Current CVS contains a couple of syntactic typos, patch below.
If fixes for these are contained in pending patches, please just ignore
this mail (I have not checked and am still catching up on mail).
Regards,
Ralf
* doc/autoconf.texi (Generic Functions): Typos.
Index: doc/autoconf.texi
Hi Paul,
Thanks for the review.
* Paul Eggert wrote on Fri, May 20, 2005 at 09:18:46PM CEST:
Ralf Wildenhues [EMAIL PROTECTED] writes:
+eval `./config.status --config`
+./config.status --recheck ./config.status
This doesn't look right to me. Shouldn't the argument to eval be
quoted
* Ralf Wildenhues wrote on Thu, Jun 09, 2005 at 09:20:53AM CEST:
| autoreconf: Entering directory `.'
| autoreconf: configure.ac: not using Gettext
| autoreconf: running: aclocal --force -I m4
| configure.ac:186: /usr/local/bin/m4: ERROR: Recursion limit of 1024
exceeded, use -LN to change
Hi Ralf,
You just made my day. :)
* Ralf Menzel wrote on Thu, Jun 09, 2005 at 06:50:18PM CEST:
Index: m4/libtool.m4
[...]
-[lt_join($@, lt_decl_varnames_tagged([$1],
[...]
Well I head a look at the traces. This indicated that lt_combine might
be the culprit. (lt_join uses
* Ralf Wildenhues wrote on Wed, Jun 22, 2005 at 06:45:35PM CEST:
3)
m4/ltoptions.m4
m4_define([_LT_SET_OPTIONS],
-[AC_FOREACH([_LT_Option], [$1], [...]
+[m4_if([$1], [], [],
+ [AC_FOREACH([_LT_Option], [$1], [...]
This change is a workaround for a bug
Hi Stepan,
* Stepan Kasal wrote on Mon, Jun 27, 2005 at 12:34:12PM CEST:
On Wed, Jun 22, 2005 at 06:45:35PM +0200, Ralf Wildenhues wrote:
*snip nice explanations*
But:
There are two cases when the second argument to m4_default is not
double quoted; I think these should be fixed.
So I
Hi Gary,
I've put autoconf-patches back in -- they are the target of this. :)
* Gary V. Vaughan wrote on Mon, Jun 27, 2005 at 03:58:43PM CEST:
Ralf Wildenhues wrote:
Stepan Kasal wrote on Mon, Jun 27, 2005 at 12:42:50PM CEST:
AC_FOREACH was never documented, so there is no problem if you
* Ralf Wildenhues wrote on Wed, Jun 29, 2005 at 02:48:57PM CEST:
* lib/autoconf/fortran.m4 (_AC_PROG_FC_V_OUTPUT): Fix also for
double-quoted -cmdline argument in Portland Group compiler.
Reported against LAM by Ole Holm Nielsen [EMAIL PROTECTED].
Gee, I got it all backwards
With Portland 5.2-4 and 6.0-4 compilers it is necessary to remove some
single quotes from verbose Fortran linking output. Steven applied a
similar patch for double quoted stuff some time ago[1]. A bug report
against LAM[2] suggests the patch below, I don't know of a better
solution right away.
Hi Stepan, others,
This is a reply to
http://lists.gnu.org/archive/html/libtool-patches/2005-08/msg2.html
which starts at
http://lists.gnu.org/archive/html/libtool-patches/2005-07/msg00128.html
and
http://lists.gnu.org/archive/html/autoconf-patches/2005-07/msg00032.html
* Stepan Kasal wrote
Hi Stepan,
* Stepan Kasal wrote on Fri, Aug 12, 2005 at 01:39:43PM CEST:
Hi,
I noticed that Makefile.maint supposes that gzip prints its
help text to stdout.
But my gzip 1.2.4a, which is still the latest stable version released,
prints its help test to stderr.
So Makefile.maint has to
Hi Stepan,
* Stepan Kasal wrote on Wed, Aug 10, 2005 at 05:05:51PM CEST:
On Wed, Aug 10, 2005 at 03:42:18PM +0200, Ralf Wildenhues wrote:
no, I don't think using ac_config_headers is a good idea.
