995 Norbert Kiesel <nk@col.sw-ley.de>
* valloc.c (valloc): change parameter definition
Sun Oct 22 14:15:44 1995 Jim Meyering (meyering@comco.com)
* getline.c, getline.h: New files.
* Makefile.in (SOURCES, OBJECTS, HEADERS): Add getline.c, getline.o,
and getline.h, respectively.
Tue Oct 10 18:01:50 1995 Karl Fogel <kfogel@totoro.cyclic.com>
* Makefile.in (cvs_srcdir): define cvs_srcdir to be ../src, then
include it with -I so save_cwd.c can find error.h (for example).
Sun Oct 8 12:27:57 1995 Peter Wemm <peter@haywire.DIALix.COM>
* system.h: define POSIX_SIGNALS or BSD_SIGNALS if configure has
located all the necessary functions for each "type".
* sighandle.c: detect/use POSIX/BSD reliable signals (especially
for blocking signals in critical sections). Helps prevent stray
locks on interruption.
Mon Oct 2 18:11:23 1995 Jim Blandy <jimb@totoro.cyclic.com>
* system.h: Doc fix.
Mon Oct 2 18:10:35 1995 Larry Jones <larry.jones@sdrc.com>
* regex.c: compile 4.2 BSD compatible functions even when
_POSIX_SOURCE is defined since we need them and we wouldn't be
compiling this file unless they don't exist.
Mon Oct 2 10:32:20 1995 Michael Finken <finken@conware.de>
* strstr.c (strstr): new file and func.
* Makefile.in (SOURCES): added strstr.c.
Sun Oct 1 21:03:40 1995 Karl Fogel <kfogel@totoro.cyclic.com>
* regex.c: reverted below change.
Thu Sep 28 13:37:04 1995 Larry Jones <larry.jones@sdrc.com>
* regexp.c: check for ISC.
Thu Sep 7 19:18:00 1995 Jim Blandy <jimb@cyclic.com>
* save-cwd.c: #include <direct.h> and <io.h>, on systems that
have them.
* getopt.c (_getopt_internal): Cast the return value of strlen,
which is unsigned, before comparing it with the difference between
two pointers, which is unsigned.
Thu Aug 31 11:31:42 1995 Jim Blandy <jimb@totoro.cyclic.com>
* getdate.y [STDC_HEADERS]: #include <stdlib.h>, for abort.
[HAVE_ALLOCA_H]: #include <alloca.h>, for alloca on Windows NT.
Wed Aug 30 18:48:44 1995 Jim Blandy <jimb@totoro.cyclic.com>
* system.h [HAVE_IO_H]: #include <io.h>, for Windows NT.
[HAVE_DIRECT_H]: #include <direct.h>, for Windows NT.
(CVS_MKDIR, FOLD_FN_CHAR, fnfold, fncmp, ISDIRSEP, OPEN_BINARY,
FOPEN_BINARY_READ, FOPEN_BINARY_WRITE): New macros/functions, for
use in system-sensitive code.
* regex.c (re_set_registers): start and end are pointers, not
integers. Cast the initializing value appropriately.
* getopt.c [HAVE_STRING_H]: #include <string.h>, to avoid
warnings.
* fnmatch.c (FOLD_FN_CHAR): Give this a dummy #definition if
config.h didn't #define it.
(fnmatch): Pass filename characters through FOLD_FN_CHAR before
comparing them.
* argmatch.c: #include <sys/types.h>.
(argmatch): Declare arglen to be a size_t, rather than an int,
to avoid signed/unsigned comparison "problems".
* .cvsignore: Remove getdate.c from this file. We want to
distribute it, for systems that don't have a Yacc-equivalent
installed (like Windows NT).
Sat Aug 19 22:00:51 1995 Jim Blandy <jimb@totoro.cyclic.com>
* error.c: Don't #define CVS_SUPPORT here. config.h takes care of
that for us.
[CVS_SUPPORT] (error_use_protocol): New variable, with apology.
(error): If error_use_protocol is set, report errors using the
client/server protocol.
* error.h [CVS_SUPPORT]: Extern decl for error_use_protocol.
Fri Aug 4 00:01:24 1995 Jim Meyering (meyering@comco.com)
* xgetwd.c: Don't declare free. A K&R style declaration gets
a conflict on some Sun systems when compiling with acc.
* save-cwd.c: New file.
* save-cwd.h: New file.
* Makefile.in (SOURCES): Add save-cwd.c
(OBJECTS): Add save-cwd.o.
(HEADERS): Add save-cwd.h.
Thu Aug 3 00:55:54 1995 Jim Meyering (meyering@comco.com)
* error.h: New file.
* Makefile.in (HEADERS): Add error.h.
Sat Jul 29 15:53:55 1995 James Kingdon <kingdon@harvey.cyclic.com>
* Makefile.in (SOURCES): Add getdate.c.
Thu Jul 27 09:11:41 1995 Robert Lipe <robertl@rjlhome.arnet.com>
* system.h: Check for PATHSIZE before falling back to _POSIX_PATH_MAX.
Thu Jul 20 12:38:03 1995 James Kingdon <kingdon@harvey.cyclic.com>
* error.c: Instead of calling cvs functions to clean up, allow cvs
to register a callback via error_set_cleanup. Avoids hassles with
include files and SERVER_SUPPORT and so on.
