A tiny fix.
2010-07-30 Bruno Haible br...@clisp.org
mbrtowc: Fix cross-compilation guess for AIX.
* m4/mbrtowc.m4 (gl_MBRTOWC_INCOMPLETE_STATE): Fix cross-compilation
guess. Leftover from 2008-12-21.
--- m4/mbrtowc.m4.orig Fri Jul 30 10:52:03 2010
+++ m4/mbrtowc.m4
A similar one.
2010-07-30 Bruno Haible br...@clisp.org
mbrlen: Fix cross-compilation guess for AIX.
* m4/mbrlen.m4 (gl_MBRLEN_INCOMPLETE_STATE): Fix cross-compilation
guess. Leftover from 2008-12-22.
--- m4/mbrlen.m4.orig Fri Jul 30 10:59:13 2010
+++ m4/mbrlen.m4
On 29 July 2010 23:18, Reuben Thomas r...@sc3d.org wrote:
On 29 July 2010 23:14, Bruno Haible br...@clisp.org wrote:
Reuben Thomas wrote:
Thanks. I guess I should report this documentation bug somewhere, but
where? (As mentioned in a recent message to this list, I've not had
much mileage from
On 07/30/2010 11:33 AM, Reuben Thomas wrote:
I have submitted a bug report with patch at
http://sourceware.org/bugzilla/show_bug.cgi?id=11857
This addresses regex.h, but doesn't address the texinfo manual.
Unfortunately the texinfo manual is not currently shipped with glibc.
I have written to
On 30 July 2010 11:50, Paolo Bonzini bonz...@gnu.org wrote:
On 07/30/2010 11:33 AM, Reuben Thomas wrote:
I have submitted a bug report with patch at
http://sourceware.org/bugzilla/show_bug.cgi?id=11857
This addresses regex.h, but doesn't address the texinfo manual.
Unfortunately the
On 07/30/2010 01:01 PM, Reuben Thomas wrote:
On 30 July 2010 11:50, Paolo Bonzinibonz...@gnu.org wrote:
On 07/30/2010 11:33 AM, Reuben Thomas wrote:
I have submitted a bug report with patch at
http://sourceware.org/bugzilla/show_bug.cgi?id=11857
This addresses regex.h, but doesn't address
On 30 July 2010 12:02, Paolo Bonzini bonz...@gnu.org wrote:
I think that's gone at the moment. Your options are to add it to glibc
(harder) or to add it the gnulib manual. The latter not optimal, but it
could be good enough for now.
I have made enquiries to glibc, but as you say, it's
On 07/30/2010 01:09 PM, Reuben Thomas wrote:
On 30 July 2010 12:02, Paolo Bonzinibonz...@gnu.org wrote:
I think that's gone at the moment. Your options are to add it to glibc
(harder) or to add it the gnulib manual. The latter not optimal, but it
could be good enough for now.
I have made
On 07/30/2010 12:21 AM, Rainer Tammer wrote:
configure:25948: checking whether strtod obeys C99
configure:26060: cc -qlanglvl=extc89 -qlanglvl=extc99 -o conftest -g
conftest.c 5
configure:26060: $? = 0
configure:26060: ./conftest
configure:26060: $? = 3
configure: program exited with
On 07/29/2010 05:52 PM, Paolo Bonzini wrote:
echo 'regexp(,^)' | src/m4
should output 0, not complain about memory failure.
Ah I missed that upthread. Problem is, there's no other patch between
February 3 and today.
On 07/29/2010 05:03 AM, Rainer Tammer wrote:
Hello,
I now have tested M4 1.4.14 on AIX 7.1BETA. I have replaced
build-aux/config.guess with the new AIX 7.1 aware version.
Now fixed in gnulib, and therefore in my current m4 snapshot:
Grab m4-1.4.14.6-76d1 from
On 07/30/2010 04:17 PM, Eric Blake wrote:
7873eb96dregex: avoid new dead-code warning with gcc-4.6.0
It's really dead.
5a4965c01regex: Don't require alloca.