I knew it. ;)
So I passed the test. :-)
Other way round. I tried to make you do
[ only to autoconf-patches ]
* Stepan Kasal wrote on Sat, Aug 13, 2005 at 09:16:08AM CEST:
On Sat, Aug 13, 2005 at 12:00:13AM +0200, Ralf Wildenhues wrote:
Hmm. But this:
| AC_CONFIG_COMMANDS_PRE([dnl
| m4_ifndef([AH_HEADER], [AC_CONFIG_HEADERS([config.h])])
| AC_SUBST([CONFIG_H
The following (arguably ugly-written) test shows that
./testsuite -k key1 -k key2
do not work as advertised.
I can guess that my patch[1] which fixed the behavior of
./testsuite --help
in the presence of multiple keys per test was wrong or insufficient;
./testsuite -k key1 -k key2 --help
This is continuing an old old thread from the autoconf list:
http://lists.gnu.org/archive/html/autoconf/2005-03/msg00117.html
I moved it to autoconf-patches now.
* Noah Misch wrote on Thu, Mar 24, 2005 at 03:47:45AM CET:
On Tue, Mar 22, 2005 at 10:05:04AM +0100, Ralf Wildenhues wrote:
I'm
Hi Noah,
* Noah Misch wrote on Tue, Aug 16, 2005 at 01:50:15AM CEST:
On Mon, Aug 15, 2005 at 06:21:16PM +0200, Ralf Wildenhues wrote:
* tests/autoscan.at (autoscan): New file. Test autoscan
traversal of subpackages Makefile's.
Reported against Libtool by Gideon Go
Hi Noah,
* Noah Misch wrote on Tue, Aug 16, 2005 at 01:50:57AM CEST:
On Mon, Aug 15, 2005 at 06:22:09PM +0200, Ralf Wildenhues wrote:
* tests/autotest.at (Keywords): Test keywords combinations.
Marked XFAIL and Applied. Thanks.
And here's the fix for this one, three other bugs
Hi Stepan,
* Stepan Kasal wrote on Tue, Aug 16, 2005 at 04:28:04PM CEST:
I noticed a problem with your patch:
*snip excerpt*
So --keyword might select several banners, but they are 0 for the purpose
of sort -nu. This means that one of the banners will be at the top, and all
others are
* Stepan Kasal wrote on Tue, Aug 16, 2005 at 07:15:13PM CEST:
On Tue, Aug 16, 2005 at 06:10:52PM +0200, Ralf Wildenhues wrote:
I happen not to understand the exact nature of the bug report
that led to the breakage of this code in the first place:
http://lists.gnu.org/archive/html/autoconf
Hi Stepan,
* Stepan Kasal wrote on Wed, Aug 17, 2005 at 12:58:39PM CEST:
On Tue, Aug 16, 2005 at 08:33:12PM +0200, Ralf Wildenhues wrote:
Sure. Feel free to adjust the patch accordingly.
OK, I've prepared a variation of your patch.
It passes make check on my GNU/Linux.
Could you please
Hi Stepan,
* Stepan Kasal wrote on Tue, Aug 16, 2005 at 12:15:51PM CEST:
+Since they have to appear before @file{config.status}
+is created, they may be called no later than in an argument to
[EMAIL PROTECTED]
This formulation has two problems:
1) Many users would not understand
Hi Stepan,
* Stepan Kasal wrote on Wed, Aug 17, 2005 at 07:31:12PM CEST:
On Wed, Aug 17, 2005 at 06:36:46PM +0200, Ralf Wildenhues wrote:
[ solaris awk ]
| 21: m4sh.at:252AS_VERSION_COMPARE
loops endlessly (in awk).
This is for calling AS_VERSION_COMPARE with empty
Hi Paul,
* Paul Eggert wrote on Thu, Aug 25, 2005 at 11:31:30PM CEST:
Ralf Wildenhues [EMAIL PROTECTED] writes:
- if (v1 == ) d1 = v1; else { d1 = substr(v1, 1, 1); v1 = substr(v1,2) }
- if (v2 == ) d2 = v2; else { d2 = substr(v2, 1, 1); v2 = substr(v2,2) }
+ if (v1 == ) d1 = v1
Hi Stepan, Paul,
* Paul Eggert wrote on Thu, Sep 08, 2005 at 01:09:43AM CEST:
Stepan Kasal [EMAIL PROTECTED] writes:
2005-08-31 Stepan Kasal [EMAIL PROTECTED]
* lib/autoconf/lang.m4 (_AC_COMPILER_EXEEXT_DEFAULT): Don't try to
detect exeext, it will be done by
* Ralf Wildenhues wrote on Sat, Oct 01, 2005 at 09:12:43PM CEST:
This patch fixes that -- I'm not sure whether `chmod -R' is portable
enough though, but it's a lot less expensive than
find $at_group_dir -exec chmod u+w \{\} \;
and I don't think you can expect file names within a test
Hi Paul,
* Paul Eggert wrote on Sun, Oct 02, 2005 at 11:46:06PM CEST:
Ralf Wildenhues [EMAIL PROTECTED] writes:
* lib/autotest/general.m4 (AT_INIT): Make test dir content
writable before removing.
