Paul Eggert [Fri, 3 Feb 2023 20:25:08 +0000 (12:25 -0800)]
Clarify NEWS
* NEWS: Clarify recent items about AC_SYS_LARGEFILE_REQUIRED,
AC_SYS_YEAR2038_REQUIRED, AC_PROG_MKDIR_P. Reported by Bruno Haible in:
https://lists.gnu.org/r/bug-autoconf/2023-02/msg00005.html
Paul Eggert [Thu, 2 Feb 2023 21:18:24 +0000 (13:18 -0800)]
Improve year-2038 documentation
* NEWS, doc/autoconf.texi (System Services):
Improve documentation for behavior of largefile and year-2038 support.
Say that in the current implementation, year-2038 support
requires largefile support. Say that year-2038 support
matters only for GNU/Linux glibc 2.34+ on 32-bit x86 and ARM.
Prefer brevity when this does not hurt understandability;
for example, prefer active to passive voice.
Prefer “wider” to “larger” when talking about the number of
bits in an integer, as this terminology is more standard.
Tone down the wording in warnings about enabling year-2038 support,
use similar wording in warnings about enabling largefile support,
and warn also about disabling largefile and year-2038 support.
No need for @emph. Also mention rlim_t.
Be a bit more careful about saying “2 GiB” rather than “2 GB”.
Mention that a future version of Autoconf might change
AC_SYS_LARGEFILE to default to --enable-year2038, since
something has gotta happen before 2038.
Coalesce descriptions of --enable-largefile and --enable-year2038
to simplify documentation. Mention that the only system where
AC_SYS_LARGEFILE changes CC is IRIX and that these systems
are obsolete. Say that ‘stat’ can fail due to time_t
overflow. Say that you can’t portably print time_t with %ld.
Say that binary compatibilty problems also can occur when one
library is linking to amother; it’s not just apps vs libraries.
Mention the possibility of modifying libraries to support both
32- and 64-bit interfaces. Warn more consistently about
ABI compatibility issues, but put the bulk of this text
in one location that the other locations refer to.
Todd C. Miller [Mon, 26 Dec 2022 17:35:39 +0000 (10:35 -0700)]
fix AC_CHECK_HEADER_STDBOOL regression
Commit 6dcecb780a69bd208088d666b299e92aa7ae7e80 "Port
AC_CHECK_HEADER_STDBOOL to C23" causes AC_CHECK_HEADER_STDBOOL to
always fail, even on systems with a conforming stdbool.h.
There is no longer an 'a' variable so it should not be referenced
in the return statement.
Paul Eggert [Sun, 25 Dec 2022 07:24:54 +0000 (23:24 -0800)]
Omit just-added *_REQUIRED macros
They are not needed for Gnulib, and users have an easy way to get
their effect, so for now omit them and just document the easy way.
Also, redo documentation to make it clear that AC_YEAR_2038 is
like AC_SYS_LARGEFILE except with a different year-2038 default.
* NEWS, doc/autoconf.texi: Document the above.
* lib/autoconf/specific.m4 (AC_SYS_YEAR2038_REQUIRED):
(AC_SYS_LARGEFILE_REQUIRED): Remove. Remove some support code.
Perhaps further simplification could be done but I quit while
I was ahead.
Zack Weinberg [Sun, 25 Dec 2022 07:24:51 +0000 (23:24 -0800)]
AC_SYS_LARGEFILE: Don’t enlarge time_t by default
Having AC_SYS_LARGEFILE enlarge time_t means that any program that has
already requested large file support will be abruptly migrated to
64-bit time_t (on 32-bit systems) as soon as its configure script is
regenerated with a sufficiently new Autoconf. We’ve received reports
of several widely used programs and libraries that are not prepared
for this migration, with breakage ranging from annoying (garbage
timestamps in messages) through serious (binary compatibility break
in security-critical shared library) to catastrophic (on-disk data
corruption).
Partially revert f6657256a37da44c987c04bf9cd75575dfca3b60: in the
absence of AC_SYS_YEAR2038, AC_SYS_LARGEFILE will now only add an
--enable-year2038 command line option to configure. If this option is
used, time_t will be enlarged, allowing people to experiment with the
migration without needing to *edit* the configure script in question,
only regenerate it.
In the process, AC_SYS_LARGEFILE and AC_SYS_YEAR2038 were drastically
overhauled for modularity; it should now be much easier to add support
for platforms that offer large off_t / time_t but not with the standard
feature selection macros. Also, new macros AC_SYS_LARGEFILE_REQUIRED and
AC_SYS_YEAR2038_REQUIRED can be used by programs for which large off_t /
time_t are essential.
The implementation is a little messy because it needs to gracefully
handle the case where AC_SYS_LARGEFILE and AC_SYS_LARGEFILE_REQUIRED
are both used in the same configure script — or, probably more common,
AC_SYS_LARGEFILE (which invokes _AC_SYS_YEAR2038_OPT_IN) followed by
AC_SYS_YEAR2038 — but if macro B is invoked after macro A, there’s no
way for B to change *what macro A expanded to*. The best kludge I
managed to find is to AC_CONFIG_COMMANDS_PRE as a m4-level hook that
sets shell variables in an early diversion.
* lib/autoconf/functions.m4 (AC_FUNC_FSEEKO): Rewrite to avoid dependency
on internal subroutines of AC_SYS_LARGEFILE.
* lib/autoconf/specific.m4 (_AC_SYS_YEAR2038_TEST_INCLUDES): Renamed to
_AC_SYS_YEAR2038_TEST_CODE.
(_AC_SYS_YEAR2038): Refactor into subroutines: _AC_SYS_YEAR2038_OPTIONS,
_AC_SYS_YEAR2038_PROBE, _AC_SYS_YEAR2038_ENABLE.
