Re: AC_CHECK_DECL has wrong results on macOS

Two questions about this patch installed a couple days ago . Should Gnulib's gl_CHECK_FUNCS_MACOS be adjusted to work better with future Autoconf versions that have the patch? Is there a s

Re: fix quoting in trap command

Thanks for that quote cleanup. I was struck by its simplification of "s/'\''/'\'''\'''\''/g" to "s/'/'''/g"; only cognoscenti can appreciate minor victories like these. I installed it.

Re: AC_CHECK_DECL has wrong results on macOS

Thanks for checking the revised patch, and I installed it with the additional line "Co-authored-by: Paul Eggert ".

Re: AC_CHECK_DECL has wrong results on macOS

On 2025-05-24 04:37, Bruno Haible via Bug reports for autoconf wrote: Here's a proposed fix for the issue along these lines. Thanks. It'd be nicer if AC_CHECK_DECL didn't require AC_CANONICAL_HOST. We can do that by checking inside the C program rather than using AC_CANONICAL_HOST. Also, sh

Re: \{n\} are not recognized as repetition counter in regular expressions.

On 2025-05-22 10:42, Eric Blake wrote: BSD has tried hard to make their m4 be a drop-in replacement enough that autoconf will use it instead of mandating GNU m4 Is this a real problem, though? I just now tried running Autoconf's 'configure' on FreeBSD 15, and FreeBSD 15's /usr/bin/m4 quickly f

[sr #111243] stdbool conftest program (and possibly others) is broken with GCC 15

Follow-up Comment #5, sr #111243 (group autoconf): [comment #4 comment #4:] >> I cannot reproduce the problem with stdbool.h using Autoconf 2.72 (as >> shipped with Fedora 42), when building cifs-tools commit >> 5ba5b16a73e70b586355b591f9feefbe3c493bc2 (aka version 7.3). > > Hm, that's interestin

[sr #111243] stdbool conftest program (and possibly others) is broken with GCC 15

Update of sr #111243 (group autoconf): Severity: 4 - Important => 3 - Normal Status:None => Works For Me Open/Closed:Open => Closed ___ Follow-

Re: fedora 42 doesn't have awk: how to deal with autoconf subst?

, configure with AWK='busybox awk', etc., etc.From fa76d00401519123a782af1963b7d7ba677347b3 Mon Sep 17 00:00:00 2001 From: Paul Eggert Date: Sun, 20 Apr 2025 10:36:19 -0700 Subject: [PATCH] Fall back on busybox awk * lib/autoconf/programs.m4 (AC_PROG_AWK): Fall back 'bu

Re: fedora 42 doesn't have awk: how to deal with autoconf subst?

-n "\$AWK" || { + awk '' >"$CONFIG_STATUS" <<\_ACEOF || ac_write_fail=1 -- 2.48.1 From 3d79cd9ff455d85e788398bfe4fdc4544fabceb5 Mon Sep 17 00:00:00 2001 From: Paul Eggert Date: Fri, 18 Apr 2025 14:01:38 -0700 Subject: [PATCH 2/2] Avoid grep in mini

Re: fedora 42 doesn't have awk: how to deal with autoconf subst?

g with 'cmp', 'gcc', 'make', and 'touch'.From 95b849dd837210dc3109adfbd52a0929fbec141e Mon Sep 17 00:00:00 2001 From: Paul Eggert Date: Thu, 17 Apr 2025 12:14:41 -0700 Subject: [PATCH] Improve list of 'configure' prereqs * doc/install.texi (Installation Pre

Re: fedora 42 doesn't have awk: how to deal with autoconf subst?

On 2025-04-17 04:31, Simon Josefsson wrote: Paul Eggert writes: [ awk cat cmp cp diff echo expr false grep ls mkdir mv printf pwd rm rmdir sed sort test touch tr true Of those, I'm guessing awk, grep, sed and ls are the most complex. The 'diff' tool also stands out, and most

Re: fedora 42 doesn't have awk: how to deal with autoconf subst?