I have a few qualms about invoking chmod u+w as I worry it might run
afoul
Hi Stepan,
* Stepan Kasal wrote on Tue, Oct 04, 2005 at 11:15:01AM CEST:
On Mon, Oct 03, 2005 at 12:41:04PM -0700, Paul Eggert wrote:
I was worried about the case where it wasn't erased, and where the
test contained a symlink to /usr/bin or something like that.
Admittedly I'm waving my
Hi Paul,
* Paul Eggert wrote on Tue, Oct 04, 2005 at 09:17:54PM CEST:
Ralf Wildenhues [EMAIL PROTECTED] writes:
+ test -d $at_suite_dir
+ find $at_suite_dir -type l -prune -o -type d -exec chmod u+rwx \{\}
\;
rm -f -r $at_suite_dir $at_suite_log
Why do you need the -type
Hi there,
I'm not quite certain about this patch, but I get the impression that
users may need to be pushed a bit more prominently to the location of
testsuite.{log,dir}. The patch below is a suggestion to this extent.
Surely for users not executing the testsuite from the toplevel build
dir the
Hi there,
The AC_CHECK_TOOL/AC_PATH_TOOL and similar macros cause variables of the
form
ac_ct_*
ac_pt_*
to be AC_SUBSTed. This looks unnecessary to me: the variables are not
documented, neither are they used outside configuration; they are merely
helper variables to avoid conflicts with the
Darn.
* Ralf Wildenhues wrote on Sat, Nov 05, 2005 at 07:42:32AM CET:
* lib/autoconf/programs.m4 (AC_CHECK_PROG, AC_PATH_PROG):
Factor functionality to..
(_AC_CHECK_PROG, _AC_PATH_PROG): these new macros, but only
`AC_SUBST($1)' in the public version
[ I've Cc:ed Dan as the author of this ]
Sorry for answering myself all the time.
* Ralf Wildenhues wrote on Tue, Nov 08, 2005 at 12:59:45PM CET:
It's an off-by-one in status.m4, I just don't know where.
Has nothing to do with Libtool.
I can reproduce it with
[ see shell script below
Hi Paul,
* Paul Eggert wrote on Thu, Nov 10, 2005 at 01:03:59AM CET:
That patch looks good to me, except:
+ CONFIG_SHELL=$SHELL exec $SHELL $[0] $ac_configure_args
\$ac_configure_extra_args --no-create --no-recursion
Many shells treat A=B exec COMMAND differently; they ignore the
I have applied this tiny patch, in the hope that it was ok to do so.
Aside: by the old copyright year rules, this would be a small patch
not needing an update; by the new ones, apparently either none or
all files could be changed at once; it may be good to know how
Autoconf developers choose to
A bit ugly. But works for AS_HELP_STRING. :-)
OK to apply (including copyright year updates)?
Does this break any other m4_text_wrap uses?
If yes, are they important enough to worry about? I could only find
the autoheader one, which should not be impacted by the patch below.
Cheers,
Ralf
* Paul Eggert wrote on Wed, Jan 11, 2006 at 07:41:04AM CET:
Ralf Wildenhues [EMAIL PROTECTED] writes:
OK to apply (including copyright year updates)?
Looks good to me.
Thanks. Applied as below. I took liberty to also add a NEWS entry.
Cheers,
Ralf
* lib/m4sugar/m4sugar.m4
Hi Greg,
* Greg Lindahl wrote on Wed, Jan 11, 2006 at 06:47:53PM CET:
--lang* | -lcrt*.o | -lc | -lgcc | -libmil | -LANG:=*)
+-lang* | -lcrt*.o | -lc | -lgcc | -libmil | -LANG:=* | -LIST:* |
-LNO:*)
This looks good to me. Note that pathcc/pathCC also take -LANG, -LIST,
In http://thread.gmane.org/gmane.comp.sysutils.autoconf.general/5849
the move of CONFIG_SHELL to the environment of `configure' wasn't done.