This one looks more interesting on the surface. However, config.log
shows HAVE_ALLOCA and HAVE_ALLOCA_H, so it doesn't
Hello Rainer,
Thanks for the logs. This is exactly what I needed in order to get an
overview of the problems on this platforms.
There are a number of issues. I'll handle them, one per mail.
Bruno
The logs show:
checking whether chown honors trailing slash... no
gl_cv_func_chown_slash_works=${gl_cv_func_chown_slash_works=no}
I'm updating the documentation.
2010-07-30 Bruno Haible br...@clisp.org
chown: Update doc regarding AIX.
* doc/posix-functions/chown.texi:
On 30 July 2010 13:06, Paolo Bonzini bonz...@gnu.org wrote:
A chapter maybe. But Bruno is the doc guru.
Thanks. While waiting for Bruno to say what he'd like, one more
question: in regex.h there's the documentation:
/* If this bit is set, then anchors do not match at newline
characters in
On 07/30/2010 08:35 PM, Reuben Thomas wrote:
So am I right in concluding that the comment for REG_NEWLINE in
regex.h is the wrong way around?
Yes.
Paolo
The logs show:
checking whether fopen recognizes a trailing slash... no
checking whether open recognizes a trailing slash... no
gl_cv_func_fopen_slash=${gl_cv_func_fopen_slash=no}
gl_cv_func_open_slash=${gl_cv_func_open_slash=no}
I'm updating the doc and cross-compilation guesses:
2010-07-30
The logs say:
checking whether frexpl works... yes
gl_cv_func_frexpl_works=${gl_cv_func_frexpl_works=yes}
2010-07-30 Bruno Haible br...@clisp.org
frexpl: Update regarding AIX.
* m4/frexpl.m4 (gl_FUNC_FREXPL_WORKS): When cross-compiling, guess yes
on AIX 7.
The logs say:
checking whether ldexpl works... yes
gl_cv_func_ldexpl_works=${gl_cv_func_ldexpl_works=yes}
2010-07-30 Bruno Haible br...@clisp.org
ldexpl; Update regarding AIX.
* m4/ldexpl.m4 (gl_FUNC_LDEXPL_WORKS): When cross-compiling, guess yes
on AIX 7.
On 29.07.2010 18:56, Eric Blake wrote:
[again, please keep the lists in the loop; and splitting my reply to
multiple topics]
Sorry, sometimes I forget to use reply all instead of reply...
On 07/29/2010 10:46 AM, Rainer Tammer wrote:
Hello Eric,
source: m4-1.4.14.6-ffa0
1. The test-strtod
Hello,
sorry I have to correct this post in respect to test test-strtod.c part...
On 29.07.2010 18:56, Eric Blake wrote:
[again, please keep the lists in the loop; and splitting my reply to
multiple topics]
Sorry, sometimes I forget to use reply all instead of reply...
On 07/29/2010 10:46
Hello,
just to be sure I checked the glib from M4 1.4.14 again. All test were
successful:
# gmake -k -C checks check
gmake: Entering directory `/daten/source/m4-1.4.14/checks'
PATH=`pwd`/../src:$PATH; export PATH; \
./check-them -I ./../examples ./*[0-9][0-9][0-9].* ./stackovf.test
m4 (GNU M4)
The logs say:
configure:14572: checking whether getopt is POSIX compatible
configure:14680: xlc -o conftest -g -D_THREAD_SAFE conftest.c 5
configure:14680: $? = 0
configure:14680: ./conftest
configure:14680: $? = 16
configure: program exited with status 16
On 30 July 2010 19:37, Paolo Bonzini bonz...@gnu.org wrote:
On 07/30/2010 08:35 PM, Reuben Thomas wrote:
So am I right in concluding that the comment for REG_NEWLINE in
regex.h is the wrong way around?
Yes.
Thanks, I will update the patch for the bug I submitted to glibc earlier.
--
The logs say:
checking for working iconv... yes
am_cv_func_iconv_works=${am_cv_func_iconv_works=yes}
But they also say
checking whether iconv supports conversion between UTF-8 and
UTF-{16,32}{BE,LE}... no
so the iconv() being used must be the system iconv(), not GNU libiconv.
