GNU bug report logs -
#17598
Emacs pretest 24.3.91 - AIX - SIGSEV when customize
Previous Next
Reported by: Gilles Pion <gpion <at> lfdj.com>
Date: Mon, 26 May 2014 06:54:02 UTC
Severity: important
Done: Paul Eggert <eggert <at> cs.ucla.edu>
Bug is archived. No further changes may be made.
To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 17598 in the body.
You can then email your comments to 17598 AT debbugs.gnu.org in the normal way.
Toggle the display of automated, internal messages from the tracker.
Report forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Mon, 26 May 2014 06:54:02 GMT)
Full text and
rfc822 format available.
Acknowledgement sent
to
Gilles Pion <gpion <at> lfdj.com>:
New bug report received and forwarded. Copy sent to
bug-gnu-emacs <at> gnu.org.
(Mon, 26 May 2014 06:54:03 GMT)
Full text and
rfc822 format available.
Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
sequences of actions making emacs crash:
* lauch "emacs --no-site-file --no-init-file" with $DISPLAY pointing
to a running X server.
* M-X customize-group
* icomplete
* expand "icomplete mode" node (should be initially "off")
* toggle
* goto "state"
* press "return"
Note: this sequence of actions always leads crash but stack
backtraces are never the same.
==> config.log and gdb session typescript attached
--
Gilles PION
[anonymized-gdb_session.txt (text/plain, attachment)]
[anonymized-config.log (application/octet-stream, attachment)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Mon, 26 May 2014 12:31:01 GMT)
Full text and
rfc822 format available.
Message #8 received at 17598 <at> debbugs.gnu.org (full text, mbox):
Just noticed that when using "temacs" instead of "emacs" (which the
same sequence of actions) everything runs fine.
--
Gilles PION
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Mon, 26 May 2014 18:25:03 GMT)
Full text and
rfc822 format available.
Message #11 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
I can't easily reproduce the problem since I don't use X. Perhaps the
problem is due to cleanup changes in src/unexaix.c. To test this
hypothesis, can you please try the attached file in place of the 24.3.91
src/unexaix.c? The attached file is as close to 24.3's unexaix.c as
possible; the files can't be identical due to other changes in Emacs.
Thanks.
[unexaix.c (text/x-csrc, attachment)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Tue, 27 May 2014 07:36:02 GMT)
Full text and
rfc822 format available.
Message #14 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
It worked!
That's a miracle: being able to fix a such obscure crash with so few
information although not being able to reproduce on you own configuration
is amazing.
I will keep on testing anyway
Thanks,
--
Gilles PION
Ingénieur système
Industrialisation des systèmes, réseaux et middleware
Tél. (direct):
+33 (0) 4 42 78 56 59
De : Paul Eggert <eggert <at> cs.ucla.edu>
A : Gilles Pion <gpion <at> lfdj.com>,
Cc : 17598 <at> debbugs.gnu.org
Date : 26/05/2014 20:24
Objet : Re: Emacs pretest 24.3.91 - AIX - SIGSEV when customize
I can't easily reproduce the problem since I don't use X. Perhaps the
problem is due to cleanup changes in src/unexaix.c. To test this
hypothesis, can you please try the attached file in place of the 24.3.91
src/unexaix.c? The attached file is as close to 24.3's unexaix.c as
possible; the files can't be identical due to other changes in Emacs.
Thanks.
[pièce jointe "unexaix.c" supprimée par Gilles Pion/DPR/SI/FDJ]
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
[Message part 3 (image/gif, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Tue, 27 May 2014 14:59:02 GMT)
Full text and
rfc822 format available.
Message #17 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
Still crashing (but not always at the same moment)
M-X customize-group
then some stuff, sometime crash immediately, sometime after some cursor
movement/
gdb session attached
--
Gilles PION
[Message part 2 (text/html, inline)]
[anonymized-gdb_session_2.txt (text/plain, attachment)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Tue, 27 May 2014 15:39:02 GMT)
Full text and
rfc822 format available.
Message #20 received at 17598 <at> debbugs.gnu.org (full text, mbox):
If it's still crashing, that suggests the problem is not in unexaix.c
after all, unfortunately.
What happens if you run "./configure --enable-checking=all && make"?
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 28 May 2014 07:49:02 GMT)
Full text and
rfc822 format available.
Message #23 received at 17598 <at> debbugs.gnu.org (full text, mbox):
On 05/27/2014 07:38 PM, Paul Eggert wrote:
> If it's still crashing, that suggests the problem is not in unexaix.c after all, unfortunately.
Is it possible to run undumped (i.e. temacs) on AIX? If there is an issue with
unexec, it shouldn't affect undumped executable.
> What happens if you run "./configure --enable-checking=all && make"?
Latest backtrace (with SIGSEGV in malloc) is probably caused by heap corruption.
Are there some built-in debugging facilities in AIX malloc implementation
(something similar to glibc mcheck, for example)?
Dmitry
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 28 May 2014 08:41:03 GMT)
Full text and
rfc822 format available.
Message #26 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
> De : Dmitry Antipov <dmantipov <at> yandex.ru>
> A : Gilles PION <gpion <at> lfdj.com>,
> Cc : Paul Eggert <eggert <at> cs.ucla.edu>, 17598 <at> debbugs.gnu.org
> Date : 28/05/2014 09:48
> Objet : Re: bug#17598: Emacs pretest 24.3.91 - AIX - SIGSEV when
customize
>
> On 05/27/2014 07:38 PM, Paul Eggert wrote:
>
> > If it's still crashing, that suggests the problem is not in
> unexaix.c after all, unfortunately.
>
> Is it possible to run undumped (i.e. temacs) on AIX? If there is an
issue with
> unexec, it shouldn't affect undumped executable.
Yes, it crashes too:
$ gdb ./emacs-24.3.91/src/temacs
/../
(gdb) run --no-site-file --no-init-file
/../
Pure-hashed: 23622 strings, 3448 vectors, 37051 conses, 3241 bytecodes, 82
others
# emacs windows appears here , doing M-x configure then moving cursor, as
usual
Fatal error 6: Resource lost
Program received signal SIGABRT, Aborted.
[Switching to Thread 1]
0xd0508b20 in pthread_kill () from /usr/lib/libpthread.a(shr_xpg5.o)
So you're right, the issue is not unexec related
>
> > What happens if you run "./configure --enable-checking=all && make"?
>
> Latest backtrace (with SIGSEGV in malloc) is probably caused by heap
> corruption.
> Are there some built-in debugging facilities in AIX malloc
implementation
> (something similar to glibc mcheck, for example)?
The previous run uses "temacs" from a new build using those suggested
"configure" parameters,
backtrace:
(gdb) where
#0 0xd0508b20 in pthread_kill () from /usr/lib/libpthread.a(shr_xpg5.o)
#1 0xd0507f8c in _p_raise () from /usr/lib/libpthread.a(shr_xpg5.o)
#2 0x1000c4f8 in terminate_due_to_signal ()
#3 0x1003483c in emacs_abort ()
#4 0x10020a04 in xfree ()
#5 0x1029c5ec in x_uncatch_errors ()
#6 0x10288fe8 in xfont_list_pattern ()
#7 0x1028a214 in xfont_list ()
#8 0x10271940 in font_list_entities ()
#9 0x1027480c in font_find_for_lface.constprop.12 ()
#10 0x10274f48 in font_load_for_lface ()
#11 0x1025a8d0 in realize_face ()
#12 0x1025cd80 in lookup_face.isra.22 ()
#13 0x10260b3c in face_at_buffer_position ()
#14 0x100c1230 in handle_face_prop ()
#15 0x100a89c0 in handle_stop ()
#16 0x100c1d64 in next_element_from_buffer ()
#17 0x100acf80 in get_next_display_element ()
#18 0x100bd448 in display_line ()
#19 0x100ccba0 in try_window.constprop.53 ()
#20 0x100e8f28 in redisplay_window ()
#21 0x100ebfbc in redisplay_window_0 ()
#22 0x10053f4c in internal_condition_case_1 ()
#23 0x1008a4b4 in redisplay_windows ()
#24 0x100d08e0 in redisplay_internal ()
#25 0x100497c8 in read_char ()
#26 0x1004ba00 in read_key_sequence.constprop.46 ()
#27 0x1004e354 in command_loop_1 ()
#28 0x10053d94 in internal_condition_case ()
#29 0x10036c00 in command_loop_2 ()
#30 0x10053c3c in internal_catch ()
#31 0x1003e414 in recursive_edit_1 ()
#32 0x1003e630 in Frecursive_edit ()
#33 0x10000ee4 in main ()
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 28 May 2014 09:46:01 GMT)
Full text and
rfc822 format available.