In http://thread.gmane.org/gmane.comp.gnu.libtool.patches/6368 we
concluded it to be necessary though. One instance of this was missed.
OK to apply?
Hi Eric,
* Eric Blake wrote on Thu, Feb 16, 2006 at 09:35:28PM CET:
According to Ralf Wildenhues on 2/15/2006 5:22 PM:
+dnl Make sure we do not divide by zero:
You probably also ought to have tests:
m4_for([myvar], 1, 2, 0, [ myvar])
m4_for([myvar], 1, 1, 0, [ myvar])
m4_for([myvar], 2
2) Made AS_IF n-ary. Then, I extended the test suite because I tripped
in the why in the world should AS_IF need to be m4_defun'ed? trap
and don't want that to happen again.
1) Add a couple of notes to the docs to this end. I'm not certain
whether they are too confusing to the (ahem)
Hi Stepan,
* Stepan Kasal wrote on Sun, Feb 19, 2006 at 12:58:31PM CET:
On Fri, Feb 17, 2006 at 06:02:15PM +0100, Ralf Wildenhues wrote:
(Obsolete Macros): Document AC_FOREACH.
Ii'm not sure whether AC_FOREACH should be documented.
``Obsolete Macros'' is an ``attic'' for macros which
Hi Paul,
* Paul Eggert wrote on Mon, Feb 20, 2006 at 11:09:29AM CET:
Ralf Wildenhues [EMAIL PROTECTED] writes:
[EMAIL PROTECTED] be more precise, the required macro will be
+expanded before the outermost @code{AC_DEFUN}'d macro in the current
+expansion stack.}
Can you please reword
Hi Stepan,
* Stepan Kasal wrote on Mon, Feb 20, 2006 at 03:15:33PM CET:
On Sun, Feb 19, 2006 at 05:16:55PM +0100, Ralf Wildenhues wrote:
Look, the chapter says:
|You may completely skip this chapter if you are new to Autoconf. Its
| intention is mainly to help maintainers updating
Hi Eric,
* Eric Blake wrote on Mon, Feb 20, 2006 at 09:36:53PM CET:
Are minor optimizations such as this worthwhile? It leads to
smaller generated configures, and allows the shell to perform
fewer pattern matches during operation.
If you ask me, definitely: $LIBOBJS may easily be a few K
Hi Eric,
* Eric Blake wrote on Tue, Feb 21, 2006 at 02:44:33PM CET:
According to Ralf Wildenhues on 2/21/2006 12:16 AM:
Note your actual diff posted had some whitespace differences wrt. the
actual code so it did not apply cleanly (TAB vs spaces).
Sorry about that - my ISP's webmail
OK to apply?
m4_bpatsubsts (sic) and m4_text_wrap should probably be documented, too.
Likewise, m4_join and m4_car/m4_cdr would be helpful. I refrained from
documenting those for now because I think these need more investigation
(remembering how bad Libtool's lt_join scaled some time ago). IMHO
Hi Paul,
Thanks for the review, and sorry for the broken patch. :-/
* Paul Eggert wrote on Wed, Feb 22, 2006 at 07:10:09AM CET:
Ralf Wildenhues [EMAIL PROTECTED] writes:
+Grow strings (without duplicating substrings).
This doesn't make it clear that m4_append grows, and m4_append_uniq
Real life Autoconf macros sometimes are too complex to fit their
results in one cache variable. But they are also too expensive
to be run several times. Several of these exist and are widely used.
The patch below adds an example of how such a test may be written
properly. I'm a bit uncertain
Hi Paul,
* Paul Eggert wrote on Wed, Mar 01, 2006 at 01:48:54AM CET:
Several macros in Gnulib really belong in Autoconf. I took a first
cut at migrating the code, and came up with the following proposed
patch to Autoconf. Not all the gnulib macros made the cut, and some
needed to be renamed
OpenBSD grep has a bug in that it mistreats patterns separated by
newlines. I will report more on this after finishing a bug report to
this end. For now, it's only important that Autoconf does not use this
in lib/*/*m4, but in tests/mktests.sh.