(AC_SYS_YEAR2038): Update for refactoring.
(_AC_SYS_YEAR2038_OPT_IN): New sorta-top-level macro, for use by
AC_SYS_LARGEFILE, that probes for large time_t only if the
--enable-year2038 option is given.
(AC_SYS_YEAR2038_REQUIRED): New top-level macro that insists on
support for large time_t.
(_AC_SYS_LARGEFILE_TEST_INCLUDES): Renamed to _AC_SYS_LARGEFILE_TEST_CODE.
(_AC_SYS_LARGEFILE_MACRO_VALUE, AC_SYS_LARGEFILE): Refactor along same
lines as above: _AC_SYS_LARGEFILE_OPTIONS, _AC_SYS_LARGEFILE_PROBE,
_AC_SYS_LARGEFILE_ENABLE. Invoke _AC_SYS_YEAR2038_OPT_IN at end of
_AC_SYS_LARGEFILE_PROBE. MinGW-specific logic moved to YEAR2038
macros as it has nothing to do with large file support.
(AC_SYS_LARGEFILE_REQUIRED): New top-level macro that insists on
support for large off_t.
* tests/local.at (_AT_CHECK_ENV): Also allow changes in CPPFLAGS,
enableval, enable_*, withval, with_*.
* doc/autoconf.texi, NEWS: Update documentation to match above changes.
Fix typo in definition of @dvarv.
Paul Eggert [Wed, 16 Nov 2022 18:19:43 +0000 (10:19 -0800)]
Cater better to underquoted callers
Problem reported by Frederic Berat in:
https://lists.gnu.org/r/autoconf/2022-11/msg00092.html
* lib/autoconf/c.m4 (AC_LANG_CALL(C)): Omit comma in comment,
which triggers bugs in underquoted callers.
Paul Eggert [Fri, 11 Nov 2022 08:28:49 +0000 (00:28 -0800)]
Modernize and regularize doc for C89 etc.
In documentation and comments, prefer the more-common “C89” to the
equivalent “C90”, and use 2-digit years for C standards as that’s
common usage. Remove some confusing old doc for pre-C89 systems,
as Autoconf assumes C89 or later. Mention C17 and C23 briefly.
Improve doc for malloc, realloc.
Zack Weinberg [Thu, 10 Nov 2022 17:05:30 +0000 (12:05 -0500)]
More fixes for compilers that reject K&R function definitions.
This fixes all of the remaining failures exposed by running the
testsuite with GCC 12 and
CC='cc -Wimplicit-function-declaration -Wold-style-definition
-Wimplicit-int -Werror'
.
* lib/autoconf/c.m4 (_AC_C_C89_TEST_GLOBALS): Don’t use K&R function
definitions.
* lib/autoconf/specific.m4 (AC_SYS_RESTARTABLE_SYSCALLS): Likewise.
(AC_FUNC_MKTIME): Declare functions taking no arguments as ‘fn (void)’
not ‘fn ()’.
* lib/autoconf/c.m4 (_AC_C_C99_TEST_GLOBALS): Declare free().
Paul Eggert [Fri, 28 Oct 2022 22:19:16 +0000 (15:19 -0700)]
Warn about TMPDIR
* doc/autoconf.texi (Initialization Macros): Warn about
setting TMPDIR, following up on this Paul Smith remark:
https://lists.gnu.org/r/autoconf/2022-10/msg00024.html
Paul Eggert [Wed, 12 Oct 2022 19:01:50 +0000 (12:01 -0700)]
MKDIR_P now falls back on mkdir -p
* lib/autoconf/programs.m4 (AC_PROG_MKDIR_P):
Fall back on mkdir -p instead of on a relative path to
install-sh, as the latter now seems to be more of a problem
than the former.
Xi Ruoyao [Fri, 30 Sep 2022 15:36:03 +0000 (11:36 -0400)]
Fix testsuite failures with bash 5.2.
Bash v5.2 includes several new optimizations to the number of
subshells used for various constructs; as a side effect, the value of
SHLVL is less stable than it used to be. Add SHLVL to the list of
built-in shell variables with unstable values, to be ignored when
checking for inappropriate differences to the shell environment
before and after a macro invocation / between two configure runs.
Problem and solution reported by Xi Ruoyao in
https://lists.gnu.org/archive/html/autoconf/2022-09/msg00015.html
Problem also reported by Bruce Dubbs in
https://lists.gnu.org/archive/html/bug-autoconf/2022-09/msg00010.html
* tests/local.at (_AT_CHECK_ENV, AT_CONFIG_CMP): Ignore changes in
value of SHLVL.
Copyright-paperwork-exempt: yes
Paul Eggert [Sun, 18 Sep 2022 02:43:04 +0000 (19:43 -0700)]
Allow users to clear BINSH
Problem and solution reported by Roger Crew in:
https://lists.gnu.org/r/bug-autoconf/2022-09/msg00005.html
* lib/m4sugar/m4sh.m4 (_m4_divert(BINSH)): Change from 0 to 1,
and add 2 to HEADER-REVISION, HEADER-COMMENT, HEADER-COPYRIGHT,
M4SH-SANITIZE, M4SH-INIT-FN, and M4SH-INIT.
Copyright-paperwork-exempt: yes
autoconf: store autom4te request keys in sorted order
Notced unstable key order when debugging unrelated bug.
Data::Dumper's SortKeys() makes ordering stable and decreases
diffs from run to run. No functional change otherwise.
* lib/Autom4te/C4che.pm
* lib/Autom4te/Request.pm: Sort request keys in serialization.
Paul Eggert [Fri, 9 Sep 2022 21:54:11 +0000 (16:54 -0500)]
Port to GNU make 4.4
* cfg.mk (PATH): Assign eagerly, and simplify shell use, avoiding
use of the shell entirely if PWD is set, as it should be.