Message #29 received at 17598 <at> debbugs.gnu.org (full text, mbox):
On 05/28/2014 12:40 PM, Gilles PION wrote:
> So you're right, the issue is not unexec related
So you have a few irregular crashes, more or less related to memory allocation.
If this is applicable to your AIX version, can you try some recipes from the
following article:
http://www.ibm.com/developerworks/aix/library/au-mallocdebug.html
> > > What happens if you run "./configure --enable-checking=all && make"?
[skip]
> The previous run uses "temacs" from a new build using those suggested "configure" parameters,
These parameters assumes '-O2 -g3' or at least '-g', so your gdb backtrace should
include file names, line numbers and (some) function parameters. But your backtrace
looks like from stripped binary or the binary compiled without debug info.
> backtrace:
> (gdb) where
> #0 0xd0508b20 in pthread_kill () from /usr/lib/libpthread.a(shr_xpg5.o)
> #1 0xd0507f8c in _p_raise () from /usr/lib/libpthread.a(shr_xpg5.o)
> #2 0x1000c4f8 in terminate_due_to_signal ()
> #3 0x1003483c in emacs_abort ()
> #4 0x10020a04 in xfree ()
How the emacs_abort() was called here? AFAICS xfree() doesn't call it directly.
Dmitry
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 28 May 2014 09:59:02 GMT)
Full text and
rfc822 format available.
Message #32 received at 17598 <at> debbugs.gnu.org (full text, mbox):
Dmitry Antipov <dmantipov <at> yandex.ru> writes:
>> (gdb) where
>> #0 0xd0508b20 in pthread_kill () from /usr/lib/libpthread.a(shr_xpg5.o)
>> #1 0xd0507f8c in _p_raise () from /usr/lib/libpthread.a(shr_xpg5.o)
>> #2 0x1000c4f8 in terminate_due_to_signal ()
>> #3 0x1003483c in emacs_abort ()
>> #4 0x10020a04 in xfree ()
>
> How the emacs_abort() was called here? AFAICS xfree() doesn't call it directly.
But via unblock_input.
Andreas.
--
Andreas Schwab, SUSE Labs, schwab <at> suse.de
GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE 1748 E4D4 88E3 0EEA B9D7
"And now for something completely different."
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 28 May 2014 12:04:02 GMT)
Full text and
rfc822 format available.
Message #35 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
> These parameters assumes '-O2 -g3' or at least '-g', so your gdb
> backtrace should
> include file names, line numbers and (some) function parameters. But
> your backtrace
> looks like from stripped binary or the binary compiled without debug
info.
You're right, I was using "CFLAGS=-O3",
Now rebuild with "CFLAGS="-O3 -g3" here's the backtrace:
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 1]
0xd013d94c in rightmost () from /usr/lib/libc.a(shr.o)
(gdb) where
#0 0xd013d94c in rightmost () from /usr/lib/libc.a(shr.o)
#1 0xd013eb28 in free_y () from /usr/lib/libc.a(shr.o)
#2 0xd0129a74 in free_common () from /usr/lib/libc.a(shr.o)
#3 0xd11ff294 in XFreeFontNames <at> AF6_3 () from /usr/lib/libX11.a(shr4.o)
#4 0x1028a04c in xfont_list_pattern (display=0x20837f98,
pattern=0x2ff1da48 "-misc-fixed-*-*-*--*-*-*-*-*-*-iso8859-1",
registry=542476962, script=1210196514) at xfont.c:142
#5 0x1028aef4 in xfont_list (f=0x1, spec=542396093) at xfont.c:516
#6 0x102725c0 in font_list_entities (f=0x20ce1e88, spec=552058821)
at font.c:2758
#7 0x1027548c in font_find_for_lface.constprop.12 (f=0x20ce1e88,
attrs=0x2ff1e058, spec=5) at font.c:129
#8 0x10275bc8 in font_load_for_lface (f=0x20ce1e88, attrs=0x2ff1e058,
spec=552075205) at font.c:3302
#9 0x1025b508 in realize_face (cache=0x20ae2c18, attrs=0x2ff1e058,
former_face_id=542425624) at xfaces.c:5568
#10 0x1025d9b8 in lookup_face.isra.22 (ISRA.742=0x20ce1ee0,
attr=0x2ff1e058)
at xfaces.c:4453
#11 0x10261774 in face_at_buffer_position (w=0x0, pos=1,
endptr=0x2ff1e128,
limit=542062760, mouse=560775400, base_face_id=0) at xfaces.c:6057
#12 0x100c1928 in handle_face_prop (it=0x2ff1f4f8) at xdisp.c:3873
#13 0x100a90b8 in handle_stop (it=0x2ff1f4f8) at xdisp.c:3377
#14 0x100c245c in next_element_from_buffer (it=0x2ff1f4f8) at xdisp.c:8175
#15 0x100ad668 in get_next_display_element (it=0x2ff1f4f8) at xdisp.c:6807
#16 0x100bdb30 in display_line (it=0x2ff1f4f8) at xdisp.c:19902
#17 0x100cd2a8 in try_window.constprop.53 (window=1) at xdisp.c:16719
#18 0x100e9650 in redisplay_window (window=550387341,
just_this_one_p=false)
at xdisp.c:16198
#19 0x100ec6e4 in redisplay_window_0 (window=1) at xdisp.c:14191
#20 0x10054464 in internal_condition_case_1 (
bfun=@0x202233e4: 0x100ec6a0 <redisplay_window_0>, arg=550387341,
handlers=-1431589715,
hfun=@0x202233f0: 0x100838f0 <redisplay_window_error>) at eval.c:1378
#21 0x1008ab3c in redisplay_windows (window=-260271624) at xdisp.c:14171
#22 0x100d0fe8 in redisplay_internal () at xdisp.c:13770
#23 0x10049cd0 in read_char (commandflag=1, map=562883446,
prev_event=542216242, used_mouse_menu=0x2ff225bc, end_time=0x0)
at keyboard.c:2570
#24 0x1004bf08 in read_key_sequence.constprop.46 (keybuf=0x2ff22698,
prompt=542216242, dont_downcase_last=false,
can_return_switch_frame=true,
fix_current_buffer=true, prevent_redisplay=false) at keyboard.c:9087
#25 0x1004e85c in command_loop_1 () at keyboard.c:1452
#26 0x100542ac in internal_condition_case (
bfun=@0x2021eab8: 0x1004e568 <command_loop_1>, handlers=-260271624,
hfun=@0x2021eac4: 0x1003c678 <cmd_error>) at eval.c:1354
#27 0x100370e8 in command_loop_2 (ignore=1) at keyboard.c:1177
#28 0x10054154 in internal_catch (tag=542282210,
func=@0x202240bc: 0x10037098 <command_loop_2>, arg=542216242)
at eval.c:1118
#29 0x1003e8ec in recursive_edit_1 () at keyboard.c:1156
#30 0x1003eb08 in Frecursive_edit () at keyboard.c:848
#31 0x10000ee4 in main (argc=542368394, argv=0x2ff22a8c) at emacs.c:1646
(gdb)
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 28 May 2014 12:14:01 GMT)
Full text and
rfc822 format available.