Now, tests/Makefile.am is not too portable, due to
Hi Paul,
* Paul Eggert wrote on Sun, Mar 05, 2006 at 09:32:18AM CET:
Ralf Wildenhues [EMAIL PROTECTED] writes:
I'd guess you prefer the unified behavior but a bit slower everywhere
instead, done in the patch below.
Thanks. I installed the following instead, as it avoids a couple
Hi Paul,
* Paul Eggert wrote on Sat, Feb 25, 2006 at 09:03:50AM CET:
Ralf Wildenhues [EMAIL PROTECTED] writes:
The patch below adds an example of how such a test may be written
properly. I'm a bit uncertain as to whether it should be applied:
we would constrain Autoconf in a way that we
* Ralf Wildenhues wrote on Mon, Mar 06, 2006 at 10:13:24PM CET:
Applied.
Hrmpf. That required regenerating tests/Makefile.in with
Automake-1.9.6. That exposed an interesting bug in that version (since
fixed in CVS Automake): {tests,man}/Makefile.in contain code like this:
| $(mkdir_p
I apologize for posting ever more patches in what was supposed to be
some stabilization work..
This patch and discussion
http://lists.gnu.org/archive/html/autoconf-patches/2005-08/msg00038.html
http://lists.gnu.org/archive/html/autoconf-patches/2005-08/msg00027.html
changed a comment in
(1)
The Tru64 UNIX 4.0D sed(1) manpage claims
| [Digital] The sed command can process up to 99 commands in a file.
The /bin/sed on this alphaev5-dec-osf4.0d supports 999 commands and 48
labels, though. Empty lines do not count as commands, neither does a
closing brace `}' nor a comment, nor
In tests/autotest.at, the checks
| Parenthetical command substition
and
| Multiline parenthetical command substition
use the $(cmd) style command substition, which is not portable.
So, let's put in an AT_XFAIL_IF. Hmm. Doesn't work:
AT_CHECK_AT creates a `micro_suite' which is executed as
* Paul Eggert wrote on Fri, Mar 10, 2006 at 12:04:42AM CET:
Ralf Wildenhues [EMAIL PROTECTED] writes:
+except that this will fail on Solaris and OpenBSD 3.8.
Please install, but could you please broaden that a bit first?
Done. Thanks!
Cheers,
Ralf
* Paul Eggert wrote on Thu, Mar 09, 2006 at 10:35:46PM CET:
Ralf Wildenhues [EMAIL PROTECTED] writes:
* lib/autoconf/lang.m4 (AC_LANG_CONFTEST): AC_DEFUN this, not
m4_define, so that the requirements of `AC_INCLUDES_DEFAULT' are
expanded outside.
Thanks; please apply.
Done
Hi Paul,
* Paul Eggert wrote on Thu, Mar 09, 2006 at 11:48:10PM CET:
Ralf Wildenhues [EMAIL PROTECTED] writes:
* doc/autoconf.texi (autoconf Invocation): Fix typos in trace
example. Do not emphasize `$%', it is hardly new and special.
Reported by Edouard BECHETOILLE [EMAIL
Hi Noah,
* Noah Misch wrote on Sun, Mar 05, 2006 at 07:46:49AM CET:
On Wed, Mar 01, 2006 at 08:34:56PM +0100, Ralf Wildenhues wrote:
* tests/autotest.at (AT_NO_CMDSUBST): New macro to determine
XFAIL condition for `$(cmd)' style command substitutions.
(Parenthetical command
`make html' invokes
texi2html -split_chapter $(srcdir)/autoconf.texi
to create a multi-page HTML version of the manual. Issues
observed with texi2html 1.64 (as distributed with tetex-2.0.2):
- The @insertcopying is not replaced. Presumably this is
why the web page does not link to a copy of
Hi Stepan, Noah,
* Noah Misch wrote on Wed, Mar 15, 2006 at 05:03:57PM CET:
On Wed, Mar 15, 2006 at 12:47:11PM +0100, Stepan Kasal wrote:
well, the following patch would save 170 bytes.
A space optimization of this magnitude is irrelevant at the scope of a
configure
script. Let us
* Ralf Wildenhues wrote on Wed, Mar 15, 2006 at 05:30:43PM CET:
Second, your change to _AC_CANONICAL_SPLIT is more than just a change of
names, but a latent bugfix (presumably the one that you worked around
with some change implemented soon after you started using $IFS in
_AC_CANONICAL_SPLIT
[ from autoconf and autoconf-patches ]
* Ralf Wildenhues wrote on Thu, Mar 16, 2006 at 10:19:18AM CET:
This patch comes with a sibling patch to fix support for OBJC in
Automake.