Tue Jul 18 21:18:00 1995 Jim Blandy <jimb@cyclic.com>
* system.h: Include <sys/param.h> only if HAVE_SYS_PARAM_H
is #defined. We've added a test to configure.in to #define this
on most systems.
Thu Jul 13 11:22:21 1995 Jim Meyering (meyering@comco.com)
* xgetwd.c: New file.
* Makefile.in (SOURCES): Add xgetwd.c
(OBJECTS): Add xgetwd.o.
Wed Jul 12 09:18:49 1995 Jim Meyering (meyering@comco.com)
* Makefile.in (OBJECTS): Remove fnmatch.o. Now configure adds it
to LIBOBJS when necessary.
Fri Jun 30 16:27:18 1995 James Kingdon <kingdon@harvey.cyclic.com>
* rename.c (rename): If MVDIR is not defined, just give an error
on attempt to rename a directory.
Thu Jun 29 00:46:31 1995 James Kingdon <kingdon@harvey.cyclic.com>
* system.h: Check HAVE_SYS_TIMEB_H not non-existent HAVE_TIMEB_H.
* system.h: Don't define alloca if it is already defined.
Wed Jun 28 15:24:51 1995 James Kingdon <kingdon@harvey.cyclic.com>
* system.h: If NeXT, define utimbuf ourself.
Mon May 29 22:32:40 1995 J.T. Conklin <jtc@rtl.cygnus.com>
* system.h: Handle time and directory headers as recommended in
the autoconf manual.
Undefine the S_FOO() macros if STAT_MACROS_BROKEN is set.
Don't define mode_t, as it is handled by config.h.
Sat May 27 08:46:00 1995 Jim Meyering (meyering@comco.com)
* Makefile.in (Makefile): Regenerate only Makefile in current
directory when Makefile.in is out of date. Depend on ../config.status.
Fri Apr 28 22:49:25 1995 Jim Blandy <jimb@totoro.bio.indiana.edu>
* Makefile.in (SOURCES, OBJECTS): Updated.
(HEADERS): New variable.
(DISTFILES): Updated.
(dist-dir): Renamed from dist; changed to work with DISTDIR
variable passed from parent.
Wed Feb 8 06:37:53 1995 Roland McGrath <roland@churchy.gnu.ai.mit.edu>
* system.h (S_IRUSR et al): Define if not already defined.
* waitpid.c [HAVE_CONFIG_H]: Include "config.h".
(ualloc): Return OLDPTR rather than running off the end.
Mon Aug 22 22:48:19 1994 Ken Raeburn (raeburn@kr-pc.cygnus.com)
* error.c (strerror): Replaced conditional static definition
(always used, since the condition variable was never set) with an
extern declaration, since it's provided by libc or strerror.c.
Wed Aug 10 14:54:25 1994 Ken Raeburn (raeburn@cujo.cygnus.com)
* Makefile.in (SOURCES): Add waitpid.c.
* waitpid.c: New file.
Tue Aug 9 16:00:12 1994 Ken Raeburn (raeburn@cujo.cygnus.com)
* md5.h (uint32): If SIZEOF_LONG isn't 4, don't define this to be
"unsigned long"; try SIZEOF_INT and "unsigned int", otherwise
complain.
* md5.c: Include config.h.
(const): Don't bother defining here, config.h should take care of
it.
* valloc.c (malloc): Declare.
Fri Jul 15 12:57:20 1994 Ian Lance Taylor (ian@sanguine.cygnus.com)
* getopt.c: Do not include <stdlib.h> unless __GNU_LIBRARY__ is
defined. On Irix 5.2, <stdlib.h> includes <getopt.h>, which
causes a multiple definition of struct option.
Fri Jul 8 10:04:59 1994 Jim Kingdon (kingdon@lioth.cygnus.com)
* md5.h, md5.c: Remove ANSI-isms.
Thu Jul 7 20:24:18 1994 Ian Lance Taylor (ian@sanguine.cygnus.com)
* md5.h, md5.c: New files.
* Makefile.in (SOURCES): Add md5.c.
(OBJECTS): Add md5.o.
(DISTFILES): Add md5.h.
(md5.o): New target; depend upon md5.h.
Fri May 27 18:15:34 1994 Ian Lance Taylor (ian@tweedledumb.cygnus.com)
* valloc.c: New file.
Tue May 17 08:18:26 1994 Jim Kingdon (kingdon@lioth.cygnus.com)
* error.c (error, fperror): If server_active, call server_cleanup
as well as Lock_Cleanup.
Thu Jan 6 13:45:04 1994 Ken Raeburn (raeburn@cujo.cygnus.com)
* system.h: Fix Dec 27 change to work correctly. Makes Sep 9
change unnecessary, so backed that one out. Never define PATH_MAX
in terms of pathconf, because that doesn't produce a constant, and
PATH_MAX is used to set array sizes.
Mon Dec 27 14:22:07 1993 Mark Eichin (eichin@cygnus.com)
* system.h: don't touch PATH_MAX or MAXPATHLEN if *both* of them
are already defined, as one may be defined in terms of the other.