. Maybe Autoconf should be changed so that 'configure' fails quickly if there's no awk. That shouldn't be much work.From 56860215d1af490eddb821d73154a190137aa8ec Mon Sep 17 00:00:00 2001 From: Paul Eggert Date: Thu, 17 Apr 2025 01:08:52 -0700 Subject: [PATCH] Document t

[sr #111221] _AC_DEFINE_UNQUOTED has broken regex

Update of sr #111221 (group autoconf): Status:None => Done Open/Closed:Open => Closed ___ Follow-up Comment #2: Thanks for reporting that. I installed [https://git.sava

[sr #111172] AC_FC/F77_LIBRARY_LDFLAGS incorrectly picks up flags from 'llc'

Update of sr #72 (group autoconf): Status:None => Done Open/Closed:Open => Closed ___ Follow-up Comment #2: I installed [https://git.savannah.gnu.org/cgit/autoconf.

Re: [GNU Autoconf 2.72] testsuite: 19 failed

On 2025-01-01 15:03, Sean Denney wrote: The patch failed to apply to the 2.72 release directory (even with a clean install from the tarball). It failed at doc/autoconf.texi:18532. Yes, master has evolved too much for that to work without human assistance in installing the patch. I just did

Re: [GNU Autoconf 2.72] testsuite: 19 failed

On 2024-12-31 14:43, Sean Denney wrote: The output from echo $? In that case was zero (0). Thanks for checking. I installed the attached patch to Autoconf master on Savannah; please give it a try.From 1a46ea891587e697be27e2d4d8e3eb47d9615321 Mon Sep 17 00:00:00 2001 From: Paul Eggert Date

Re: [GNU Autoconf 2.72] testsuite: 19 failed

On 12/31/24 05:00, Sean Denney wrote: That returns 1 as well. Timestamps are: sh-3.2$ stat -f %Fm B A 1735649785.553643117 1735649797.635429442 touch B sleep 2 touch A test B -nt A echo $? Thanks, and let's try the opposite direction, to make sure the diagnosis is correct. Please try this:

Re: [GNU Autoconf 2.72] testsuite: 19 failed

On 12/31/24 04:49, Sean Denney wrote: When I set the shell to sh, then echo $? returns 1 after running test -nt OK, I think we've found the problem: plain sh doesn't fully support test -nt. It'd help to see how bad the problem is. If you run sh, does 'test A -nt B' work when A is much newer

Re: [GNU Autoconf 2.72] testsuite: 19 failed

On 12/31/24 04:22, Sean Denney wrote: if test ${ZSH_VERSION+y} && (emulate sh) >/dev/null 2>&1 then : emulate sh Those lines say that if testsuite is being run by zsh, then zsh should execute 'emulate sh'. But is testsuite actually being run by zsh? What is the first line of the testsuite

Re: [GNU Autoconf 2.72] testsuite: 19 failed

On 12/31/24 03:38, Paul Eggert wrote: Also the echo output was zero (0) Oh, but the log in your first message said "test configure -nt configure.1" had exit status 1. Possibly you're using a different shell than Autoconf uses to run its test script, and your shell supports t

Re: [GNU Autoconf 2.72] testsuite: 19 failed

On 12/31/24 03:27, Sean Denney wrote: As for the timestamps, ls doesn’t have ‘—full-time’ but does have ‘-T’ which I hope is what you are after I was after the subseconds part of the file timestamp, which ls -T doesn't provide. Although I don't use macOS, as I understand it apfs is supposed

Re: [GNU Autoconf 2.72] testsuite: 19 failed

Thanks for the bug report. What kind of filesystem was the test run on? Can you investigate the timestamps of the failed test? E.g., something like the following? cd tests/testsuite.dir/019 test configure -nt configure.1 echo $? ls -l --full-time configure configure.1

[sr #111140] autoscan gets confused by Makefile rules that start with '-' to ignore errors

Update of sr #40 (group autoconf): Priority: 5 - Unprioritized => 2 - Eventually Severity: 3 - Normal => 1 - Wish ___ Follow-up Comment #1: Unfortunately the su

Re: Not portable echo with backslash character in AC_SUBST_FILE

Thanks for reporting that. I installed the attached patch.From 53c23bb4e1cd8aa35248c73277fa08b7c38a6ad9 Mon Sep 17 00:00:00 2001 From: Paul Eggert Date: Mon, 21 Oct 2024 10:49:30 -0700 Subject: [PATCH] Avoid some echo '...\...' gotchas Problem with AC_SUBST_FILE reported by Dylan P