Message #38 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
> So you have a few irregular crashes, more or less related to memory
> allocation.
> If this is applicable to your AIX version, can you try some recipes from
the
> following article:
>
> http://www.ibm.com/developerworks/aix/library/au-mallocdebug.html
Ouch!
$ MALLOCTYPE=debug MALLOEBUG=validate_ptrs ./emacs-24.3.91/src/temacs
--no-site-file --no-init-file
Fatal error 11: Segmentation violationMemory fault
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 28 May 2014 14:09:02 GMT)
Full text and
rfc822 format available.
Message #41 received at 17598 <at> debbugs.gnu.org (full text, mbox):
On 05/28/2014 04:13 PM, Gilles PION wrote:
> $ MALLOCTYPE=debug MALLOEBUG=validate_ptrs ./emacs-24.3.91/src/temacs --no-site-file --no-init-file
> Fatal error 11: Segmentation violationMemory fault
Are there some more specific messages about memory-related errors? Most malloc
debuggers can handle typical errors like double free or writing before begin /
past end of block. It's almost impossible to understand what's wrong with
#0 0xd013d94c in rightmost () from /usr/lib/libc.a(shr.o)
#1 0xd013eb28 in free_y () from /usr/lib/libc.a(shr.o)
#2 0xd0129a74 in free_common () from /usr/lib/libc.a(shr.o)
#3 0xd11ff294 in XFreeFontNames <at> AF6_3 () from /usr/lib/libX11.a(shr4.o)
without a solid cooperation from the malloc debugging facilities.
And, as usual with weird problems, please try '-O0 -g3' (i.e. avoid any optimizations).
Dmitry
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 28 May 2014 14:28:02 GMT)
Full text and
rfc822 format available.
Message #44 received at 17598 <at> debbugs.gnu.org (full text, mbox):
Gilles PION wrote:
> $ MALLOCTYPE=debug MALLOEBUG=validate_ptrs ./emacs-24.3.91/src/temacs
> --no-site-file --no-init-file
> Fatal error 11: Segmentation violationMemory fault
>
Have you gone back to the 24.3.91 unexaix.c (with the short fix I sent
you), or are you using the much bigger change I sent you later? For now
I'll assume the former. Also, I assume you're still configuring with
--enable-checking=all.
Unfortunately AIX uses the relocating allocator in ralloc.c, not the
system malloc, so MALLOCTYPE and MALLOCDEBUG (misspelled above) aren't
likely to be helpful. If you want to try the system malloc, please
comment out "#define REL_ALLOC 1" in src/config.h, change "#undef
SYSTEM_MALLOC" to "#define SYSTEM_MALLOC 1" in the same file, and
comment out "RALLOC_OBJ=ralloc.o" in src/Makefile (though now of course
we're debugging a different program....).
Also, did you actually use "MALLOEBUG"? That looks like a typo.
Once you've fixed the above, how about this?
MALLOCTYPE=debug
MALLOCDEBUG=postfree_checking,validate_ptrs,stack_depth:1000 ./temacs -Q
or this?
gdb temacs
(gdb) set env MALLOCTYPE debug
(gdb) set env MALLOCDEBUG postfree_checking,validate_ptrs,stack_depth:200
(gdb) run -Q
Please also try with the "-nw" option, e.g., "./temacs -Q -nw" or "run
-Q -nw".
A couple of issues that may or may not be related. On my AIX 7 host I
have to configure --with-jpeg=no --with-png=no --with-gif=no
--with-tiff=no because those libraries are not installed. On my
platform pkg-config dumps core (don't know why) so I'm not using any of
the fancy libraries you're probably using. This (along with the fact
that I can't run X) makes it hard for me to reproduce the problem.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 28 May 2014 16:24:02 GMT)
Full text and
rfc822 format available.
Message #47 received at 17598 <at> debbugs.gnu.org (full text, mbox):
> Date: Wed, 28 May 2014 07:26:49 -0700
> From: Paul Eggert <eggert <at> cs.ucla.edu>
> Cc: 17598 <at> debbugs.gnu.org
>
> Unfortunately AIX uses the relocating allocator in ralloc.c, not the
> system malloc
Does it? The backtrace Gilles posted, viz.:
> (gdb) where
> #0 0xd013d94c in rightmost () from /usr/lib/libc.a(shr.o)
> #1 0xd013eb28 in free_y () from /usr/lib/libc.a(shr.o)
> #2 0xd0129a74 in free_common () from /usr/lib/libc.a(shr.o)
> #3 0xd11ff294 in XFreeFontNames <at> AF6_3 () from /usr/lib/libX11.a(shr4.o)
> #4 0x1028a04c in xfont_list_pattern (display=0x20837f98,
> pattern=0x2ff1da48 "-misc-fixed-*-*-*--*-*-*-*-*-*-iso8859-1",
> registry=542476962, script=1210196514) at xfont.c:142
doesn't seem to have gmalloc and ralloc functions anywhere in sight,
it looks like a call to 'free' in the system libc to me. Am I missing
something?
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 28 May 2014 16:35:02 GMT)
Full text and
rfc822 format available.
Message #50 received at 17598 <at> debbugs.gnu.org (full text, mbox):
Eli Zaretskii wrote:
>> >Unfortunately AIX uses the relocating allocator in ralloc.c, not the
>> >system malloc
> Does it?
It does in the AIX 7 version that I built, yes.
> it looks like a call to 'free' in the system libc to me. Am I missing something?
It's linker magic. Both the system malloc and the ralloc.c-supplied
malloc are available, and both are used: Emacs uses the ralloc version
and library functions like XFreeFontNames use the system malloc.
This mixture of system and ralloc-supplied malloc might be causing
Gilles's problem, which is partly why I suggested that he build with the
system malloc only.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 28 May 2014 16:41:02 GMT)
Full text and
rfc822 format available.
Message #53 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
> > $ MALLOCTYPE=debug MALLOEBUG=validate_ptrs ./emacs-24.3.91/src/temacs
> > --no-site-file --no-init-file
> > Fatal error 11: Segmentation violationMemory fault
> >
>
> Have you gone back to the 24.3.91 unexaix.c (with the short fix I sent
> you), or are you using the much bigger change I sent you later? For now
> I'll assume the former. Also, I assume you're still configuring with
> --enable-checking=all.
I was using the latest unexaix.c, going back to previous one now
>
> Unfortunately AIX uses the relocating allocator in ralloc.c, not the
> system malloc, so MALLOCTYPE and MALLOCDEBUG (misspelled above) aren't
> likely to be helpful. If you want to try the system malloc, please
> comment out "#define REL_ALLOC 1" in src/config.h, change "#undef
> SYSTEM_MALLOC" to "#define SYSTEM_MALLOC 1" in the same file, and
> comment out "RALLOC_OBJ=ralloc.o" in src/Makefile (though now of course
> we're debugging a different program....).
OK, recompiling with those fixes (after configuring using "CFLAGS="-O0
-g3" ):
1rst step: ./configure
2nd step: fix src/config.h as indicated
3rd step make
>
> Also, did you actually use "MALLOEBUG"? That looks like a typo.
You're right, sorry.
>
> Once you've fixed the above, how about this?
>
> MALLOCTYPE=debug
> MALLOCDEBUG=postfree_checking,validate_ptrs,stack_depth:1000 ./temacs -Q
$ MALLOCTYPE=debug
MALLOCDEBUG=postfree_checking,validate_ptrs,stack_depth:1000 ./temacs -Q
Fatal error 11: Segmentation violationMemory fault
>
> or this?
