_This_ branch does not contain actively-developed code. More likely,
you want to check out branch-1.4 (latest stable), branch-1.6 (some
speedups over 1.4, but not yet released), or branch-2.0 (many more
experimental changes, and even further from release).
* README: Document the status of this git repository.
Follow the practice set in numerous other GNU projects, where
the ChangeLog (since 2015) is generated from git commit messages.
This avoids duplication or subtle differences between the two,
as well as making it easier to merge patches across branches
(as good as Bruno Haible's 'git-merge-changelog' helper program
is, it still doesn't handle cross-branch cherry-picks very well).
* Makefile.am (EXTRA_DIST): Rely on gnulib to ship gitlog-to-changelog.
($(changelog)): Update rule to match branch-1.4 style.
* .gitattributes: Drop stale mention of Changelog.
* HACKING: Reword to match new procedure, and simplify by
referring to an external description of ChangeLog style.
Signed-off-by: Eric Blake <eblake@redhat.com>
(cherry picked from commit 5456676ea90dfedd9b54c843f15b006d02150d06)
[commit 6bdb0a4b75 already started the conversion, but this updates
the approach to be more like other GNU projects, and branch-1.4]
Commit b29e2867 added gnulib's src/assure.h during bootstrap to fix
development builds, but forgot to check that tarballs also build.
With this patch, things get further, although I'm still facing a
failure because the VPATH 'make check' portion of distcheck is
failing to load modules, where an in-tree build succeeded.
* Makefile.am (src_m4_SOURCES): Add assure.h.
Signed-off-by: Eric Blake <eblake@redhat.com>
* NEWS: Add some blurbs.
* .prev-version: Record previous version.
* cfg.mk (old_NEWS_hash): Update.
Signed-off-by: Eric Blake <eblake@redhat.com>
[this is a mish-mash of two patches from branch-1.4:]
(cherry picked from commit 658427104bbe515249f728d8524e975167e6533d)
(cherry picked from commit e0140bd91ce0d2dfcdd1df22472d81ed13489d46)
The user still has full control over verbosity levels, both setting
their per-project defaults at configure time (or even in a config.site
file), as well as a per-run override. But these days, most projects
are defaulting to silent rules without user intervention.
* configure.ac (AM_SILENT_RULES): Add, to match what most projects
are doing these days.
Signed-off-by: Eric Blake <eblake@redhat.com>
(cherry picked from commit b4a08ecfd34e703bd846f1e85b4e2a94a34b16d9)
[For the cherry-pick, all we have to do is normalize the spelling
of the line - now that we require new-enough automake, it does not
have to be conditional]
* configure.ac (AM_INIT_AUTOMAKE): Require 1.11.6 or newer.
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
(cherry picked from commit a82e6b74a462b16effb7304b850f72faf727aec5)
Ever since commit f1cf390 (1.4.14 release), we no longer create
diff files as part of the release process. These days, it is
assumed that it is easier to download a fresh tarball rather
than to try and use a diff file to patch an older tarball.
Signed-off-by: Eric Blake <eblake@redhat.com>
(cherry picked from commit c8a4b7c32db3a6d8d5dc7a9997df2c80ac95b030)
texinfo 6.1 complains (during 'make dvi'):
/home/eblake/m4-1.4/doc/./m4.texi:9: @setcontentsaftertitlepage has been remove
d as a Texinfo command; move your @contents command if you want the contents af
ter the title page..
It turns out that eliminating the command has no effect - modern
tools correctly emit the contents in-place, right after the title
page, so it was leftover cruft from an older time.
* doc/m4.texi: Satisfy newer texinfo.
Signed-off-by: Eric Blake <eblake@redhat.com>
(cherry picked from commit c7a3591347238aa34defded34609007b326a0984)
Needed to reduce number of 'make syntax-check' failures.
* gnulib: Update to latest.
* bootstrap.conf (src_modules): Add 'assure', to match gnulib module
changes in the meantime.
* .gitignore: Ignore results of assure module.
* cfg.mk (gnulib_dir): Declare our alternate location.
* bootstrap: Regenerate.
* all files: Use 'make update-copyright' to bump year.
Signed-off-by: Eric Blake <eblake@redhat.com>
(cherry picked from commit 102376916e49b6aae85dc6ff0b59699205a74bdf)
[Ultimately, this short-circuits all other commits from other branches
that touched gnulib to any intermediate version]
* m4/path.c (m4_path_search): s/FILE_TRUNCATE/TRUNCATE_FILENAME/
to match definition.
Reported by KO Myung-Hun
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
* m4/path.c (m4_path_search): Ensure that a null-terminator is
appended. Re-calculate a length of path after path_truncate().
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
Notably this moves to the new 64bit obstack implementation.
* gnulib: Sync with upstream, for maint.mk bug fixes et al.
* build-aux/m4/gnulib-cache.m4, po/Makevars: Regenerate.
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
In preparation for gnulib API change to obstacks, where
obstack_blank cannot shrink memory use any more.
* m4/macro.c (trace_flush): `start - len` can be negative, so
call obstack_blank_fast, which accepts negative arguments.