Here it is. Since CVS Automake requires CVS Autoconf anyway, I guess
it's ok to rely on presence of AC_PROG_OBJC
This started out here and here:
http://lists.gnu.org/archive/html/libtool/2004-07/msg00070.html
http://lists.gnu.org/archive/html/autoconf-patches/2004-07/msg3.html
and led to (CVS) Autoconf exporting BIN_SH=xpg4 (don't be fooled by the
patches posted; they do not fully reflect what was
* Stepan Kasal wrote on Fri, Mar 17, 2006 at 06:16:43PM CET:
on Tue, Mar 14, 2006 at 07:44:47PM +0100, Ralf Wildenhues wrote:
- The @insertcopying is not replaced. [...]
Both issues seem to be fixed in texi2html 1.76 [...]
do I understand correctly that to add the HTML manual to the web
This changes all usage of $0 (and thus $as_me and $as_myself) in
autoconf/lib/*/*.m4 to not mind white space. OK to apply?
I've tested with the _AS_LINENO_PREPARE part with the OSF /usr/bin/sh
Bourne shell (which does not support LINENO).
Note that surely there may be other reasons that
Hi Stepan,
* Stepan Kasal wrote on Mon, Mar 20, 2006 at 01:24:12PM CET:
using your fossile config.cache files, you ended up with
| INSTALL=''/usr/bin/install -c''
in your config status.
When searching for a victim to blame, you found:
- INSTALL=$INSTALL
+ INSTALL='$INSTALL'
Hi Paul,
* Paul Eggert wrote on Fri, Mar 17, 2006 at 07:17:25AM CET:
Ralf Wildenhues [EMAIL PROTECTED] writes:
Further question: document Objective C right after C++ (logical) or
after Erlang node (alphabetical)? Try 'gcc' before the other compilers?
I'd say logical (which is what you
* Stepan Kasal wrote on Wed, Mar 22, 2006 at 12:26:16AM CET:
There is a second patch which documents the ^ issue.
This is already in the docs:
| On some Posix-like platforms, `!' and `^' are special too, so they
| should be avoided.
[EMAIL PROTECTED] ^
[EMAIL PROTECTED] ^ quoting
+Always
[ this thread starts at:
http://lists.gnu.org/archive/html/autoconf-patches/2006-03/msg00118.html ]
Hi Stepan,
* Stepan Kasal wrote on Wed, Mar 22, 2006 at 12:49:20AM CET:
On Mon, Mar 20, 2006 at 09:33:50PM +0100, Ralf Wildenhues wrote:
AC_LANG_SAVE
AC_LANG_SAVE
will lead to m4
only address your suggestion:
* Stepan Kasal wrote on Wed, Mar 22, 2006 at 07:51:12PM CET:
On Tue, Mar 21, 2006 at 07:44:37PM +0100, Ralf Wildenhues wrote:
the real question is one of commitment to these interfaces.
Exactly.
First, the question of _AS_PREPARE. I agree this hack shall
Hi Laurence, Bruce,
* Laurence Darby wrote on Sun, Mar 26, 2006 at 05:01:19PM CEST:
Bruce Korb wrote:
No. Autoscan should:
if test -f configure.ac
else
echo NOTE: creating configure.ac file 2
configure_file=configure.ac
touch configure.ac
fi
[ Cc:ing because the list server is so unreliable ATM; sorry ]
* Ralf Wildenhues wrote on Mon, Mar 20, 2006 at 08:29:48PM CET:
* Paul Eggert wrote on Mon, Mar 20, 2006 at 07:47:57PM CET:
Ralf Wildenhues [EMAIL PROTECTED] writes:
- ' $as_myself |
+ ' $as_myself |
These two changes
Hi Stepan,
* Stepan Kasal wrote on Wed, Mar 22, 2006 at 12:49:20AM CET:
I have a patch prepared on my disk.
The problem was that if configure.ac started with
AC_PREREQ(2.54)
m4_define([gnumeric_version_epoch], [1])
...
then the m4_define disappeared after an autoupdate
Hi Eric,
* Eric Blake wrote on Wed, Mar 22, 2006 at 02:35:05PM CET:
After my recent patch submission to coreutils, GNU dirname/basename will
now return // on platforms (for now, cygwin) that have // distinct from /.