>
> gdb temacs
> (gdb) set env MALLOCTYPE debug
> (gdb) set env MALLOCDEBUG
postfree_checking,validate_ptrs,stack_depth:200
> (gdb) run -Q
(gdb) set env MALLOCTYPE debug
(gdb) set env MALLOCDEBUG postfree_checking,validate_ptrs,stack_depth:200
(gdb) run -Q
Starting program:
/sg/paxdev01/D1stunix/src/emacs/24.3.91/emacs-24.3.91/src/temacs -Q
[New Thread 1]
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 1]
0xd0128c78 in extend_brk () from /usr/lib/libc.a(shr.o)
(gdb) where
#0 0xd0128c78 in extend_brk () from /usr/lib/libc.a(shr.o)
>
> Please also try with the "-nw" option, e.g., "./temacs -Q -nw" or "run
> -Q -nw".
It's the same
Note that the immediate crash occurs only when MALLOC* variables are
defined.
>
> A couple of issues that may or may not be related. On my AIX 7 host I
> have to configure --with-jpeg=no --with-png=no --with-gif=no
> --with-tiff=no because those libraries are not installed. On my
> platform pkg-config dumps core (don't know why) so I'm not using any of
> the fancy libraries you're probably using. This (along with the fact
> that I can't run X) makes it hard for me to reproduce the problem.
I understand, I've also disabled graphic libraries execpt png (using a
emacs dedicated version)
I'm on aix 6.1, since I've been unable to compile on AIX 7.1:
configure: error: Emacs does not support `powerpc-ibm-aix7.1.0.0' systems.
If you think it should, please send a report to bug-gnu-emacs <at> gnu.org.
Check `etc/MACHINES' for recognized configuration names.
Used gcc binary from 6.1 there ^^^
Since I'missing native 7.1 compatible compiler to re-boostrap gcc.
NB: I won't be reachable at my office tomorrow, will return only friday
Next week, won't be very available until Thursday too.
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 28 May 2014 16:44:01 GMT)
Full text and
rfc822 format available.
Message #56 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
> Eli Zaretskii wrote:
> >> >Unfortunately AIX uses the relocating allocator in ralloc.c, not the
> >> >system malloc
> > Does it?
>
> It does in the AIX 7 version that I built, yes.
I'm running on AIX 6.1, "configure" fails on 7.1 for me:
checking build system type... powerpc-ibm-aix7.1.0.0
checking host system type... powerpc-ibm-aix7.1.0.0
configure: error: Emacs does not support `powerpc-ibm-aix7.1.0.0' systems.
If you think it should, please send a report to bug-gnu-emacs <at> gnu.org.
Check `etc/MACHINES' for recognized configuration names.
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 28 May 2014 17:33:02 GMT)
Full text and
rfc822 format available.
Message #59 received at 17598 <at> debbugs.gnu.org (full text, mbox):
On 05/28/2014 08:43 PM, Gilles PION wrote:
> I'm running on AIX 6.1, "configure" fails on 7.1 for me:
>
> checking build system type... powerpc-ibm-aix7.1.0.0
> checking host system type... powerpc-ibm-aix7.1.0.0
> configure: error: Emacs does not support `powerpc-ibm-aix7.1.0.0' systems.
> If you think it should, please send a report to bug-gnu-emacs <at> gnu.org.
> Check `etc/MACHINES' for recognized configuration names.
Hopefully this may be fixed with just:
--- emacs-24.3.91/configure.ac 2014-05-28 12:28:29.000000000 -0500
+++ emacs-24.3.91/configure.ac 2014-05-28 12:28:58.000000000 -0500
@@ -550,7 +550,7 @@
rs6000-ibm-aix[56]* )
opsys=aix4-2
;;
- powerpc-ibm-aix[56]* )
+ powerpc-ibm-aix[5-9]* | powerpc-ibm-aix[1-9][0-9]* )
opsys=aix4-2
;;
Shouldn't we backport this from trunk?
Dmitry
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 28 May 2014 17:47:01 GMT)
Full text and
rfc822 format available.
Message #62 received at 17598 <at> debbugs.gnu.org (full text, mbox):
On 05/28/2014 10:32 AM, Dmitry Antipov wrote:
> Shouldn't we backport this from trunk?
I installed it into the trunk because it's not a regression from 24.1
<http://lists.gnu.org/archive/html/emacs-devel/2014-05/msg00309.html>.
It'd make sense to install it into emacs-24 too, though; I'll CC: this
to Stefan to ask him.
Stefan, is it OK to fix emacs-24's configure.ac to treat AIX 7+ like
earlier AIX releases, as Dmitry suggests?
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 28 May 2014 17:55:02 GMT)
Full text and
rfc822 format available.
Message #65 received at 17598 <at> debbugs.gnu.org (full text, mbox):
On 05/28/2014 09:43 AM, Gilles PION wrote:
> configure: error: Emacs does not support `powerpc-ibm-aix7.1.0.0'
> systems.
One way to work around that bug, without editing 'configure', is to
configure this way:
./configure powerpc-ibm-aix6
This should work even on AIX 7.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 28 May 2014 18:02:02 GMT)
Full text and
rfc822 format available.
Message #68 received at 17598 <at> debbugs.gnu.org (full text, mbox):
On 05/28/2014 09:40 AM, Gilles PION wrote:
>
> OK, recompiling with those fixes (after configuring using "CFLAGS="-O0
> -g3" ):
> 1rst step: ./configure
> 2nd step: fix src/config.h as indicated
> 3rd step make
Did you remember to remove the line "RALLOC_OBJ=ralloc.o" from
src/Makefile? Also, to play it safe, please remove src/ralloc.o, run
'make', and make sure src/ralloc.o is not created.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 28 May 2014 23:45:02 GMT)
Full text and
rfc822 format available.
Message #71 received at 17598 <at> debbugs.gnu.org (full text, mbox):
> Stefan, is it OK to fix emacs-24's configure.ac to treat AIX 7+ like earlier
> AIX releases, as Dmitry suggests?
Yes, that sounds fine,
Stefan
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Thu, 29 May 2014 04:26:02 GMT)
Full text and
rfc822 format available.
Message #74 received at 17598 <at> debbugs.gnu.org (full text, mbox):
Stefan Monnier wrote:
>> Stefan, is it OK to fix emacs-24's configure.ac to treat AIX 7+ like earlier
>> >AIX releases, as Dmitry suggests?
> Yes, that sounds fine,
Ah, this fix was already in emacs-24 so I didn't need to bother you.
The fix should appear in the next pretest. Sorry about the noise.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Fri, 30 May 2014 07:55:01 GMT)
Full text and
rfc822 format available.
Message #77 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
> On 05/28/2014 09:43 AM, Gilles PION wrote:
> > configure: error: Emacs does not support `powerpc-ibm-aix7.1.0.0'
> > systems.
>
> One way to work around that bug, without editing 'configure', is to
> configure this way:
>
> ./configure powerpc-ibm-aix6
>
> This should work even on AIX 7.
Tried this, now "configure" is successful but "make" fails:
Warning: arch-independent data dir
`<TOOLSDIR>/emacs-24.3.91/share/emacs/24.3.91/etc/': Invalid argument
Warning: Lisp directory `<HOMEDIR>/src/emacs/24.3.91/emacs-24.3.91/lisp':
Invalid argument
Error: charsets directory not found:
<TOOLSDIR>/emacs-24.3.91/share/emacs/24.3.91/etc/charsets
Emacs will not function correctly without the character map files.
Please check your installation!
Directories are present but investigating further show that strange system
call ("faccessat") failures (EINVAL) occurs when temacs is running.
I've attached truss output to this message
Note: I'm using gcc 7.8.2 compiler boostraped and build on an AIX 6.1 host
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
[anonymized-truss-temacs-run.log (application/octet-stream, attachment)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Fri, 30 May 2014 08:28:01 GMT)
Full text and
rfc822 format available.
Message #80 received at 17598 <at> debbugs.gnu.org (full text, mbox):
Gilles PION wrote:
> Tried this, now "configure" is successful but "make" fails:
>
> Warning: arch-independent data dir
> `<TOOLSDIR>/emacs-24.3.91/share/emacs/24.3.91/etc/': Invalid argument
This is odd; it looks like some third-party value (InfoCrafter, say?)
got into your environment. I don't see this on AIX 7.1.