2010-07-30
The logs say:
gl_cv_func_printf_directive_a=${gl_cv_func_printf_directive_a=no}
gl_cv_func_printf_directive_f=${gl_cv_func_printf_directive_f=yes}
gl_cv_func_printf_directive_ls=${gl_cv_func_printf_directive_ls=yes}
gl_cv_func_printf_directive_n=${gl_cv_func_printf_directive_n=yes}
The logs say:
configure:44820: checking whether rename honors trailing slash on destination
configure:44854: xlc -o conftest -g -D_THREAD_SAFE conftest.c 5
configure:44854: $? = 0
configure:44854: ./conftest
configure:44854: $? = 1
configure: program exited with status 1
The logs show:
checking whether truncl works... no
configure:54938: checking whether truncl works
configure:54962: xlc -o conftest -g -D_THREAD_SAFE conftest.c 5
ld: 0711-317 ERROR: Undefined symbol: .trunc
So here it's the configure test which is broken: it does not use -lm.
This should fix
Hello,
Bruno Haible wrote:
Hello Rainer,
Thanks for the logs. This is exactly what I needed in order to get an
overview of the problems on this platforms.
no problem, thanks very much for your help. I'll will run the same tests
on AIX 6.1 / 5.3.3.
Q: Shouldn't configure automatically
Hello,
Bruno Haible wrote:
The logs say:
checking for working iconv... yes
am_cv_func_iconv_works=${am_cv_func_iconv_works=yes}
But they also say
checking whether iconv supports conversion between UTF-8 and
UTF-{16,32}{BE,LE}... no
so the iconv() being used must be the system
The logs say:
configure:49890: checking whether stat handles trailing slashes on files
configure:49920: xlc -o conftest -g -D_THREAD_SAFE conftest.c 5
configure:49920: $? = 0
configure:49920: ./conftest
configure:49920: $? = 1
configure: program exited with status 1
The logs say:
checking for working strndup... yes
gl_cv_func_strndup_works=${gl_cv_func_strndup_works=yes}
2010-07-30 Bruno Haible br...@clisp.org
strndup: Update regarding AIX.
* m4/strndup.m4 (gl_FUNC_STRNDUP): When cross-compiling, guess yes on
AIX 7.
The logs say:
configure:53029: checking whether symlink handles trailing slash correctly
configure:53051: xlc -o conftest -g -D_THREAD_SAFE conftest.c 5
configure:53051: $? = 0
configure:53051: ./conftest
configure:53051: $? = 1
configure: program exited with status 1
The logs say:
configure:12653: checking whether unlink honors trailing slashes
configure:12683: xlc -o conftest -g -D_THREAD_SAFE conftest.c 5
configure:12683: $? = 0
configure:12683: ./conftest
configure:12683: $? = 1
configure: program exited with status 1
The logs show that the autoconf test for strtok_r fails in an unexpected way:
checking whether strtok_r works... no
conftest.c, line 745.2: 1506-766 (S) The universal character name � is not
in the allowable range for an identifier.
conftest.c, line 745.5: 1506-766 (S) The universal character
The logs say:
checking whether wcwidth works reasonably in UTF-8 locales... yes
This means that the wcwidth replacement is not necessary.
2010-07-30 Bruno Haible br...@clisp.org
wcwidth: Drop replacement on AIX 7.
* m4/wcwidth.m4 (gl_FUNC_WCWIDTH): When cross-compiling,
Hi Eric,
Rainer Tammer's logs say:
checking whether futimens works... no
configure:30237: $? = 2
configure: program exited with status 2
Exit code 2 from this test program comes from the line
if (futimens (AT_FDCWD, NULL)) return 2;
What is this code trying to do? POSIX [1] says that
On 07/30/2010 01:45 PM, Bruno Haible wrote:
Hi Eric,
Rainer Tammer's logs say:
checking whether futimens works... no
configure:30237: $? = 2
configure: program exited with status 2
Exit code 2 from this test program comes from the line
if (futimens (AT_FDCWD, NULL)) return
The log says this:
checking whether getline is declared... no
checking for getline... yes
I'm updating the doc:
2010-07-30 Bruno Haible br...@clisp.org
getline: Update regarding AIX.