Problem reported by Sergei Trofimovich in:
https://lists.gnu.org/r/autoconf-patches/2022-09/msg00007.html
Paul Eggert [Thu, 1 Sep 2022 21:19:50 +0000 (16:19 -0500)]
Port to compilers that moan about K&R func decls
* lib/autoconf/c.m4 (AC_LANG_CALL, AC_LANG_FUNC_LINK_TRY):
Use '(void)' rather than '()' in function prototypes, as the latter
provokes fatal errors in some compilers nowadays.
* lib/autoconf/functions.m4 (AC_FUNC_STRTOD):
* tests/fortran.at (AC_F77_DUMMY_MAIN usage):
* tests/semantics.at (AC_CHECK_DECLS):
Don’t use () in a function decl.
Paul Eggert [Mon, 8 Aug 2022 21:20:41 +0000 (14:20 -0700)]
AC_SYS_LARGEFILE now widens time_t if possible
This change is taken from Gnulib, and is needed for apps like GDB.
* lib/autoconf/specific.m4 (_AC_SYS_YEAR2038_TEST_INCLUDES)
(_AC_SYS_YEAR2038, AC_SYS_YEAR2038): New macros, taken (with
renaming) from Gnulib.
(_AC_SYS_LARGEFILE_MACRO_VALUE): #undef before #define.
(AC_SYS_LARGEFILE): Prefer AS_IF and AS_CASE to doing it by hand.
Widen time_t if possible, too. Define __MINGW_USE_VC2005_COMPAT
early if needed.
Paul Eggert [Mon, 8 Aug 2022 16:08:20 +0000 (09:08 -0700)]
Fix compilation errors in 32-bit C++ mode on HP-UX 11/ia64
This should render unnecessary Gnulib’s gl_SET_LARGEFILE_SOURCE hack
https://lists.gnu.org/r/bug-gnulib/2019-12/msg00046.html
* lib/autoconf/functions.m4 (AC_FUNC_FSEEKO):
On that platform, simply set _LARGEFILE_SOURCE.
Paul Eggert [Wed, 27 Jul 2022 15:53:35 +0000 (08:53 -0700)]
Port AC_C_BIGENDIAN to cross gcc -std=c11 -flto
* lib/autoconf/c.m4 (AC_C_BIGENDIAN): Improve the
inherently-unportable grep trick well enough to survive gcc
-std=c11 -flto when cross-compiling (sr#110687).
Paul Eggert [Thu, 14 Jul 2022 19:24:20 +0000 (12:24 -0700)]
Handle z/OS Unix ‘set’ output
Problem and fix reported by Jeff Shimoda in:
https://lists.gnu.org/r/autoconf-patches/2022-06/msg00007.html
* tests/local.at (AT_CONFIG_CMP): Treat ‘var=""’ like ‘var=''’.
Copyright-paperwork-exempt: Yes
Tests of AC_{CHECK_,}HEADER_STDBOOL were failing on Darwin for two
reasons: an m4 quoting bug in tests/local.at causing Darwin sed to
throw syntax errors, and an excessively precise interpretation of how
C99 and C++11 interact. The latter is worth mentioning in NEWS.
* tests/local.at (_AT_DEFINES_CMP_PRUNE): Insert [] before ‘dnl’ in
two places to keep it separate from the result of the m4_bpatsubsts
operation that immediately precedes it.
* lib/autoconf/headers.m4 (AC_CHECK_HEADER_STDBOOL): Allow ‘bool’,
‘true’, and ‘false’ not to be macros, after including stdbool.h,
whenever __cplusplus is defined.
* NEWS: Document change to AC_{CHECK_,}HEADER_STDBOOL.
The main issue was use of s///r; the /r modifier was added in 5.14.
Since the baseline is now 5.10.0 instead of 5.6.0, a couple places
can be tidied up using slightly newer constructs.
* help-extract.pl: Use File::Spec::Functions instead of File::Spec
for reduced clunkiness.
(eval_qq_no_interpolation): Don’t use s///r. Do use \K and (?=...)
to match the empty string (but only in just the right locations),
so we don’t need to use $& in the replacement.
(main): Don’t use s///r.
Remove ancient backward compatibility code from m4/m4.m4.
We have had AC_PREREQ([2.69]) in our own configure.ac since 2021,
and before that it was AC_PREREQ([2.62]) since 2012, so backcompat
code for 2.61 is not necessary anymore.
Also add logging of the name of each candidate executable and bump the
serial number.
Unfortunately, while the Time::HiRes *module* was added in 5.8.0, it
did not export a ‘stat’ function until 5.8.9. More precisely, this
feature was added to Time::HiRes in version 1.92 of that module; Perl
core 5.8.8 shipped Time::HiRes 1.86, 5.8.9 shipped 1.9715. The only
system I have convenient access to, that has older Perls installed,
offers me a choice of 5.8.8 or 5.10.1, so the new requirement means
I cannot test with 5.8.x anymore.
Per https://perldoc.perl.org/perlhist the release history of these
versions of Perl is, in chronological order,
5.8.0 2002-Jul-18
5.8.8 2006-Jan-31
5.10.0 2007-Dec-18
5.8.9 2008-Dec-14 <-- almost a year later than 5.10.0
5.10.1 2009-Aug-22
5.12.0 2010-Apr-12
Per https://perldoc.perl.org/perl5101delta the differences between
5.10.0 and 5.10.1 are small, and do not make me worry about
accidentally introducing code that works on my test boxes but not for
our users, unlike the gulf between 5.8.x and 5.10.x.