> Note: I'm using gcc 7.8.2 compiler boostraped and build on an AIX 6.1 host
I guess you mean GCC 4.8.2?
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Fri, 30 May 2014 09:03:02 GMT)
Full text and
rfc822 format available.
Message #83 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
>
> On 05/28/2014 09:40 AM, Gilles PION wrote:
> >
> > OK, recompiling with those fixes (after configuring using "CFLAGS="-O0
> > -g3" ):
> > 1rst step: ./configure
> > 2nd step: fix src/config.h as indicated
> > 3rd step make
>
> Did you remember to remove the line "RALLOC_OBJ=ralloc.o" from
> src/Makefile?
No, I forgot
>Also, to play it safe, please remove src/ralloc.o, run
> 'make', and make sure src/ralloc.o is not created.
Just redone the 3 steps plus the Makefile fix:
$ grep RALLOC emacs-24.3.91/src/Makefile
####RALLOC_OBJ=ralloc.o
RALLOC_OBJ=
otherobj= $(TERMCAP_OBJ) $(PRE_ALLOC_OBJ) $(GMALLOC_OBJ) $(RALLOC_OBJ) \
and ralloc.o wasn't there
but still having SIGSEV:
+ cd ./emacs-24.3.91/src
+ export MALLOCTYPE=debug
+ export MALLOCDEBUG=postfree_checking,validate_ptrs,stack_depth:1000
+ ./temacs -Q
Fatal error 11: Segmentation violation./debug-malloc-run.sh[5]: 19136714
Memory fault
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Fri, 30 May 2014 09:04:02 GMT)
Full text and
rfc822 format available.
Message #86 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
> Gilles PION wrote:
> > Tried this, now "configure" is successful but "make" fails:
> >
> > Warning: arch-independent data dir
> > `<TOOLSDIR>/emacs-24.3.91/share/emacs/24.3.91/etc/': Invalid argument
>
> This is odd; it looks like some third-party value (InfoCrafter, say?)
> got into your environment. I don't see this on AIX 7.1.
>
> > Note: I'm using gcc 7.8.2 compiler boostraped and build on an AIX 6.1
host
>
> I guess you mean GCC 4.8.2?
Exact, typo again.
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Fri, 30 May 2014 10:42:02 GMT)
Full text and
rfc822 format available.
Message #89 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
Can you try the development trunk as well? Since I guess you have no bzr, there is
a link to recently enough tarball: http://37.139.80.10/tmp/trunk-r117204.tgz.
First, try to compile as usual, i.e.
CFLAGS='-O0 -g3' ./configure --enable-checking
If that fails, try to use system malloc by applying an attached patch. This requires
you to regenerate build files with 'sh autogen.sh', which means that you will need
recently enough autoconf, automake and m4:
http://www.oss4aix.org/download/RPMS/autoconf/autoconf-2.69-1.aix5.1.ppc.rpm
http://www.oss4aix.org/download/RPMS/automake/automake-1.14.1-1.aix5.1.ppc.rpm
http://www.oss4aix.org/download/RPMS/m4/m4-1.4.17-1.aix5.1.ppc.rpm
(after installing, make sure /usr/bin/m4 is symlinked to /opt/freeware/bin/m4).
With system malloc, dumped executable will not work. But temacs should, and
you should be able to try it with MALLOCTYPE and MALLOCDEBUG.
Dmitry
[aix_system_malloc.patch (text/x-patch, attachment)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Fri, 30 May 2014 12:16:01 GMT)
Full text and
rfc822 format available.
Message #92 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
>
> Can you try the development trunk as well? Since I guess you have no
> bzr, there is
> a link to recently enough tarball:
http://37.139.80.10/tmp/trunk-r117204.tgz.
Ok, downloading now,
stay tuned
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Fri, 30 May 2014 13:32:02 GMT)
Full text and
rfc822 format available.
Message #95 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
>
> First, try to compile as usual, i.e.
>
> CFLAGS='-O0 -g3' ./configure --enable-checking
>
> If that fails,
It did:
[1] + Memory fault ./trunk/src/emacs&
> try to use system malloc by applying an attached
> patch.
Ok, tring this now
(please be patient, byte-compile of lisp modules with a non optimized
emacs build is pretty long)
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Fri, 30 May 2014 14:46:01 GMT)
Full text and
rfc822 format available.
Message #98 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
> With system malloc, dumped executable will not work. But temacs should,
and
> you should be able to try it with MALLOCTYPE and MALLOCDEBUG.
>
Still crashing.
I can provide a truss output of the "temacs -Q run if it can help but may
become rather large.
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Fri, 30 May 2014 14:58:02 GMT)
Full text and
rfc822 format available.
Message #101 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
>
> > With system malloc, dumped executable will not work. But temacs
should, and
> > you should be able to try it with MALLOCTYPE and MALLOCDEBUG.
> >
> Still crashing.
> I can provide a truss output of the "temacs -Q run if it can help
> but may become rather large.
Note that, with this latest build, I've been unable to have "temacs" crash
yet, when run without malloc debug env variables
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Fri, 30 May 2014 21:41:02 GMT)
Full text and
rfc822 format available.
Message #104 received at 17598 <at> debbugs.gnu.org (full text, mbox):
On 05/30/2014 07:57 AM, Gilles PION wrote:
> Note that, with this latest build, I've been unable to have "temacs"
> crash yet, when run without malloc debug env variables .
This suggests that it's related to memory allocation somehow, sigh.
On the off-chance that it's also related to the code that the compiler
generates, I installed changes into the emacs-24 branch to get 'xlc'
working again on 32-bit AIX. If you have access to xlc, please try
'./configure CC=xlc' from the latest emacs-24 branch (and without any
other patches).
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Sat, 31 May 2014 04:20:01 GMT)
Full text and
rfc822 format available.
Message #107 received at 17598 <at> debbugs.gnu.org (full text, mbox):
By the way, the attachment anonymized-config.log at the start of this
bug report
<http://debbugs.gnu.org/cgi/bugreport.cgi?msg=5;filename=anonymized-config.log;att=2;bug=17598>
is from Emacs 24.3, not from Emacs pretest 24.3.91. Could you send a
config.log from the latest emacs-24 branch, once you've built it? Thanks.
Another idea: if it's a bug in the use of some of the libraries, perhaps
you can make it go away with something like this:
./configure --without-all --with-xft=no --with-x-toolkit=lucid \
--disable-acl \
ac_cv_path_PKG_CONFIG=no \
ac_cv_header_X11_extensions_Xrandr_h=no \
ac_cv_header_X11_extensions_Xinerama_h=no
(yes, this is painful; sorry).
The idea is to build something with minimal library dependencies; run:
ldd src/temacs
to see what libraries you have linked in, and make that list as small as
you can while still keeping X. Presumably you have an old version of
Emacs around that works; you can use "ldd" too see what libraries it's
using and try to make sure you aren't using any new ones.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Sat, 31 May 2014 13:31:08 GMT)
Full text and
rfc822 format available.
Message #110 received at 17598 <at> debbugs.gnu.org (full text, mbox):
> Date: Fri, 30 May 2014 14:39:50 -0700
> From: Paul Eggert <eggert <at> cs.ucla.edu>
> Cc: 17598 <at> debbugs.gnu.org
>
> On 05/30/2014 07:57 AM, Gilles PION wrote:
> > Note that, with this latest build, I've been unable to have "temacs"
> > crash yet, when run without malloc debug env variables .
>
> This suggests that it's related to memory allocation somehow, sigh.
>
> On the off-chance that it's also related to the code that the compiler
> generates, I installed changes into the emacs-24 branch to get 'xlc'
> working again on 32-bit AIX. If you have access to xlc, please try
> './configure CC=xlc' from the latest emacs-24 branch (and without any
> other patches).