[sr #111111] AC_SYS_LARGEFILE system wide control and autoconf releases

Update of sr #11 (group autoconf): Status:None => Done ___ Reply to this item at: ___ M

[sr #111111] AC_SYS_LARGEFILE system wide control and autoconf releases

Update of sr #11 (group autoconf): Open/Closed:Open => Closed ___ Reply to this item at: ___ M

[sr #110286] Make it possible to request a specific (non-latest) version of a language standard

Follow-up Comment #5, sr #110286 (group autoconf): As far as I know the original bug (#110879, which led to the Ixion issue for example) has been fixed in bleeding-edge Autoconf on Savannah. However, nobody has yet volunteered to implement the feature request "Make it possible to request a specif

[sr #111111] AC_SYS_LARGEFILE system wide control and autoconf releases

Follow-up Comment #3, sr #11 (group autoconf): [comment #2:] > ac_cv_sys_largefile_opts='none needed' > ac_cv_sys_file_offset_bits=no That should work. I'd go with that rather than anything more complicated. Still not sure it's worth documenting, but there are no plans to change it. ___

[sr #111111] AC_SYS_LARGEFILE system wide control and autoconf releases

Follow-up Comment #1, sr #11 (group autoconf): The cache variables for AC_SYS_LARGEFILE aren't documented, so they aren't part of the stable API. In Autoconf 2.71 they were ac_cv_sys_largefile_CC, ac_cv_sys_file_offset_bits, ac_cv_sys_large_files. In bleeding-edge Autoconf things are simpler:

[sr #111102] ksh93 oddity w.r.t. ${1+"$@"} on AIX (for section "Portable Shell"?)

Follow-up Comment #3, sr #02 (group autoconf): Thanks, I fixed the doc typos you flagged. I didn't find more-general flaws, though I was testing AIX 7.1 (also ksh93 Version M 93t+ 2009-05-01). ___ Reply to this item at:

[sr #111102] ksh93 oddity w.r.t. ${1+"$@"} on AIX (for section "Portable Shell"?)

Update of sr #02 (group autoconf): Status:None => Done Open/Closed:Open => Closed ___ Follow-up Comment #1: Thanks for reporting

Re: make AC_PROG_OBJC work out-of-the-box on RHEL 9 and compatible systems

Thanks, I installed that.

[sr #111072] "[[]] attributes are a C23 extension warning" with clang

Update of sr #111072 (group autoconf): Status:None => Invalid Open/Closed:Open => Closed ___ Follow-up Comment #1: The problems you men

Re: Improve Autoconf MSVC support

Thanks for your helpful comments. Would you be up to writing and testing Autoconf patches to support MSVC better? I don't use MSVC and so would not be good for that. Thanks.

[sr #111054] Typo in texi file breaks build

Update of sr #111054 (group autoconf): Status:None => Done Open/Closed:Open => Closed ___ Follow-up Comment #1: Thanks for reporting

Re: ac_cv_sys_largefile_opts undocumented?

On 2024-04-27 15:39, dmitrii.pasech...@cs.ox.ac.uk wrote: It's for nauty, a well-known program to deal with graph isomorphisms etc. I've made a Gentoo patch herehttps://bugs.gentoo.org/921138 and I wanted to upstream it. Oh my, that is indeed in an undocumented/hacky part of autoconf, one that

Re: ac_cv_sys_largefile_opts undocumented?

On 2024-04-27 14:06, dmitrii.pasech...@cs.ox.ac.uk wrote: Thus I got questions whether a patch for a build system I submitted for a project is OK, as it uses an undocumented variable (thus, perhaps, unstable). Although it's poorly designed and not documented and not stable, the patch may be go

[sr #110983] AC_SYS_LARGEFILE doesn't honor AC_LANG setting

Update of sr #110983 (group autoconf): Status: Confirmed => Done ___ Follow-up Comment #6: I installed [https://git.savannah.gnu.org/cgit/autoconf.git/commit/?id=b71143738516017f0e0d347a40

Re: autoreconf --force seemingly does not forcibly update everything

On 4/10/24 13:36, Simon Josefsson via Gnulib discussion list wrote: Is bootstrap intended to be reliable from within a tarball? I thought the bootstrap script was not included in tarballs because it wasn't designed to be ran that way, and the way it is designed may not give expected results. I