Requiring 5.10 will mean that we have access to Digest::SHA, the //
and //= operators, the regexp \K escape, and ‘state’ variables
(lexical scope, persistent value) all of which I can think
of uses for (but none of them are actually used in this patch).
Putting it all together, I think a requirement bump to version 5.10.0
is justified. We are already chopping off the trailing edge at 2006
due to the requirement for M4 1.4.8 (/de facto/ since 2.70) and late
2007 is still 15 years ago.
This patch also makes configure search $PATH for executables named
‘perl5*’ and ‘perl-5.*’ if bare ‘perl’ is too old. To do this, it
introduces a helper macro AClocal_PATH_PROG_GLOBS_FEATURE_CHECK, which
I would *like* to promote to a new Autoconf feature, but I got stuck
on quoting issues — see comments in m4/perl-time-hires.m4.
* NEWS: Document requirement for Perl 5.10.
* m4/perl-time-hires.m4: New file.
* configure.ac: Use AC_PATH_PERL_WITH_TIME_HIRES_STAT to probe for
perl.
* build-aux/fetch.pl (fetch): For .pm files fetched from Automake,
rewrite “use 5.006” to “use 5.010”.
The former minimum version was 1.4.6. 1.4.6 and 1.4.7 do not track
the original location of text fed to ‘m4wrap’, which breaks autom4te’s
ability to trace macros invoked from _AC_FINALIZE.
Right now, the only _user_ visible effect of this is that autoconf
running on M4 1.4.6 or 1.4.7 will emit an internal error, instead of
the intended warning message, when it processes a configure.ac that
neglects to invoke AC_INIT or AC_OUTPUT. Perhaps more importantly,
it causes a bunch of scary-sounding failures in our own testsuite,
which deliberately doesn’t use AC_OUTPUT sometimes.
M4 1.4.6 and 1.4.7 also have bugs in location tracking of macro
invocations spread over multiple lines, which, guess what, causes
even more testsuite failures.
1.4.8 came out in 2006. As a practical matter, this change to our
requirements means that people using macOS *to run autoconf* (not just
to run generated configure scripts) cannot use the system-provided m4
anymore. {Free,Net,Open}BSD already don’t ship GNU M4 as their system
m4, so users of those OSes should not be affected.
Paul Eggert [Tue, 28 Jun 2022 22:46:33 +0000 (17:46 -0500)]
Port _AC_COMPILER_EXEEXT to IBM z/OS Unix
Problem reported by Jeff Shimoda in:
https://lists.gnu.org/r/autoconf-patches/2022-06/msg00001.html
* lib/autoconf/lang.m4 (_AC_COMPILER_EXEEXT):
Also remove conftest.o and conftest.obj.
We can't use $ac_objext as it might not be defined yet.
Paul Eggert [Tue, 28 Jun 2022 22:10:53 +0000 (17:10 -0500)]
Refactor recent AC_FUNC_ALLOCA fix
* lib/autoconf/functions.m4 (AC_FUNC_ALLOCA): Simplify and use
AS_IF. Had I done it this way originally I would have avoided the
bug that Jim just fixed.
Jim Meyering [Mon, 27 Jun 2022 05:38:04 +0000 (22:38 -0700)]
AC_FUNC_ALLOCA: fix a misplaced (now fatal) closing "fi"
"autoconf quoting is a pain"
* lib/autoconf/functions.m4 (AC_FUNC_ALLOCA): Its AC_CACHE_CHECK
contains an if/else block, but the closing "fi" lay just after its ")".
Before, this error didn't trigger any failure because the if/else code was
in the "else" block of AC_CACHE_CHECK's AS_IF invocation and AS_IF was also
implemented using an if..fi block. So the ostensibly-"outer" "fi" provided
by AS_IF matched the inner "if/else", and that stray-after-end "fi" served
to close the AS_IF block. However, when AS_IF switched from if..fi to
case..esac, this became a nesting error: no matching "fi".
Initially-harmless error introduced by v2.69-52-gfd29dbd7 in 2012.
Error exposed by v2.72a-30-gc8d6d6eb.
* tests/mktests.pl (scan_m4_files): Do not elide direct test
of AC_FUNC_ALLOCA.
FTR, here's the list of macros whose direct tests were being suppressed:
AC_CANONICAL_BUILD AC_CANONICAL_HOST AC_CANONICAL_TARGET
AC_CHECK_INCLUDES_DEFAULT AC_DISABLE_OPTION_CHECKING
AC_ERLANG_NEED_ERL AC_ERLANG_NEED_ERLC AC_ERLANG_SUBST_ERTS_VER
AC_ERLANG_SUBST_INSTALL_LIB_DIR AC_ERLANG_SUBST_LIB_DIR AC_F77_DUMMY_MAIN
AC_F77_LIBRARY_LDFLAGS AC_FC_DUMMY_MAIN AC_FC_LIBRARY_LDFLAGS
AC_FUNC_ALLOCA AC_FUNC_LSTAT_FOLLOWS_SLASHED_SYMLINK AC_GNU_SOURCE
AC_HEADER_DIRENT AC_HEADER_SYS_WAIT AC_PATH_X AC_PROG_CPP AC_PROG_CXX
AC_PROG_CXXCPP AC_PROG_F77 AC_PROG_FC AC_PROG_GO AC_PROG_GREP
AC_PROG_OBJC AC_PROG_OBJCPP AC_PROG_OBJCXX AC_PROG_OBJCXXCPP
AC_STRUCT_TM AC_TYPE_GETGROUPS AC_TYPE_LONG_LONG_INT AC_TYPE_MBSTATE_T
AC_TYPE_PID_T AC_TYPE_SIZE_T AC_TYPE_UID_T AC_TYPE_UNSIGNED_LONG_LONG_INT
AC_USE_SYSTEM_EXTENSIONS
Jim Meyering [Sun, 26 Jun 2022 22:37:51 +0000 (15:37 -0700)]
tests: avoid test failures due to new EGREP_TRADITIONAL
Some tests require that filtered pre- and post-test
environments be identical. The filtering removes assignments
to known (set-by-test) envvar names, like GREP, SED, YACC.