The changes to lisp.h intended to enhance debugging on AIX broke
debugging when GCC 3.x is used to compile the code: I get "No symbol
"VALMASK" in current context." when xbacktrace is invoked.
So I reverted that part of the changes. I hope there's a way to
define the relevant macros only for AIX, so that other platforms
aren't affected, certainly not during the pretest.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Sun, 01 Jun 2014 07:13:02 GMT)
Full text and
rfc822 format available.
Message #113 received at 17598 <at> debbugs.gnu.org (full text, mbox):
Eli Zaretskii wrote:
> The changes to lisp.h intended to enhance debugging on AIX broke
> debugging when GCC 3.x is used to compile the code: I get "No symbol
> "VALMASK" in current context." when xbacktrace is invoked.
Sorry about that. It appears you ran into GCC bug 23336, an old bug I
wasn't familiar with. By the time this bug was fixed in 2006, the GCC
folks had stopped maintaining GCC 3.x. See:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=23336
> I hope there's a way to
> define the relevant macros only for AIX, so that other platforms
> aren't affected, certainly not during the pretest.
I gave that a shot in emacs-24 bzr 117185.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Sun, 01 Jun 2014 15:04:01 GMT)
Full text and
rfc822 format available.
Message #116 received at 17598 <at> debbugs.gnu.org (full text, mbox):
> Date: Sun, 01 Jun 2014 00:12:24 -0700
> From: Paul Eggert <eggert <at> cs.ucla.edu>
> CC: gpion <at> lfdj.com, dmantipov <at> yandex.ru, 17598 <at> debbugs.gnu.org
>
> Eli Zaretskii wrote:
>
> > The changes to lisp.h intended to enhance debugging on AIX broke
> > debugging when GCC 3.x is used to compile the code: I get "No symbol
> > "VALMASK" in current context." when xbacktrace is invoked.
>
> Sorry about that. It appears you ran into GCC bug 23336, an old bug I
> wasn't familiar with. By the time this bug was fixed in 2006, the GCC
> folks had stopped maintaining GCC 3.x. See:
>
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=23336
Yes, I think you're right.
> > I hope there's a way to
> > define the relevant macros only for AIX, so that other platforms
> > aren't affected, certainly not during the pretest.
>
> I gave that a shot in emacs-24 bzr 117185.
Thank you, this works for me.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Mon, 02 Jun 2014 15:54:02 GMT)
Full text and
rfc822 format available.
Message #119 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
>
> On the off-chance that it's also related to the code that the compiler
> generates, I installed changes into the emacs-24 branch to get 'xlc'
> working again on 32-bit AIX. If you have access to xlc, please try
> './configure CC=xlc' from the latest emacs-24 branch (and without any
> other patches).
Unfortunately, using xlc, compilation fails with the following error (I'm
using trunk-r117204.tgz):
/usr/vac/bin/xlc -qlanglvl=extc99 -DHAVE_CONFIG_H -I. -I../lib
-I../src -I../src -I/fdj/opt/emacs-24.3.91/include -g -O -c dtoastr.c
"ftoastr.c", line 80.37: 1506-045 (S) Undeclared identifier
_GL_FLT_PREC_BOUND.
"ftoastr.c", line 80.23: 1506-1324 (S) Array size must have integer type.
gmake[2]: *** [dtoastr.o] Error 1
Do I have to fetch a newer version of the sources? Note: I need a .tar
archive since I do not have bzr
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Mon, 02 Jun 2014 16:09:02 GMT)
Full text and
rfc822 format available.
Message #122 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
>
> By the way, the attachment anonymized-config.log at the start of this
> bug report
> <http://debbugs.gnu.org/cgi/bugreport.cgi?msg=5;filename=anonymized-
> config.log;att=2;bug=17598>
> is from Emacs 24.3, not from Emacs pretest 24.3.91. Could you send a
> config.log from the latest emacs-24 branch, once you've built it?
Thanks.
>
> Another idea: if it's a bug in the use of some of the libraries, perhaps
> you can make it go away with something like this:
>
> ./configure --without-all --with-xft=no --with-x-toolkit=lucid \
> --disable-acl \
> ac_cv_path_PKG_CONFIG=no \
> ac_cv_header_X11_extensions_Xrandr_h=no \
> ac_cv_header_X11_extensions_Xinerama_h=no
>
> (yes, this is painful; sorry).
Now getting the following error at link time (config and 2nd pass make log
files attached)
ld: 0711-317 ERROR: Undefined symbol: .XpmReadFileToPixmap
ld: 0711-345 Use the -bloadmap or -bnoquiet option to obtain more
information.
collect2: error: ld returned 8 exit status
gmake[1]: *** [temacs] Error 1
gmake[1]: Leaving directory `<HOMEDIR>/src/emacs/24.3.91/trunk/src'
gmake: *** [src] Error 2
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
[anonymized-config.log (application/octet-stream, attachment)]
[anonymized-make.log (application/octet-stream, attachment)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Mon, 02 Jun 2014 16:17:02 GMT)
Full text and
rfc822 format available.
Message #125 received at 17598 <at> debbugs.gnu.org (full text, mbox):
Thanks for helping to debug this, but the attachments are rather large.
Please could you compress any future logs before attaching them? TIA.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Mon, 02 Jun 2014 16:25:02 GMT)
Full text and
rfc822 format available.
Message #128 received at 17598 <at> debbugs.gnu.org (full text, mbox):
On 06/02/2014 08:53 AM, Gilles PION wrote:
>
> Unfortunately, using xlc, compilation fails with the following error
> (I'm using trunk-r117204.tgz):
If you're using the trunk, that should be fixed in trunk bzr 117219.
Please upgrade to the latest version.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Mon, 02 Jun 2014 16:36:01 GMT)
Full text and
rfc822 format available.
Message #131 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
> Thanks for helping to debug this, but the attachments are rather large.
> Please could you compress any future logs before attaching them?
Sure, I'll do it next time
Sorry for that
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Mon, 02 Jun 2014 16:39:02 GMT)
Full text and
rfc822 format available.
Message #134 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
> >
> > Unfortunately, using xlc, compilation fails with the following error
> > (I'm using trunk-r117204.tgz):
>
> If you're using the trunk, that should be fixed in trunk bzr 117219.
> Please upgrade to the latest version.
I'm unable to use bzr here,
I need link to a .tgz file.
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Mon, 02 Jun 2014 16:53:02 GMT)
Full text and
rfc822 format available.
Message #137 received at 17598 <at> debbugs.gnu.org (full text, mbox):
Gilles PION wrote:
> I need link to a .tgz file.
http://hydra.nixos.org/build/11429722/download/1/emacs-24.4.50.tar.gz
You should be able to always find the latest one by going to
http://hydra.nixos.org/jobset/gnu/emacs-trunk
Click on the uppermost number that has any green digits in its row.
Eg 1122062 at time of writing.
On the next page, click on the number for the "tarball" row.
Then click on "source distribution" where it says "build products".
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Tue, 03 Jun 2014 05:29:01 GMT)
Full text and
rfc822 format available.
Message #140 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
I had to add "LIBS='-lXpm'" do configure environment to prevent the
following error at link time:
ld: 0711-317 ERROR: Undefined symbol: .XpmReadFileToPixmap
ld: 0711-345 Use the -bloadmap or -bnoquiet option to obtain more
information.
gmake[1]: *** [temacs] Error 8
gmake[1]: Leaving directory
`<HOMEDIR>/src/emacs/24.3.91/emacs-24.4.50/src'
gmake: *** [src] Error 2
But it was worth the shot since ...
... so far so good
No crash yet
I've played a bit with "customize" options
Note that I've compiled this build without any compiler optimization.
Could you suggest sensible configure options to rebuilt a definive version
and to confirm everything is OK now?
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Tue, 03 Jun 2014 07:05:01 GMT)
Full text and
rfc822 format available.