Re: Suggestion for improving an error message

undefined or overquoted macro". Since the diagnostic can be caused by misspelled macro names, which do occur, I thought it wise to retain a mention of that in the diagnostic. I installed the attached patch.From 7a6347d1d785ee26f205154fdadf7f6f81797f92 Mon Sep 17 00:00:00 2001 From: Paul Eggert D

[sr #111044] autoconf should assert existence of all subsidiary tools at startup

Follow-up Comment #3, sr #111044 (group autoconf): [comment #2 comment #2:] > neither `diff` nor `awk` (and arguably not even `sed`) should be an implicit dependency. Not sure I'd go that far. The [https://www.gnu.org/prep/standards/html_node/Utilities-in-Makefiles.html GNU Coding Standards for u

Re: `cp -l` doesn't work correctly on some macOS versions

ble, as there are so many (and they proliferate), the manual could be better about what *is* portable and I installed the attached. From 163dade069be64df7cce5c6d48fdcb56188a6f60 Mon Sep 17 00:00:00 2001 From: Paul Eggert Date: Sat, 3 Feb 2024 21:37:49 -0800 Subject: [PATCH] =?UTF-8?q?Document=2

Re: `cp -l` doesn't work correctly on some macOS versions

On 2/2/24 12:20, Russ Allbery wrote: I thought ln || ln -s was the standard recipe here Yes, "cp -l" isn't required by POSIX and is not portable in general. For example, on AIX 7.1: $ cp -l a b cp: Not a recognized flag: l Usage: cp [-fhipHILPU][-d|-e] [-r|-R] [-E{force|ignore|warn

Re: cannot make gcc report undeclared builtins

On 1/31/24 03:23, Simon Josefsson via Bug reports for autoconf wrote: https://buildd.debian.org/status/fetch.php?pkg=libidn2&arch=loong64&ver=2.3.7-1&stamp=1706360630&raw=0 Here's the crucial part of that log: configure:15961: checking for size_t configure:15961: gcc -c -g -O2 -ffile-prefix

Re: AC_TYPE_LONG_LONG_INT busted in current Autoconf

3a813b3b2f9f3a8ec0df95e444b6fb54267775a9 Mon Sep 17 00:00:00 2001 From: Paul Eggert Date: Tue, 16 Jan 2024 21:19:31 -0800 Subject: [PATCH] Fix typo in previous patch --- lib/autoconf/types.m4 | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/lib/autoconf/types.m4 b/lib/autoconf

Re: AC_TYPE_LONG_LONG_INT busted in current Autoconf

Thanks for reporting that bug. I installed the attached on Savannah; please give it a try.From e5fae45ae39f471cba5415c21d93e5bc407eb8c7 Mon Sep 17 00:00:00 2001 From: Paul Eggert Date: Tue, 16 Jan 2024 20:20:06 -0800 Subject: [PATCH] =?UTF-8?q?Fix=20=E2=80=98long=20long=E2=80=99=20checks=20

Re: Autoconf documentation link broken

On 1/10/24 14:03, Bruno Haible wrote: When visiting https://www.gnu.org/software/autoconf/manual/ the first hyperlink (_HTML (2220K bytes)_) works, but the second hyperlink (_HTML_) lands in a 404 error page. Thanks, it should be working now. I don't know why it wasn't working earlier. The CV

Re: [GNU Autoconf 2.72e] testsuite: 11 119 261 failed on Solaris 11.4 x86

On 2024-01-02 10:40, Alan Coopersmith wrote: On 12/21/23 13:46, Alan Coopersmith wrote: On Solaris 11.4, /bin/sh is currently a somewhat modified ksh93u+ - we're working to resync with the new community upstream still. In that shell, when I run:    printf "test" /dev/full it prints nothing and

Re: [GNU Autoconf 2.72e] testsuite: 11 119 261 failed on Solaris 11.4 x86

On 2023-12-21 19:56, Nick Bowler wrote: On 2023-12-21 19:34, Paul Eggert wrote: ulimit -f 0 trap "" XFSZ printf "test" >test || echo failed with status $? which issues the following diagnostics on Solaris 10 /bin/sh: printf: write error: File too large