* tests/local.at (_AT_CHECK_ENV): Also filter out assignments to
EGREP_TRADITIONAL, added in v2.72a-35-g0c762675.
Paul Eggert [Sun, 5 Jun 2022 21:19:21 +0000 (14:19 -0700)]
Fix bug: AC_EGREP_CPP with leading '-' in pattern
Also, port AC_EGREP_CPP to AT&T UnixPC.
Latter problem reported by Alain Knaff [sr #110657].
* lib/autoconf/general.m4 (AC_EGREP_CPP): Use new _AC_PROG_GREP.
Work even if the pattern starts with "-".
* lib/autoconf/programs.m4 (_AC_PROG_GREP):
New optional arg ACTION-IF-NOT-FOUND.
(_AC_PROG_EGREP_TRADITIONAL): New macro.
Paul Eggert [Mon, 16 May 2022 05:14:57 +0000 (22:14 -0700)]
Work around AT&T 3b1 bug
* lib/m4sugar/m4sh.m4: Work around bug in ancient sed.
Problem reported by Alain Knaff [sr#110654] for AT&T 3b1.
The Autoconf manual says this sed bug also occurs for
RISC/OS and IRIX 5.3.
Paul Eggert [Tue, 10 May 2022 21:23:06 +0000 (14:23 -0700)]
Document AC_RUN_IFELSE([], ...) etc.
* doc/autoconf.texi (Running the Preprocessor)
(Running the Compiler, Running the Linker, Runtime):
Document that the _IFELSE macros reuse an existing input file
if their input is empty.
Paul Eggert [Fri, 1 Apr 2022 06:25:35 +0000 (23:25 -0700)]
Work around memory-leak false alarms
Problem reported by Raphael Isemann in:
https://lists.gnu.org/r/autoconf-patches/2022-03/msg00007.html
* lib/autoconf/c.m4 (_AC_C_C99_TEST_MAIN, AC_C_FLEXIBLE_ARRAY_MEMBER):
* tests/data/gnulib_std_gnu11_2020_08_17.m4 (_AC_C_C99_TEST_HEADER):
Free storage allocated via malloc, to pacify overly-picky
implementations that issue false alarms about memory leaks.
Paul Eggert [Fri, 1 Apr 2022 05:29:05 +0000 (22:29 -0700)]
Fix bug with AS_IF and ‘set -e’
Problem reported by Peter Johansson in:
https://lists.gnu.org/r/bug-autoconf/2022-03/msg00009.html
The bug was introduced in 2020-11-15T18:56:18Z!zackw@panix.com
with commit message saying “If anyone has a better idea for how to
make this work I will be glad to hear it”, and this patch uses a
trick with ‘case’ that should be a better idea.
* lib/m4sugar/m4sh.m4 (as_nop, as_fn_nop, _AS_EMPTY_ELSE_PREPARE):
Remove. All uses removed.
(_AS_IF_ELSE): Wrap the non-blank IF-ELSE inside a no-op case,
which works since a case branch can be empty in the shell.
* tests/m4sh.at (AS_IF and AS_CASE): Test the fix.
Paul Eggert [Fri, 18 Feb 2022 23:03:00 +0000 (15:03 -0800)]
Fix unwanted globbing of sed script
Problem reported by mirabilos via Eric Blake in:
https://www.austingroupbugs.net/view.php?id=1558
https://lists.gnu.org/r/bug-autoconf/2022-02/msg00003.html
* lib/m4sugar/m4sh.m4 (_AS_TR_SH_PREPARE):
Define as_sed_sh as the sed script. All uses of as_tr_sh changed.
Define as_tr_sh as before, but only for compatibility with scripts
that mistakenly use as_tr_sh directly.
(AS_TR_CPP_PREPARE): Likewise for as_sed_cpp and as_tr_cpp.
Mike Frysinger [Mon, 22 Feb 2021 16:09:16 +0000 (11:09 -0500)]
add new AC_PROG_AR helper
This mirrors the existing RANLIB and related toolchain macros that
autoconf already exports. Some projects assume `ar` is available
which isn't always safe, so provide a macro that probes the full
toolchain settings.
This also makes it easier to use AC_REQUIRE with the macro instead
of duplicating the AC_CHECK_TOOL call in projects.
* lib/autoconf/programs.m4 (AC_PROG_AR): New macro.
* doc/autoconf.texi: Document it.
* tests/local.at (_AT_CHECK_ENV): Allow $AR output variable.
Zack Weinberg [Thu, 4 Feb 2021 18:11:28 +0000 (13:11 -0500)]
Add a bootstrap script like Automake has.
The bootstrap script generates the same files ‘autoreconf -vi’ would,
in a normal package, but it uses autoconf *from the git sources* to do
it. This means people building from git do not need autoconf to be
installed already. More importantly, it eliminates the extra steps
when building from git, of re-generating autoconf’s own configure
script with the just-built autoconf, then rebuilding the entire tree.
(This process still requires Automake to be installed already, and
Automake’s bootstrap script requires Autoconf to be installed already,
so there is still a dependency loop between Autoconf and Automake when
building from git—you need at least one of them installed from a
tarball to get started.)
The bootstrap script works by creating a partial installation tree in
a temporary directory, containing bin/autoconf, bin/autom4te, and just
enough of the usual contents of $(pkgdatadir) for autoconf and
autom4te to work. It then runs a hardcoded list of commands,
corresponding to what ‘autoreconf -i -Wall,error’ would run, but
setting environment variables AUTOCONF and AUTOM4TE to ensure the
bootstrap versions of these tools are used. (We have to create both,
because automake runs autoconf, not autom4te, to trace configure.ac.)