So the documentation needs an update; in the process, I noticed that
Hi Paul,
Thanks for the review!
* Paul Eggert wrote on Fri, Mar 24, 2006 at 10:27:59PM CET:
Ralf Wildenhues [EMAIL PROTECTED] writes:
+# Do not use TABs in .texi files.
Hmm, come to think of it, the Texinfo manual says it's OK to use tabs
in verbatim modes, and I use that in some of my
The patch below fixes the autoscan buglet reported a while ago.
So, if the user already has a working configure.ac, autoscan won't add
unnecessary checks of AC_CONFIG_FILES for stuff found in subpackages.
If configure.ac is present but not parseable (due to syntax errors, for
example), then
Hi Paul,
Here is a preliminary version of the proposed rewrite; I'll try to
finish it tomorrow, unless beaten to.
* Ralf Wildenhues wrote on Fri, Mar 31, 2006 at 09:31:38AM CEST:
* Paul Eggert wrote on Fri, Mar 31, 2006 at 09:10:25AM CEST:
Ralf Wildenhues [EMAIL PROTECTED] writes
Hi Eric, Stepan,
* Stepan Kasal wrote on Sun, Apr 02, 2006 at 12:49:25AM CEST:
2006-04-01 Eric Blake [EMAIL PROTECTED]
* lib/m4sugar/m4sh.m4 (_AS_PATH_WALK): Optimize nonexistent
directories, unless optional third argument supplied.
(AS_UNAME): Don't optimize PATH walk.
Hi Stepan,
* Stepan Kasal wrote on Sat, Apr 01, 2006 at 11:25:19PM CEST:
On Sat, Apr 01, 2006 at 01:08:57PM -0800, Noah Misch wrote:
On Sat, Apr 01, 2006 at 11:05:58PM +0200, Stepan Kasal wrote:
Hello,
does anyone know why are generated tests/*.at files
(see TESTSUITE_GENERATED_AT
Hi Paul,
* Paul Eggert wrote on Sun, Apr 02, 2006 at 07:22:54PM CEST:
Ralf Wildenhues [EMAIL PROTECTED] writes:
OK, so I removed the generated *.at files from the CVS.
Please undo that. The creation of these files is not portable:
But these files aren't intended to be built
Hi Noah,
* Noah Misch wrote on Sun, Apr 02, 2006 at 09:10:09PM CEST:
autoconf-patches: ok?
Looks fine to me!
Cheers, and thanks!
Ralf
2006-04-02 Noah Misch [EMAIL PROTECTED]
* lib/autoconf/general.m4 (AC_CACHE_SAVE): Use AC_MSG_NOTICE.
Hi Noah,
* Noah Misch wrote on Sun, Apr 02, 2006 at 08:17:55PM CEST:
This is an old patch from a thread of autoconf@gnu.org beginning with
[ http://lists.gnu.org/archive/html/autoconf/2005-08/msg00065.html ]
Message-ID: [EMAIL PROTECTED]. It adds a macro,
AC_PRESERVE_HELP_ORDER, that makes
Hi Eric,
* Eric Blake wrote on Mon, Apr 03, 2006 at 03:03:09PM CEST:
I'm feeling left out
I'm truly sorry. I think I never checked for your presence in THANKS
because I assumed it.
(does anyone have a good script that can scan
ChangeLog to find anyone else that is inadvertantly left out?)
Hi Noah,
* Noah Misch wrote on Mon, Apr 03, 2006 at 07:23:54PM CEST:
If a package uses the zero- or one-argument form of AC_INIT, the warning for
present-but-not-usable headers looks like this:
configure: WARNING: ## -- ##
configure: WARNING:
* Noah Misch wrote on Mon, Apr 03, 2006 at 08:18:36PM CEST:
On Mon, Apr 03, 2006 at 07:59:24PM +0200, Ralf Wildenhues wrote:
* Noah Misch wrote on Mon, Apr 03, 2006 at 07:23:54PM CEST:
This patch makes `configure' omit the report request when we have no
bug report address to provide. Ok
Hi Stepan,
* Stepan Kasal wrote on Tue, Apr 04, 2006 at 05:55:20PM CEST:
On Tue, Apr 04, 2006 at 05:32:40PM +0200, Ralf Wildenhues wrote:
So it seems to me that `$egrep -f' has a good chance of success, as
it seems to work on all systems I looked at.
settled; thank you for helping
using CVS for Autoconf development.