Re: [GNU Autoconf 2.72e] testsuite: 11 119 261 failed on Solaris 11.4 x86

On 2023-12-21 13:46, Alan Coopersmith wrote: On Solaris 11.4, /bin/sh is currently a somewhat modified ksh93u+ - we're working to resync with the new community upstream still. In that shell, when I run:   printf "test" /dev/full it prints nothing and $? is set to 0. This appears to be a regre

Re: [GNU Autoconf 2.72e] testsuite: 11 119 261 failed on Solaris 11.4 x86

On 2023-12-21 13:19, Zack Weinberg wrote: Sorry, I'm with GNU here: failure to report errors on writing to stdout is a bug. No excuses will be accepted. Agreed. printf commands that silently succeed when they can't do the requested action are simply broken. Even if one is not convinced by

Re: m4_foreach_w(...AC_DEFINE...) expansion change in 2.72e?

I'm seeing the same bug with Autoconf 2.69 on RHEL 7. Tarball attached with the configure.ac that I used, and the output of "autoheader" and "autoconf". "autoconf" complains: configure.ac:5: warning: The macro `AC_FOREACH' is obsolete. configure.ac:5: You should run autoupdate. ./lib/autoconf/g

[sr #110983] AC_SYS_LARGEFILE doesn't honor AC_LANG setting

Follow-up Comment #2, sr#110983 (group autoconf): Do you have an idea for a patch? We could put the patch into Gnulib's copy of AC_SYS_LARGEFILE now, and it'll see more real-worldish testing that way. ___ Reply to this item at:

[sr #110612] Re-exec of $as_myself chooses wrong configure script from PATH

Follow-up Comment #3, sr#110612 (group autoconf): [comment #1 comment #1:] > If you or anyone else has ideas for how to tell "configure" from "sh configure", I am all ears. If it's Bash, "$_" (at the start of the script) should tell you. This is true for other shells (e.g., it works with Solari

[sr #110971] busybox mkdir is misdetected

Follow-up Comment #0, sr#110971 (group autoconf): [comment #4 comment #4:] > It seems to me it will be more robust to filter out the troublesome error message Not really, as BusyBox plausibly might change that diagnostic's wording and then the filtering will stop working. Conversely, if the outpu

[sr #110971] busybox mkdir is misdetected

Follow-up Comment #3, sr #110971 (project autoconf): Wouldn't this patch be simpler? diff --git a/lib/autoconf/programs.m4 b/lib/autoconf/programs.m4 index d06d18c4..8226a7c7 100644 --- a/lib/autoconf/programs.m4 +++ b/lib/autoconf/programs.m4 @@ -700,7 +700,7 @@ if test -z "$MKDIR_P"; then

Re: autoconf-2.72d on CentOS 8: testsuite: 11 failed

lled the attached patch, which fixes that bug. However, I don't know whether that will address the problem you observed.From 49ab3a4c5756d7ef40b03111f4c8a351d6bca7b8 Mon Sep 17 00:00:00 2001 From: Paul Eggert Date: Fri, 1 Dec 2023 09:58:37 -0800 Subject: [PATCH] Be more conservative about cache tim

[sr #110961] 3 Vulnerabilities Result in Code Execution upon running `autoconf` with crafted `configure.ac` file

Update of sr #110961 (project autoconf): Status:None => Done Open/Closed:Open => Closed ___ Follow-up Comment #1: Thanks for the bug

Re: [PATCH] Quote 'like this', not `like this'.

ep 17 00:00:00 2001 From: Paul Eggert Date: Sat, 24 Jun 2023 15:00:37 -0700 Subject: [PATCH] =?UTF-8?q?*=20build-aux/gitlog-to-changelog:=20don?= =?UTF-8?q?=E2=80=99t=20quote=20`like=20this'?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit ---

Re: AC_FUNC_MMAP lacks include?

On 2023-05-17 12:25, Florian Weimer wrote: But then HAVE_GETPAGESIZE probably should not succeed spuriously because of an implicit function declaration. I hope that the only platforms that should use that old getpagesize code, are platforms so old that implicit function declarations work, wh

Re: AC_FUNC_MMAP lacks include?

fe6e9a5afbb7b2ae13ae0 Mon Sep 17 00:00:00 2001 From: Paul Eggert Date: Wed, 17 May 2023 11:50:27 -0700 Subject: [PATCH] Improve AC_FUNC_MMAP comments * lib/autoconf/functions.m4 (AC_FUNC_MMAP): Add comment. --- lib/autoconf/functions.m4 | 2 ++ 1 file changed, 2 insertions(+) diff --git a/lib/a

Re: AC_FUNC_MMAP lacks include?