* doc/posix-functions/getline.texi: Mention bug on AIX 7.1.
Reported by Rainer
These errors all occur in regex 0.12, i.e. have been in regex.h for at
least 18 years! That's rather sad. Hopefully I can prevent their
lasting much longer...
Hello Bruno,
Bruno Haible wrote:
Hello Rainer,
Regarding this failure:
test-readlink.h:54: assertion failed
FAIL: test-readlink
Could you please show what is the value of errno between these ASSERT
statements:
ASSERT (func (BASE file/, buf, sizeof buf) == -1);
printf (errno =
On 07/30/2010 09:58 AM, Rainer Tammer wrote:
Not quite... but it's getting better...
Test 0xg case is fixed.
Thanks for the positive feedback - we're getting closer.
# ./test-strtod
test-strtod.c:435: assertion failed
{
const char input[] = 0xp;
char *ptr;
double
Hello Rainer,
Rainer Tammer wrote:
1. libm
Some tests need libm (sample):
xlc -g -o test-asin test-asin.o ../gllib/libgnu.a
ld: 0711-317 ERROR: Undefined symbol: .asin
ld: 0711-317 ERROR: Undefined symbol: asin
This is caused by an assumption that gnulib/m4/mathfunc.m4 makes,
On 07/30/2010 03:04 PM, Rainer Tammer wrote:
Hello Bruno,
Bruno Haible wrote:
Hello Rainer,
Regarding this failure:
test-readlink.h:54: assertion failed
FAIL: test-readlink
Could you please show what is the value of errno between these ASSERT
statements:
ASSERT (func (BASE
Hello Eric,
Eric Blake wrote:
... cut for clarity ...
test-strtod.c:762: assertion failed
{
const char input[] = 0x1p;
char *ptr;
double result;
errno = 0;
result = strtod (input, ptr);
ASSERT (result == 1.0);
ASSERT (ptr == input + 3); - fail
Hi Paolo,
The poll test fails like this on AIX 7.1:
Pipe test... failed (expecting POLLIN | POLLRDNORM after writing)
failed (expecting POLLIN after writing)
failed (expecting POLLRDNORM after writing)
failed (expecting POLLHUP after shutdown)
FAIL: test-poll
It also fails on
Hello Bruno,
Bruno Haible wrote:
Hello Rainer,
Rainer Tammer wrote:
1. libm
Some tests need libm (sample):
xlc -g -o test-asin test-asin.o ../gllib/libgnu.a
ld: 0711-317 ERROR: Undefined symbol: .asin
ld: 0711-317 ERROR: Undefined symbol: asin
This is caused by an
Hello,
just one addition, this is the contents of AIX 7.1BETA libm.a
libm.a:
***Archive Symbol Table***
OffsetName
128._classl128
128_classl128
2356._isfinitel128
2356_isfinitel128
3812._isinfl128
3812_isinfl128
2. dirfd
xlc -g -o test-canonicalize-lgpl test-canonicalize-lgpl.o
../gllib/libgnu.a
ld: 0711-317 ERROR: Undefined symbol: .dirfd
The logs say:
checking for dirfd... no
checking whether dirfd is declared... (cached) yes
checking whether dirfd is a macro... no
The
* lib/strtod.c (strtod): Work around AIX bug of parsing p with no
exponent.
* tests/test-strtod.c (main): Enhance tests.
* doc/posix-functions/strtod.texi (strtod): Document next bug.