Message #143 received at 17598 <at> debbugs.gnu.org (full text, mbox):
Gilles PION wrote:
> No crash yet
Yay!
> I had to add "LIBS='-lXpm'" do configure environment to prevent the
> following error at link time:
Sorry, I'm missing some context here. How did you configure Emacs (what
were the arguments to 'configure')? Which version of Emacs were you
building?
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Tue, 03 Jun 2014 07:12:01 GMT)
Full text and
rfc822 format available.
Message #146 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
> > I had to add "LIBS='-lXpm'" do configure environment to prevent the
> > following error at link time:
>
> Sorry, I'm missing some context here. How did you configure Emacs (what
> were the arguments to 'configure')? Which version of Emacs were you
> building?
Version 24.4.50, downloaded yesterday, download link given by Glen Morris.
Configure:
CFLAGS="-O0" LIBS="-lXpm"./configure --prefix=<TOOLSDIR>/emacs-24.4.50
--exec-prefix=<TOOLSDIR>/emacs-24.4.50 --sysconfdir=<ETCDIR>
--localstatedir=<LOCALSTATEDIR> --without-all --with-x-toolkit=lucid
--enable-checking=all
NB: I'm going back to training session for a few hours
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Tue, 03 Jun 2014 14:03:01 GMT)
Full text and
rfc822 format available.
Message #149 received at 17598 <at> debbugs.gnu.org (full text, mbox):
On 06/03/2014 09:28 AM, Gilles PION wrote:
> Could you suggest sensible configure options to rebuilt
> a definitive version and to confirm everything is OK now?
Hm...here is what I can suggest for trunk on AIX.
1. Choose the compiler.
IIUC gcc can't create 64-bit binaries on AIX, but xlc can (-q64).
Note build with gcc is far more well tested, and I have no ideas
whether unexec will work for 64-bit XCOFF binary.
2. Choose CFLAGS.
If you don't want to debug, just use '-O2' for both gcc and xlc.
If you want to debug, add '-g3' for gcc and up to '-g9' for xlc
(for the latter, please report your results!).
If you want to fight for 1% with the very unexpected results, try:
for gcc:
- determine CPU type with 'prtconf | grep -i "Processor Type"'
==> Processor Type: PowerPC_POWERX (where X is Power core
generation) and use CFLAGS='-O3 -mcpu=powerX -mtune=powerX'
for xlc, try '-O5'.
3. Choose configuration options.
For some weird reasons, you can't link without -lXpm (probably some
other library from vendor's X requires it). So use:
--without-all --with-xpm
to get the bare minimum executable with X support. This is pretty
enough to do basic programming tasks (i.e. editing source code).
If you want more features, look through 'configure --help' output,
choose the features you want, install required external libraries,
and then try '--without-all --with-FEATURE'. Note that just
'configure' (with no options) means the most feature-rich configuration
which may be compiled with installed headers and libraries.
Dmitry
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Tue, 03 Jun 2014 16:55:02 GMT)
Full text and
rfc822 format available.
Message #152 received at 17598 <at> debbugs.gnu.org (full text, mbox):
On 06/03/2014 12:11 AM, Gilles PION wrote:
>
> CFLAGS="-O0" LIBS="-lXpm"./configure --prefix=<TOOLSDIR>/emacs-24.4.50
> --exec-prefix=<TOOLSDIR>/emacs-24.4.50 --sysconfdir=<ETCDIR>
> --localstatedir=<LOCALSTATEDIR> --without-all --with-x-toolkit=lucid
> --enable-checking=all
Thanks, this suggests that it's one of the optional libraries (perhaps
an image library) that's causing the problem. Please try to narrow it
down to which library it is. If an ordinary build links 37 libraries
and the above build links 5 libraries (use "ldd src/temacs" to find
out), then perhaps one of the 32 omitted libraries is causing the
problem; so try linking half of them (16 + the 5 essential ones) and see
whether the problem recurs. Then you can use binary search to narrow it
down to the guilty library.
> m...here is what I can suggest for trunk on AIX.
>
> 1. Choose the compiler.
> IIUC gcc can't create 64-bit binaries on AIX, but xlc can (-q64).
> Note build with gcc is far more well tested, and I have no ideas
> whether unexec will work for 64-bit XCOFF binary.
I suggest using 32-bit gcc. 64-bit unexec definitely does not work.
xlc should work on the trunk now, but as you mentioned it's less
well-tested.
> For some weird reasons, you can't link without -lXpm (probably some
> other library from vendor's X requires it). So use:
It's IBM's -lXaw; its Pixmap.o defines a function XPixmapLoader that
calls XpmReadFileToPixmap, but it doesn't have the proper information in
the shared library to let the linker find -lXpm on its own. Presumably
we can work around this problem by the approriate hacking in
configure.ac but that's lower priority than fixing the real bug here.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Tue, 03 Jun 2014 17:36:01 GMT)
Full text and
rfc822 format available.
Message #155 received at 17598 <at> debbugs.gnu.org (full text, mbox):
On 06/03/2014 08:54 PM, Paul Eggert wrote:
> It's IBM's -lXaw; its Pixmap.o defines a function XPixmapLoader that
> calls XpmReadFileToPixmap, but it doesn't have the proper information
> in the shared library to let the linker find -lXpm on its own.
Ugh. This looks like a new etc/PROBLEMS entry, isn't it?
Dmitry
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 04 Jun 2014 06:33:01 GMT)
Full text and
rfc822 format available.
Message #158 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
> > No crash yet
>
> Yay!
>
Oh no!
After having recompiled with IBM xlc default optimization level (-O) ,
entering "customize-group" crashes again ("Fatal error 11: Segmentation
violationMemory fault")
using either temacs or dumped emacs makes no difference.
Configure options:
./configure --prefix=<TOOLSDIR>/emacs-24.4.50
--exec-prefix=<TOOLSDIR>/emacs-24.4.50 --sysconfdir=<ETCDIR>
--localstatedir=<LOCALSTATEDIR> --without-all --with-x-toolkit=lucid
--with-xpm --enable-locallisppath=<LOCALLISPPATH> --enable-checking=all
will try to remove "--with xpm"
ldd results:
./emacs-24.4.50/src/temacs needs:
/usr/lib/libc.a(shr.o)
/usr/lib/libpthread.a(shr_xpg5.o)
/usr/lib/libXt.a(shr4.o)
/usr/lib/libX11.a(shr4.o)
/usr/lib/libcurses.a(shr42.o)
/usr/lib/libXpm.a(shr.o)
/usr/lib/libpthread.a(shr_comm.o)
/usr/lib/libXext.a(shr.o)
/usr/lib/libICE.a(shr.o)
/usr/lib/libSM.a(shr.o)
/usr/lib/libperfstat.a(shr.o)
/unix
/usr/lib/libcrypt.a(shr.o)
/usr/lib/libpthreads.a(shr_comm.o)
/usr/lib/libXi.a(shr.o)
/usr/lib/libpthreads.a(shr_xpg5.o)
/usr/lib/libiconv.a(shr4.o)
/usr/lib/libIM.a(shr.o)
/usr/lib/libgaimisc.a(shr.o)
/usr/lib/libgair4.a(shr.o)
/usr/lib/libcfg.a(shr.o)
/usr/lib/libodm.a(shr.o)
/usr/lib/liblvm.a(shr.o)
/usr/lib/libcorcfg.a(shr.o)
/usr/lib/libsrc.a(shr.o)
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 04 Jun 2014 07:00:03 GMT)
Full text and
rfc822 format available.
Message #161 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
> will try to remove "--with xpm"
>
Done and no crashes yet
I'm thinking testing a gcc build using same configure settings or alse aix
higher optimization level (O5)
Please be patient, going to be out of office for a few hours again.
(NB: added Dmitryi in CC list)
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 04 Jun 2014 07:08:02 GMT)
Full text and
rfc822 format available.