The ‘Autom4te’, ‘autoconf’, and ‘m4sugar’ subdirectories of the
partial installation tree are symlinked back to the source tree; this
is why version.m4 needed to be moved out of the m4sugar subdirectory,
so the bootstrap script can create it without scribbling on the source
tree. autom4te is run in --melt mode, so we don’t need to create
freeze files in any subdirectories either. All of the substitution
variables that are needed for autoconf and autom4te to both run, and
create the same output that they would have if fully configured, are
honored (unfortunately this does involve digging around in
configure.ac with sed expressions).
Zack Weinberg [Thu, 4 Feb 2021 16:14:37 +0000 (11:14 -0500)]
Rewrite bin/autoconf in Perl.
Of all the installed programs (autoconf, autoheader, autom4te,
autoreconf, autoscan, autoupdate, ifnames) autoconf is the only one
that is a shell script instead of a Perl script. This means it has to
do a lot of fiddly quoting and requoting to assemble an autom4te
command line, it doesn’t get to use the shared option handling code in
Autom4te/{General,Getopt}.pm, and it has to duplicate usage text that
properly should only be in Autom4te/ChannelDefs.pm. It also means
there’s extra code in build-aux/help-extract.pl just for it, and a
special two-phase generation process in bin/local.mk.
This also paves the way for the bootstrap script mentioned in the
previous commit; it will only have to know how to deal with
substitution variables, not generation of m4sh scripts.
The new script winds up being slightly longer on disk but that’s
because our boilerplate for Perl scripts is quite long. The code is
visibly simpler.
* bin/autoconf.as: Rename to bin/autoconf.in and rewrite in Perl.
* bin/local.mk (EXTRA_DIST): Change autoconf.as to autoconf.in.
(MOSTLYCLEANFILES): Don’t delete autoconf.in.
(bin/autoconf.in): Delete rule.
(ETAGS_PERL): Add autoconf.in and sort list.
(ETAGS_SH): Delete, no longer needed.
* tests/tools.at: Syntax-check autoconf as a Perl script.
* build-aux/help-extract.pl: Remove all code for extracting usage text
from shell scripts.
* man/autoconf.w, man/local.mk: Refer to autoconf.in, not autoconf.as.
Zack Weinberg [Thu, 4 Feb 2021 15:48:36 +0000 (10:48 -0500)]
Generate version.m4 from config.status, in lib/ not lib/m4sugar/.
This simplifies the Makefile a bit and also paves the way for a
bootstrap script (replacing the process of re-running autoreconf with
the just-built autoconf) which will be able to cope with generated
files that get installed in @pkgdatadir@ but not in @pkgdatadir@/m4sugar.
* lib/version.in: New file, contents extracted from lib/local.mk.
* lib/local.mk: Do not generate version.m4 here.
* configure.ac: Generate lib/version.m4 from lib/version.in via
config.status.
* lib/freeze.mk, lib/local.mk, lib/m4sugar/m4sugar.m4:
version.m4 now lives in @pkgdatadir@ rather than @pkgdatadir@/m4sugar.
* lib/autoconf/fortran.m4 (_AC_FC_LIBRARY_LDFLAGS):
Filter out mingw libraries too.
Problem and fix reported by Keno Fischer in
https://lists.gnu.org/r/autoconf-patches/2021-09/msg00000.html
Copyright-paperwork-exempt: Yes
Paul Eggert [Tue, 31 Aug 2021 20:08:10 +0000 (13:08 -0700)]
Modernize discussion of integer overflow
* doc/autoconf.texi (Integer Overflow, Integer Overflow Basics)
(Signed Overflow Examples, Optimization and Wraparound):
Modernize discussion to take current compiler and Gnulib
technology into account.
Marshall Ward [Thu, 12 Aug 2021 00:43:36 +0000 (20:43 -0400)]
Fortran Cray pointer macro support
This patch adds the AC_F77_CRAY_POINTERS and AC_FC_CRAY_POINTERS macros,
which test if the Fortran compiler supports Cray pointers.
The macros are written such that the tests share a common backend
(_AC_FC_CRAY_POINTERS) which works on both F77 and FC compilers.
Wrappers are provided to address any future potential compatibility
issues.
The macros include additional tests for particular flags required by
GFortran and PGI compilers. The current set of flags is sparse, but can
be extended for other compilers if needed.
Documentation and a minimal test of the macro have been included.
Two minor variable name typos (@EXEEXT@ as @EEXEXT@) were also fixed in
two of the other Fortran tests.
Paul Eggert [Tue, 20 Jul 2021 21:01:43 +0000 (16:01 -0500)]
Autoconf now quotes 'like this' instead of `like this'
Autoconf’s diagnostics now follow current GNU coding standards,
which say that diagnostics in the C locale should quote 'like this'
with plain apostrophes instead of the older GNU style `like this'
with grave accent and apostrophe.
Paul Eggert [Mon, 19 Jul 2021 04:54:19 +0000 (23:54 -0500)]
Document unquoted here-documents better
Problem reported by Bruno Haible (sr #110518).
* doc/autoconf.texi (Generating Sources)
(Running the Preprocessor, Running the Compiler)
(Running the Linker, Runtime): Briefly mention that the program
snippets are expanded as unquoted here-documents.
(Here-Documents): Document escaping needed in unquoted
here-documents. Remove mention of \" glitch in OpenBSD 2.7 sh;
that old system has not been supported for many years.
Remove advice that every \ must be quoted, as that sort
of quoting (though harmless) is not needed and is rarely done.