* modules/m4.c (m4_make_temp): Likewise... -1 is always negative!
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
* configure.ac (fn_module_cmds): Split ~ delimited command lists,
and set globals required by $module_cmds and $archive_cmds.
(libname): Preload with 'conftmod' ready for fun_module_cmds
calls, and for reuse as an 8.3 compatible module basename in this
check.
(sys_dlsym_uscore): Use it to detect underscore prefix requirement
in some dlysm() implementations more reliably.
Reported by KO Myung-Hun
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
* m4/module.c (m4__module_open): Use LT_MODULE_EXT instead of
no longer calculated SYS_SHREXT.
* m4/path.c (FILE_SUFFIXES): Likewise.
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
* configure.ac (LT_SYS_MODULE_EXT): Use this...
(mfour_shrext): Remove. ...instead of working it out manually.
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
* Makefile.am (m4_libm4_la_LIBADD): Use modern LIBADD_DLOPEN
instead of obsolete and undocumented LIBADD_DL.
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
* configure.ac (SYS_SHREXT): Define to host system module file
extension.
* m4/module.c (suffixes), m4/path.c (FILE_SUFFIXES): Use it
instead of listing suffixes for all supported host systems.
Reported by Pavel Raiskup
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
* m4/module.c (uscore_sym): Change NEED_USCORE guarde to
LDSYM_USCORE.
* configure.ac: Move compiler characteristics checks above libtool
initialization section.
(Dynamic Loader Characteristics): Determine whether dlsym requires
a leading symbol name underscore, without requiring that self
dlopening works like LT_FUNC_DLSYM_USCORE macro does.
Reported by KO Myung-Hun
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
Moving the 'doc' related Makefile targets into doc/Makefile.am to
avoid building manual page before the m4 binary itself (suggested
by automake.info).
* Makefile.am (SUBDIRS): Add doc subdir.
* Makefile.am: Move to docs-related rules from here...
* doc/Makefile.am: New file. ...to here.
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
M4 no longer uses libltdl.
* configure.ac: Call LT_LIB_DLLOAD to set LIBADD_DL correctly.
* Makefile.am (m4_libm4_la_LIBADD): Remove LIBLTDL.
(m4_libm4_la_DEPENDENCIES): Remove LTDLDEPS.
Reported by Pavel Raiskup
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
* m4/path.c (TRUNCATE_FILENAME): New macro, defined on OS2.
(path_truncate): New function when TRUNCATE_FILENAME is defined,
otherwise a null-operation macro. Trim the basename to no more
than 8 characters, followed by the extension.
(m4_path_search): Use it.
Reported by Ko Myung-Hun
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
The original patch had security issues, and broke loading of
frozen files including modules that are only available along
the fall-back path.
This reverts commit 1d887c9b5002d8e9a4e65c192c8a1b7355cc356a.
* configure.ac (PATH_SEPARATOR): Propagate to config.h.
* m4/path.c (search_path_env_init): Respect PATH_SEPARATOR.
Reported by KO Myung-Hun
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
* configure.ac (LT_FUNC_DLSYM_USCORE): Call it to set NEED_USCORE.
* m4/module.c (uscore_sym): New function. Use this to prepend an
underscore if NEED_USCORE was set at configure time.
Reported by KO Myung-Hun
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
* m4/module.c (m4__module_open): If the module search function
cannot find a matching module, fall-back to the system dlopen
search path.
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
* m4/module.c (m4__module_open): Support .dll suffix in module
names.
* m4/path.c (FILE_SUFFIXES): Likewise for include builtin.
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
On OS/2 kLIBC, fdopen() creates a stream in a mode of a file
descriptor. So specify "t" to open a stream in a text mode explicitly
on OS/2.
* modules/gnu.c (m4_esyscmd): fdopen() in a text mode on OS/2.
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
-Zargs-resp enables a response file support on OS/2.
* configure.ac (OS2_LDFLAGS): Add -Zargs-resp on *-*-os2*.
* Makefile.am (AM_LDFLAGS): Add OS2_LDFLAGS.
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
Without $EXEEXT, /bin/sh cannot be located on OS/2.
* build-aux/m4/m4-syscmd.m4 (M4_SYSCMD): append $EXEEXT suffix
to /bin/sh.
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
Modules are loaded by libltdl. Arguably, loading from the .la
pseudo-library should work irrespective of platform library object
file-naming schemes; but in reality, M4 loads modules by basename
anyway, so we shouldn't put a platform specific version number in
the module filename anyway.
* Makefile.am (module_ldflags): Add -avoid-version flag.
Reported by KO Myung-Hun
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
For compatibility with OS/2 and cygwin, modules must be compiled
with no undefined symbols.
* Makefile.am (AM_LDFLAGS): Add -no-undefined flag.
Reported by Ko Myung-Hun
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
More automation == less time wasted on menial tasks.
* build-aux/thanks-gen: script inspired by coreutils.
* Makefile.am (THANKS): Based on rule from coreutils/Makefile.am.
* NO-THANKS: New file. Configure thanks-gen output.
* THANKS: Remove.
Signed-off-by: Gary V. Vaughan <gary@gnu.org>