OK to apply?
2006-04-04 Stepan Kasal [EMAIL PROTECTED],
Eric Blake [EMAIL PROTECTED],
Ralf Wildenhues [EMAIL PROTECTED]
* lib/autoconf/general.m4 (_AC_CACHE_DUMP): Fix the detection of
whether `set' quotes correctly
OK to apply this? It fixes some and enforces some but not all rules in
Makefile.maint's sytax-check. Some of the imported files violate
several rules (I suppose I should post a patch to doc/make-stds.texi to
gnulib?), also some rules are not so useful (the `space tab' combination
is used several
Hi Eric,
* Eric Blake wrote on Wed, Apr 05, 2006 at 03:21:46PM CEST:
According to Stepan Kasal on 4/1/2006 9:15 AM:
+ AS_MKDIR_P([$at_group_dir]) ||
AS_ERROR([cannot create $at_group_dir])
I just noticed that AS_MKDIR_P already does AS_ERROR (fatal to the script)
if it fails,
Hi Eric,
* Eric Blake wrote on Wed, Apr 05, 2006 at 05:58:31PM CEST:
Low hanging fruit (I was actually surprised it was this easy).
2006-04-05 Eric Blake [EMAIL PROTECTED]
* lib/autotest/general.m4 (AT_INIT): Prep AT_*_all, so that an
empty test suite works.
*
* Ralf Wildenhues wrote on Wed, Apr 05, 2006 at 12:19:03AM CEST:
* doc/autoconf.texi (Shell Substitutions): Document `^' vs. `|'.
FYI, I have applied this bit.
Index: doc/autoconf.texi
===
RCS file: /cvsroot/autoconf
Ping. It'd be nice to know if this has a chance of being acceptable,
or if I should just drop it.
Cheers,
Ralf
* Ralf Wildenhues wrote on Fri, Mar 17, 2006 at 05:50:37AM CET:
* doc/autoconf.texi (Specific Compiler Characteristics):
Document `#error' limitations.
Index: doc
A case study, FYI.
The Y-MP from Cray-Cyber has
- 31 value-bit shorts
- 63 value-bit ints, but also a mode in which fast arithmetic is done
with only 45 value-bits,
- 63 value-bit longs
All three report a sizeof of 8. CHAR_BIT is 8.
The 46 bit int mode is an optimization mode. The
Hi Paul,
* Paul Eggert wrote on Wed, Apr 05, 2006 at 09:15:58AM CEST:
Ralf Wildenhues [EMAIL PROTECTED] writes:
+AC_COMPILE_IFELSE(
+ [AC_LANG_BOOL_COMPILE_TRY([AC_INCLUDES_DEFAULT([$4])],
+ [$ac_max $ac_max1 ($1) $ac_max1 == $ac_max1])],
Strictly speaking, for signed
Hi Eric,
* Eric Blake wrote on Thu, Apr 06, 2006 at 05:14:06AM CEST:
I just noticed that AS_MKDIR_P already does AS_ERROR (fatal to the script)
if it fails,
Which sucks, frankly. Another badly designed interface that now is less
generally usable than it could be. And we have to
* Paul Eggert wrote on Thu, Apr 06, 2006 at 05:40:26AM CEST:
Noah Misch [EMAIL PROTECTED] writes:
On my middle-aged PC running GNU/Linux, rebuilding
`testsuite' takes 22 seconds.
It takes 16 real seconds on my 4-year-old PC (2.4 GHz P4). Of that,
15 real seconds are spent running
We have to go back once more.
* Stepan Kasal wrote on Fri, Mar 17, 2006 at 08:35:12PM CET:
On Fri, Mar 17, 2006 at 11:20:49AM -0800, Paul Eggert wrote:
Yes, except you need to modify the Autoconf documentation accordingly.
I checked in the version attached to this mail.
The BIN_SH removal
Hi Stepan,
* Stepan Kasal wrote on Thu, Apr 06, 2006 at 12:42:06PM CEST:
On Wed, Apr 05, 2006 at 01:10:40PM -0700, Noah Misch wrote:
We now trigger the TESTSUITE_GENERATED_AT update from check-local, so `make
-C
tests ./testsuite' may build a testsuite using out-of-date tests/ac*.at. I
1 - 100 of 1027 matches
Mail list logo