* lib/autoconf/c.m4 (AC_C_BIGENDIAN, _AC_LANG_OPENMP(Fortran 77)):
* lib/autoconf/types.m4 (AC_TYPE_LONG_LONG_INT): Properly quote
code snippets. The backslash-newline typos were harmless, and the
backslash-$ typo has unspecified behavior as per POSIX though the
typo is harmless on all shells I know of. However, Autoconf
should follow its own quoting advice.
_AC_PROG_CXX_STDCXX_EDITION_TRY: fix typo in variable name
The typo causes "ac_prog_cxx_stdcxx" to be always set to "cxx11",
regardless of whether the C++ compiler supports C++11.
Message-Id: <YJ4TXAeJcvU0oSec@FSAPPLE2215.fi.f-secure.com>
(tiny change)
Eric Blake [Mon, 10 May 2021 20:50:42 +0000 (15:50 -0500)]
maint: advertise GNU in README
* README: Add section to attract more people towards the GNU project.
Inspired by a suggestion from Jose E. Marchesi <jemarch@gnu.org> on
the gnu-prog-discuss mailing list.
Ben Elliston [Wed, 25 Nov 2015 10:08:19 +0000 (21:08 +1100)]
update HACKING file
Fairly obvious - be gender neutral in who gets thanked.
* HACKING: Tweak pronouns.
Message-Id: <20151125100819.GA29844@air.net.au> Signed-off-by: Eric Blake <eblake@redhat.com>
Zack Weinberg [Mon, 4 Jan 2021 22:27:37 +0000 (17:27 -0500)]
autoreconf: don’t error out when AM_GNU_GETTEXT is used alone.
Some projects intentionally don't call AM_GNU_GETTEXT_(REQUIRE_)VERSION
because they have all of the gettext infrastructure checked into version
control and they want autoreconf to _not_ run autopoint. Therefore, make
the “AM_GNU_GETTEXT is used, but not AM_GNU_GETTEXT_(REQUIRE_)VERSION”
and “AM_GNU_GETTEXT_(REQUIRE_)VERSION is used, but not AM_GNU_GETTEXT”
diagnostics be warnings again, as they were in 2.69.
(Technically these diagnostics were always errors, in the sense that
they were reported with a call to Autom4te::Channels::error(), but
prior to 2.70, error() calls made by autoreconf did not cause
autoreconf to exit unsuccessfully, due to an unrelated bug. So people
came to depend on these diagnostics not being fatal.)
Zack Weinberg [Tue, 29 Dec 2020 20:33:33 +0000 (15:33 -0500)]
autom4te: always update the output file, even if it hasn’t changed
Automake generates a Makefile rule for regenerating the configure
script, that relies on an invocation of ‘autoconf’ always bumping the
timestamp on the configure script, even if it hasn’t changed.
The patch to make autom4te update the output file atomically
(1725c947144d9bebfe7817c2c5f0d53d884b1297) broke this.
Fixes several failures in automake’s test suite.
* bin/autom4te.in (handle_output): Always call update_file with force=1.
* tests/tools.at (autoconf: timestamp changes): New test.
Zack Weinberg [Mon, 21 Dec 2020 19:04:22 +0000 (14:04 -0500)]
Restore compatibility with older std-gnu11.m4.
Gnulib’s std-gnu11.m4 backports C11 and C++11 detection to autoconf
2.69. It does this by replacing the definitions of AC_PROC_CC and
AC_PROG_CXX and most of their subroutines. In particular, it replaces
the definitions of _AC_PROG_CC_C11, _AC_PROG_CC_C99, and _AC_C_STD_TRY,
but it does *not* replace the definition of _AC_PROG_CC_C89.
Autoconf commit 131d8c69f31dc6fc8dc93abe1096d52d1fe19fd3 changed the
calling convention of _AC_C_STD_TRY, and changed the internal
definitions of _AC_PROG_CC_C{11,99,89} to match. If std-gnu11.m4 is
in use, our _AC_PROG_CC_C89 calls their _AC_C_STD_TRY with the new
calling convention, and this produces a syntactically invalid
configure script. (This is is fortunate: it could easily have been a
runtime malfunction that only manifested with compilers that only
implement C89, and then we might not have noticed the problem for
years.)
Gnulib commit a3b3fc85e3e632374811b27cb2111e50fa177e36 makes
std-gnu11.m4 do nothing when used with autoconf >=2.70, but older
versions of the file will circulate for years to come, so this patch
works around the problem in autoconf. It does this by renaming all of
the internal macros involved with C and C++ standard edition
detection, *except* _AC_PROG_CC_C89. AC_PROG_CC now calls
_AC_PROG_CC_STDC_EDITION, which loops over all supported editions
calling _AC_PROG_CC_STDC_EDITION_TRY, which uses the data provided by
the existing _AC_C_C${edition}_TEST_PROGRAM macros and a new set of
macros called _AC_C_C${edition}_OPTIONS to perform the test for that
edition of the standard. Similarly, AC_PROG_CXX calls
_AC_PROG_CXX_STDCXX_EDITION, which loops calling
_AC_PROG_CXX_STDCXX_EDITION_TRY, which uses data from
_AC_CXX_CXX${edition}_TEST_PROGRAM and _AC_CXX_CXX${edition}_OPTIONS.
_AC_PROG_CC_C89 is the only macro from the old set that we still
define, and its definition is reverted to what std-gnu11.m4 expects it
to be. Nothing in Autoconf proper uses it anymore.
foreign.at grows a test to verify that the compatibility stub version
of _AC_PROG_CC_C89 does its job. Since this is now the third test
involving an embedded copy of a third-party macro, I broke them all
out of foreign.at to separate files in test/data/.