On 5/11/23 07:08, Zack Weinberg wrote: Unless someone knows of a bug in some old kernel where the first*disk block* of an oversized mmap would be properly zeroed and then the rest of it would be memory garbage Yes, that was the reason I went back to configuring a getpagesize substitute: I wan

Re: AC_FUNC_MMAP lacks include?

suppose we should go back to invoking getpagesize despite the hassle of configuring it. I installed the attached further patch; please give it a try.From 33c26d2700f927432c756ccf7a4fc89403d35b95 Mon Sep 17 00:00:00 2001 From: Paul Eggert Date: Wed, 10 May 2023 22:57:27 -0700 Subject: [PATCH] F

Re: AC_FUNC_MMAP lacks include?

s the test so that it doesn't use getpagesize. This is a conservative-enough approach that it should be OK to squeeze it in before the next Autoconf release.From 028526149ee804617a302ccef22cc6adbda681b0 Mon Sep 17 00:00:00 2001 From: Paul Eggert Date: Wed, 10 May 2023 17:20:49 -0700 Subje

Re: two recent NEWS entries

Thanks for mentioning that. I installed the attached to try to clarify things.From 2a43cf6ab085708f900aeb7ca1c19a17b2992ece Mon Sep 17 00:00:00 2001 From: Paul Eggert Date: Fri, 3 Feb 2023 12:25:08 -0800 Subject: [PATCH] Clarify NEWS * NEWS: Clarify recent items about AC_SYS_LARGEFILE_REQUIRED

[sr #110828] configure: error: could not determine the string function for int64_t

Update of sr #110828 (project autoconf): Status:None => Invalid Open/Closed:Open => Closed ___ Follow-up Comment #2: This appears to be

[sr #110713] Use __STDC__VERSION as first test for C99 compatibility (patch attached)

Follow-up Comment #2, sr #110713 (project autoconf): I don't see how this fixes an Autoconf bug, so I'm inclined to not install this until perhaps after the next release (since we're trying to stabilize). ___ Reply to this item at:

Re: CFLAGS containing quoted arguments to compiler does not work

This is a known limitation of Autoconf; see: https://www.gnu.org/software/autoconf/manual/autoconf-2.71/html_node/Special-Chars-in-Variables.html

[sr #110740] AC_PROG_MKDIR_P has incorrect result on NetBSD

Follow-up Comment #7, sr #110740 (project autoconf): [comment #6 comment #6:] > Assuming postgresql wants to work with current Autoconf, it can put in something like "case $MKDIR_P in /*) ;; */*) MKDIR_P='mkdir -p';; esac" ... Oops, I see that Savannah botched the rendering of that shell code. He

[sr #110740] AC_PROG_MKDIR_P has incorrect result on NetBSD

Update of sr #110740 (project autoconf): Priority: 3 - Low => 5 - Normal Severity: 2 - Minor => 3 - Normal Open/Closed:Open => Closed _

Re: [sr #110740] AC_PROG_MKDIR_P has incorrect result on NetBSD

On 2022-10-11 10:04, Tom Lane wrote: it appears to me that POSIX has required race-free-ness since the nineties, so do you really need to? As I recall, NetBSD had the race bug more recently than the 1990s. Are the newer-than-1990s buggy NetBSD versions EOLed yet? ("EOL" means NetBSD doesn't s

[sr #110740] AC_PROG_MKDIR_P has incorrect result on NetBSD

Update of sr #110740 (project autoconf): Severity: 3 - Normal => 2 - Minor ___ Follow-up Comment #3: This is not a serious bug, as "make install" is supposed to fall back on install-sh in this cas

[sr #110740] AC_PROG_MKDIR_P has incorrect result on NetBSD

Update of sr #110740 (project autoconf): Priority: 5 - Normal => 3 - Low ___ Reply to this item at: ___

Re: m4sh -- BINSH diversion cannot be cleared

e suggests m4_divert([0]) in a place where a named diversion wouldn't work, and tests/m4sugar.at tests that. Can you think of a good fix for this?From 55dfa2377602e743d00b726344422acc3116d859 Mon Sep 17 00:00:00 2001 From: Paul Eggert Date: Sun, 18 Sep 2022 11:52:04 -0700 Subject: [PATCH] Mention l