Signed-off-by: Eric Blake ebl...@redhat.com
---
I think this should solve the remaining strtod bugs. I also
Sigh. I've been picking my way through this paragraph, looking at the code:
/* This data structure represents a compiled pattern. Before calling
the pattern compiler, the fields `buffer', `allocated', `fastmap',
`translate', and `no_sub' can be set. After the pattern has been
compiled,
deprecate the manual setting of those particular fields,
As I recall, Kathy and I wrote all that based on actual usage (Emacs was
the main thing we looked at). I am not in favor of deprecating
anything that has been around so long, precisely because it has been
around so long. What is
On 07/30/2010 04:09 PM, Reuben Thomas wrote:
Sigh. I've been picking my way through this paragraph, looking at the code:
/* This data structure represents a compiled pattern. Before calling
the pattern compiler, the fields `buffer', `allocated', `fastmap',
`translate', and `no_sub'
On 30 July 2010 23:23, Karl Berry k...@freefriends.org wrote:
deprecate the manual setting of those particular fields,
As I recall, Kathy and I wrote all that based on actual usage (Emacs was
the main thing we looked at). I am not in favor of deprecating
anything that has been around so
On 30 July 2010 23:33, Eric Blake ebl...@redhat.com wrote:
and `no_sub' can't be set (because re_compile
always overwrites it, as it does newline_anchor).
Did you compile a pattern with grouping ()? I'm not sure, but the
behavior on no_sub may be conditional on whether there are any
Hello Rainer,
The logs say:
test-iconv.c:81: assertion failed
FAIL: test-iconv
Can you please compile and run the following program, and show the output?
===
#include iconv.h
#include errno.h
#include stdio.h
int
On 07/30/2010 04:45 PM, Reuben Thomas wrote:
On 30 July 2010 23:33, Eric Blake ebl...@redhat.com wrote:
and `no_sub' can't be set (because re_compile
always overwrites it, as it does newline_anchor).
Did you compile a pattern with grouping ()? I'm not sure, but the
behavior on no_sub may
Hello Rainer,
Regarding this test failure:
test-link.h:117: assertion failed
FAIL: test-linkat
What's the output when you change test-link.h like this at line 116:
ASSERT (func (BASE a/, BASE c) == -1);
{ int err = errno; fprintf (stderr, errno = %d\n, err); errno = err; perror
(); }
Hi Rainer,
Regarding this test failure:
test-chown.h:88: assertion failed
FAIL: test-fchownat
What's the output if you change test-chown.h at line 87 like this:
ASSERT (func (BASE dir/file/, -1, -1) == -1);
{ int err = errno; fprintf (stderr, errno = %d\n, err); errno = err; perror
Rainer, Eric,
I leave these investigations to you:
test-utimens.h:85: assertion failed
FAIL: test-utimens
test-utimens.h:85: assertion failed
FAIL: test-utimensat
Bruno
Hello Rainer,
Regarding this test failure:
test-duplocale.c:70: assertion failed
FAIL: test-duplocale
What does the following test program print?
===
#define _GNU_SOURCE
#define _ALL_SOURCE
#include locale.h
Other than that:
About fdopendir, the logs say:
checking for fdopendir... no
checking whether fdopendir is declared without a macro... yes
checking whether fdopendir is declared... (cached) yes
So apparently they declare the function but don't implement it.
Fortunately, gnulib does the
On 07/30/2010 05:00 PM, Bruno Haible wrote:
Rainer, Eric,
I leave these investigations to you:
test-utimens.h:85: assertion failed
FAIL: test-utimens
Both probably another case of an unexpected errno value. But which?
--
Eric Blake ebl...@redhat.com+1-801-349-2682
Libvirt
On 07/30/2010 05:11 PM, Bruno Haible wrote:
Other than that:
About mknod, the logs say:
configure:40991: checking whether mknod can create fifo without root
privileges
configure:41015: xlc -o conftest -g -D_THREAD_SAFE conftest.c 5
configure:41015: $? = 0
configure:41015:
* m4/mknod.m4 (gl_FUNC_MKNOD): Make it harder to run configure as
root, since that is just asking for problems.
Based on a report by Bruno Haible.
Signed-off-by: Eric Blake ebl...@redhat.com
---
What do you think of this?
ChangeLog |7 +++
m4/mknod.m4 | 11 ---
2 files
Hi,
On the bison mailing lists, it was recently suggested that bison should
include gettext in bootstrap-tools in cfg.mk. That sounds reasonable to
me, but I see that coreutils and m4 don't do this. Is there any general
rule of thumb about which tools are significant enough to be listed?
66 matches
Mail list logo