In addition to fixing the breakage, this patch should make it easier
to extend C / C++ standard edition detection in the future, by getting
rid of the if-else chains in AC_PROG_CC/CXX and by disentangling the
lists of command-line options to test from the logic.
I also changed the manual to suggest people refer to the variables
‘ac_prog_cc_stdc’ and ‘ac_prog_cxx_stdcxx’ to learn which edition
of the C and C++ standards are selected; these are much easier to
work with than the ac_cv_prog_cc_cNN cache variables.
(_AC_PROG_CC_C89): Convert to compatibility stub for std-gnu11.m4.
(AC_PROG_CC): Use _AC_PROG_CC_STDC_EDITION.
(AC_PROG_CXX): Use _AC_PROG_CXX_STDCXX_EDITION.
* tests/data/ax_prog_cc_for_build_v18.m4
* tests/data/ax_prog_cxx_for_build_v3.m4
* tests/data/gnulib_std_gnu11_2020_08_17.m4: New files.
* tests/foreign.at (AX_PROG_CC_FOR_BUILD, AX_PROG_CXX_FOR_BUILD):
Remove embedded copy of ax_prog_cc_for_build_v18.m4,
ax_prog_cxx_for_build_v3.m4 respectively.
(gnulib-std-gnu11.m4): New test.
* tests/local.mk: Distribute tests/data/*.m4.
* doc/autoconf.texi (AC_PROG_CC, AC_PROG_CXX): Document use of
ac_prog_cc_stdc / ac_prog_cxx_stdcxx, respectively, to tell which
edition of the C / C++ standards are selected, instead of looking
through a series of cache variables with awkward definitions.
Zack Weinberg [Wed, 23 Dec 2020 17:14:06 +0000 (12:14 -0500)]
Use -fno-builtin, not -Werror, in AC_CHECK_DECLS (#110400)
clang issues only a warning, not an error, when an undeclared
identifier that names a built-in function is used: for instance
char *(*p)(const char *, int) = strchr;
(with no `#include <string.h>`) is an error with most compilers,
a warning with clang. This broke the 2.69 implementation of
AC_CHECK_DECL. In commit 82ef7805faffa151e724aa76c245ec590d174580,
we tried to work around this quirk by using -Werror, but that put us
at risk of being tripped up by other warnings. Bug 110400 reports,
for instance, that this fragment (which is roughly what you get, after
preprocessing, when AC_CHECK_DECL is applied to a function that *is*
properly declared)
extern void ac_decl (int, char *);
int main (void)
{
(void) ac_decl;
;
return 0;
}
provokes a warning from clang (and thus an error) when -Wextra-semi-stmt
has been added to CFLAGS earlier in the configure script. The extra
semicolon comes from AC_LANG_PROGRAM, and we can’t get rid of it
because we have no way of telling reliably when someone wrote
something like
with no semicolon at the end of the statement; this has been
acceptable for decades. Besides, that’s just one warning, who knows
what compilers will start complaining about tomorrow?
So: change AC_CHECK_DECL to compile its programs with -fno-builtin,
instead, when the default compilation mode fails to detect an
undeclared strchr. The code is restructured so that we can try other
options as well, if we find another compiler with the same quirk but
different command-line syntax.
(All of this logic is very C-family specific, but it appears to me
that AC_CHECK_DECL has never worked with other languages, so we can
continue to live with that for now.)
* lib/autoconf/general.m4 (_AC_UNDECLARED_WARNING): Rename to
_AC_UNDECLARED_BUILTIN. Instead of looking at diagnostic output,
loop trying to find a command-line option that makes the compiler
error out on undeclared builtins.
(_AC_CHECK_DECL_BODY): Don’t AC_REQUIRE anything here.
Make shell code language-agnostic, except for the actual test program.
Add arguments to the shell function for additional compiler options
to use.
(AC_CHECK_DECL): AC_REQUIRE _AC_UNDECLARED_BUILTIN here.
Supply $ac_{AC_LANG_ABBREV}_undeclared_builtin_options to
ac_fn_check_decl.
* tests/local.at (AT_CONFIG_CMP): Update list of variables to ignore
when comparing C and C++ configure runs.
* tests/semantics.at (AC_CHECK_DECLS): Add memcpy and strchr to
AC_CHECK_DECLS call for functions that may be known to the compiler.
* doc/autoconf.texi (AC_CHECK_DECL, AC_CHECK_DECLS): Remove note
about compiler warnings.
Zack Weinberg [Mon, 21 Dec 2020 20:29:32 +0000 (15:29 -0500)]
autom4te: correct error message when we can’t create autom4te.cache.
While testing something else, I noticed that autom4te may print a
nonsensical error message when it fails to create autom4te.cache,
because it checks again whether the directory already exists before
giving up, and this clobbers errno.
Instead of doing (the perl equivalent of)
test -d $cache || mkdir $cache || test -d $cache
call mkdir unconditionally. If it fails with an errno code other than
EEXIST, consider that a hard failure; if it fails with EEXIST, check
whether the thing that exists is in fact a directory. (A symlink to
a directory qualifies; I wouldn’t be surprised if people are moving
autom4te.cache around with symlinks.)
Either way, if we fail, report strerror(errno) from the original
mkdir failure. Also, print the current working directory as part
of the error message; this aids debugging when you’re working with a
big hairy nested tree.
* bin/autom4te.in: Don’t check whether autom4te.cache exists before
attempting to create it. Only stat autom4te.cache if mkdir fails
with EEXIST, otherwise fail immediately. Make sure to report the
errno code from mkdir, not the subsequent stat (if any). Report
the current working directory as part of the error message.
* tests/tools.at: Verify that autom4te reports the actual reason when
it fails to create autom4te.cache. Verify that failure to create
autom4te.cache because that name exists, but isn’t a directory,
is detected.
projects / thirdparty / autoconf.git / log