Re: m4sh -- BINSH diversion cannot be cleared

Thanks for the bug report and analysis. I installed into Autoconf the attached, which should fix the problem in the way you suggested.From 30bf5c49825ec2092c63b8519678e3e06148a8b5 Mon Sep 17 00:00:00 2001 From: Paul Eggert Date: Sat, 17 Sep 2022 19:43:04 -0700 Subject: [PATCH 1/2] Allow users to

[sr #110716] stdbool.h related tests failing on macOS 12 with M1 chip

Update of sr #110716 (project autoconf): Status:None => Duplicate Open/Closed:Open => Closed ___ Follow-up Comment #1: Thanks for reporti

[sr #110687] AC_C_BIGENDIAN fails when cross-compiling with -std=c11 and -flto

Update of sr #110687 (project autoconf): Status:None => Done Open/Closed:Open => Closed ___ Follow-up Comment #5: Thanks for the bug

Re: what platforms don't have `dirname`?

On 7/21/22 10:22, Werner LEMBERG wrote: it seems to me that the advice, while useful for autoconf, is very outdated. Thanks, I installed the attached. From 949e49297940529d2803a274617af8951f18e82b Mon Sep 17 00:00:00 2001 From: Paul Eggert Date: Thu, 21 Jul 2022 12:49:50 -0700 Subject

Re: the cost of due diligence

On 6/28/22 19:11, Jim Meyering wrote: ['EGR(EP|AC)_TRADITIONAL$'])])]) What is the "|AC" part intended to match? Nothing, it's just noise designed to make the pattern an ERE and not a BRE.

Re: the cost of due diligence

Thanks for fixing that old bug of mine. I installed the attached further patch to refactor and simplify. If I'd done it this way originally the bug wouldn't have existed.From 6e3282fe3893a42eb9c5cc4eba5387a3be4e8d10 Mon Sep 17 00:00:00 2001 From: Paul Eggert Date: Tue, 28 Jun 202

[sr #110657] Autoconf configure script considers absence of usable grep a fatal error, although it only uses egrep thereafter

Update of sr #110657 (project autoconf): Status: Wont Do => Done Open/Closed:Open => Closed ___ Follow-up Comment #12: [comment #11 comm

[sr #110657] Autoconf configure script considers absence of usable grep a fatal error, although it only uses egrep thereafter

Follow-up Comment #9, sr #110657 (project autoconf): [comment #8 comment #8:] > there is another way than -e to escape that dash: simply put a pair of parenthesis around the search string: > > egrep '(-z)' https://file.savannah.gnu.org/file/0001-Fix-bug-AC_EGREP_CPP-with-leading-in-pattern.patch?

[sr #110657] Autoconf configure script considers absence of usable grep a fatal error, although it only uses egrep thereafter

Follow-up Comment #7, sr #110657 (project autoconf): [comment #6 comment #6:] > How about the attached patch instead? Oops, that patch had a stray -E so I removed it. Revised patch attached. (file #53277) ___ Additional Item Attachment: F

[sr #110657] Autoconf configure script considers absence of usable grep a fatal error, although it only uses egrep thereafter

Follow-up Comment #6, sr #110657 (project autoconf): [comment #5 comment #5:] > the only reason why AC_PROG_EGREP is brought in (in my case) is due to the macro AC_EGREP_CPP, which does not ever use the -e option AC_EGREP_CPP should use -e though, since it should work even if the pattern begins wi

[sr #110656] On AT&T UnixPC (3b1), parameter $3 gets overwritten (cleared) by call to subfunction.

Follow-up Comment #6, sr #110656 (project autoconf): > dash is much less portable than bash, due to its very incomplete autoconf script Oh, I meant only that dash _once working_ should be a better shell to use for running 'configure' and similar shell scripts on the 3b1, since dash should use less

[sr #110656] On AT&T UnixPC (3b1), parameter $3 gets overwritten (cleared) by call to subfunction.

Follow-up Comment #4, sr #110656 (project autoconf): > Only snag was that it failed to write out confcache and Cache variables to config.log due to out of memory condition. I suggest using dash instead of bash for this sort of thing. The 'configure' script should be portable to any POSIX-compatibl

[sr #110657] Autoconf configure script considers absence of usable grep a fatal error, although it only uses egrep thereafter

Update of sr #110657 (project autoconf): Priority: 3 - Low => 1 - Later ___ Reply to this item at: ___

[sr #110657] Autoconf configure script considers absence of usable grep a fatal error, although it only uses egrep thereafter

Update of sr #110657 (project autoconf): Status: Need Info => Wont Do ___ Follow-up Comment #3: OK, I think I see what's happening now; you have an egrep that's good enough, but you also want

[sr #110660] AC_LANG_INT_SAVE (used by AC_CHECK_SIZEOF) includes stdlib.h without checking whether it actually exists

Update of sr #110660 (project autoconf): Priority: 5 - Normal => 3 - Low Severity: 3 - Normal => 1 - Wish ___ Follow-up Comment #3: > This fails if st

[sr #110653] configure fails on AT&T UnixPC (3b1) due to faulty LINENO substitution

Update of sr #110653 (project autoconf): Status:None => Duplicate Open/Closed:Open => Closed ___ Follow-up Comment #1: Closing this as a

[sr #110654] On AT&T UnixPC (3b1) autoconf script fails to properly replace $LINENO variable with line number

Update of sr #110654 (project autoconf): Status:None => Done Open/Closed:Open => Closed ___ Follow-up Comment #1: Thanks for the bug

[sr #110657] Autoconf configure script considers absence of usable grep a fatal error, although it only uses egrep thereafter

Update of sr #110657 (project autoconf): Priority: 5 - Normal => 3 - Low Status:None => Need Info ___ Follow-up Comment #1: > For example, gre

[sr #110655] AT&T UnixPC (3b1) shell thrown off by #( after case.

Update of sr #110655 (project autoconf): Status:None => Wont Do Open/Closed:Open => Closed ___ Follow-up Comment #1: The #( is not need

[sr #110656] On AT&T UnixPC (3b1), parameter $3 gets overwritten (cleared) by call to subfunction.

Update of sr #110656 (project autoconf): Status:None => Wont Do Open/Closed:Open => Closed ___ Follow-up Comment #2: For this one I thi

Re: set -e causes AS_IF to exit in else clause

Thanks for reporting that regression. I installed the attached patch.From 1684650184aea9ed51ded79680bb1b4fcb098f26 Mon Sep 17 00:00:00 2001 From: Paul Eggert Date: Thu, 31 Mar 2022 22:29:05 -0700 Subject: [PATCH] =?UTF-8?q?Fix=20bug=20with=20AS=5FIF=20and=20=E2=80=98set?= =?UTF-8?q?=20-e=E2=80

Re: [1003.1(2016/18)/Issue7+TC2 0001558]: require [^...] in addition to [!...] for bracket expression negation

1c14f36305836ef135dd3 Mon Sep 17 00:00:00 2001 From: Paul Eggert Date: Fri, 18 Feb 2022 15:03:00 -0800 Subject: [PATCH] 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/ms

[sr #110571] Autoconf-2.71 misconverts new-line escapes in AC_CHECK_FUNCS() argument

Update of sr #110571 (project autoconf): Status:None => Invalid Open/Closed:Open => Closed ___ Reply to this item at:

[sr #110503] Autoconf 2.70 problem: gkt-doc/gtkdocize is now unconditionally required

Follow-up Comment #4, sr #110503 (project autoconf): [comment #3 comment #3:] > Can you clarify what you mean by self-contained example? All it takes is to create a configure.ac with the snippet I quoted in the previous comment and run > autoreconf -fiv. I mean a standalone shell script that repro

[sr #110503] Autoconf 2.70 problem: gkt-doc/gtkdocize is now unconditionally required

Update of sr #110503 (project autoconf): Status:None => Need Info Summary: Autoconf 2.70 regression: gkt-doc/gtkdocize is now unconditionally required => Autoconf 2.70 problem: gkt-doc/gtkdocize is now unconditionally required

[sr #110521] autom4te does not always update cache, breaks if change happens within 1 second

Update of sr #110521 (project autoconf): Status:None => Done Open/Closed:Open => Closed ___ Follow-up Comment #2: Thanks for the bug

  1   2   3   4   5   6   >