Message #164 received at 17598 <at> debbugs.gnu.org (full text, mbox):
On 06/04/2014 10:31 AM, Gilles PION wrote:
> will try to remove "--with xpm"
As it was said, you can't use --with-x-toolkit=lucid and --without-xpm
(because Lucid toolkit uses -lXaw, and the latter one requires -lXpm on
your system). Try --with-x-toolkit=no instead (don't worry, you will have
X but with ugly menus, scroll bars, popups and other similar stuff :-().
> ldd results:
> ./emacs-24.4.50/src/temacs needs:
> /usr/lib/libc.a(shr.o)
> /usr/lib/libpthread.a(shr_xpg5.o)
> /usr/lib/libXt.a(shr4.o)
> /usr/lib/libX11.a(shr4.o)
> /usr/lib/libcurses.a(shr42.o)
> /usr/lib/libXpm.a(shr.o)
> /usr/lib/libpthread.a(shr_comm.o)
Did you use CC=xlc or CC=xlc_r? The first one may be unsafe since you
link with libpthread.a.
Dmitry
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 04 Jun 2014 11:23:02 GMT)
Full text and
rfc822 format available.
Message #167 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
>
> > ldd results:
> > ./emacs-24.4.50/src/temacs needs:
> > /usr/lib/libc.a(shr.o)
> > /usr/lib/libpthread.a(shr_xpg5.o)
> > /usr/lib/libXt.a(shr4.o)
> > /usr/lib/libX11.a(shr4.o)
> > /usr/lib/libcurses.a(shr42.o)
> > /usr/lib/libXpm.a(shr.o)
> > /usr/lib/libpthread.a(shr_comm.o)
>
> Did you use CC=xlc or CC=xlc_r? The first one may be unsafe since you
> link with libpthread.a.
just "xlc"
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 04 Jun 2014 15:58:02 GMT)
Full text and
rfc822 format available.
Message #170 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
>
>
> > will try to remove "--with xpm"
> >
> Done and no crashes yet
>
> I'm thinking testing a gcc build using same configure settings or
> alse aix higher optimization level (O5)
>
> Please be patient, going to be out of office for a few hours again.
>
> (NB: added Dmitryi in CC list)
Build with CFLAGS="-O3" seem to be OK also
Tried another build with "CC=xlc_r" as suggested by Dmitry and putting
bacj "--with-xpm"
With this setting, emacs doesn't crashes but goes to en endless loop as
soon as I enter "customize"
I thing that the working combo could be
"CC=xlc_r" and NO "--with-xpm"
Going to try this now
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 04 Jun 2014 16:22:02 GMT)
Full text and
rfc822 format available.
Message #173 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
>
> I thing that the working combo could be
>
> "CC=xlc_r" and NO "--with-xpm"
Confirmed: this one works (with CFLAGS=-O3)
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Sat, 07 Jun 2014 19:59:01 GMT)
Full text and
rfc822 format available.
Message #176 received at 17598 <at> debbugs.gnu.org (full text, mbox):
Gilles PION wrote:
> I thing that the working combo could be
>
> "CC=xlc_r" and NO "--with-xpm"
I've reread this bug report and tried to codify all the feedback we've
gotten, and have installed this as emacs-24 bzr 117216. In particular:
* --without-xpm is now the default on AIX.
* Link with -lXpm if also linking with -lXaw, even if --without-xpm.
* Define _THREAD_SAFE on AIX (that's what xlc_r does).
I've tested this on AIX 7.1 but I can't easily test X11 so please give
it a try when you have the chance.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Tue, 10 Jun 2014 06:51:02 GMT)
Full text and
rfc822 format available.
Message #179 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
>
> I've reread this bug report and tried to codify all the feedback we've
> gotten, and have installed this as emacs-24 bzr 117216. In particular:
/../
> I've tested this on AIX 7.1 but I can't easily test X11 so please give
> it a try when you have the chance.
Would you please give me the link to the .tgz file corresponding to this
release ?
Tried to browse form there: http://hydra.nixos.org/jobset/gnu/emacs-trunk
but I'm lost
Regards
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Tue, 10 Jun 2014 15:04:02 GMT)
Full text and
rfc822 format available.
Message #182 received at 17598 <at> debbugs.gnu.org (full text, mbox):
Gilles PION wrote:
> Tried to browse form there:http://hydra.nixos.org/jobset/gnu/emacs-trunk
> but I'm lost
I was talking the emacs-24 branch, so you can browse from here:
http://hydra.nixos.org/jobset/gnu/emacs-24
Click on the top "tarball →" link, then click just under "Build
products" on the next page; it should say "Source Source distribution
emacs-24.3.91.tar.gz" or something like that. I just now did that and
came up with:
http://hydra.nixos.org/build/11774896/download/1/emacs-24.3.91.tar.gz
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 11 Jun 2014 07:37:02 GMT)
Full text and
rfc822 format available.
Message #185 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
>
> I've tested this on AIX 7.1 but I can't easily test X11 so please give
> it a try when you have the chance.
Unfortunatly, this build made me discover that my previous tests were'nt
comprehensive enough
M-X shell makes emacs unresponsive (frozen)
configured with "--without-all --with-x-toolkit=lucid",
CC=/usr/vac/bin/xlc_r CFLAGS="-O3 -g"
Here's debugger backtrace after ^C:
(gdb) run --no-site-file --no-init-file
Starting program: <TOOLSDIR>/emacs-24.3.91/bin/emacs --no-site-file
--no-init-file
[New Thread 1]
Program received signal SIGINT, Interrupt.
[Switching to Thread 1]
0xd0122bec in read () from /usr/lib/threads/libc.a(shr.o)
(gdb) where
#0 0xd0122bec in read () from /usr/lib/threads/libc.a(shr.o)
#1 0x1004167c in emacs_read (fildes=1375217388, buf=0x4, nbyte=4)
at sysdep.c:2279
#2 0x101cb65c in create_process (process=270315820, new_argv=0x18,
current_dir=538786192) at process.c:1903
#3 0x101cb260 in Fstart_process (nargs=541711672,
args=0x201a0b6d <_$STATIC+629>) at process.c:5157
#4 0x1001ab98 in Ffuncall (nargs=268906868, args=0x0) at eval.c:2796
#5 0x1001a4ec in Fapply (nargs=550438488, args=0x0) at eval.c:2300
Backtrace stopped: frame did not save the PC
Note: freeze happens also when not running under X11 (emacs -nw)
Going to try this with gcc now
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 11 Jun 2014 07:55:01 GMT)
Full text and
rfc822 format available.
Message #188 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
>
> Going to try this with gcc now
Not successfull, getting same error at runtime already spotted on May 30:
Remember?
temacs "trussed" run output excerpt:
25165962: 20775055: faccessat(-3041965,
"<HOMEDIR>/src/emacs/24.3.91/emacs-24.3.91/lisp/.", 0, 4) Err#22 EINVAL
Warning: Lisp directory `<HOMEDIR>/src/emacs/24.3.91/emacs-24.3.91/lisp':
Invalid argument
25165962: 20775055: faccessat(-3041965,
"<TOOLSDIR>/emacs-24.3.91/share/emacs/24.3.91/etc/charsets/.", 0, 4)
Err#22 EINVAL
Error: charsets directory not found:
<TOOLSDIR>/emacs-24.3.91/share/emacs/24.3.91/etc/charsets
Emacs will not function correctly without the character map files.
Please check your installation!
25165962: 20775055: kfcntl(1, F_GETFL, 0x00000008) = 1
25165962: 20775055: close(1) = 0
25165962: 20775055: close(2) = 0
25165962: 20775055: __loadx(0x04400000, 0x2FF22000, 0x00000800,
0x0000D032, 0x00000000) = 0x00000000
25165962: 20775055: _exit(1)
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 11 Jun 2014 09:59:02 GMT)
Full text and
rfc822 format available.
Message #191 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
> >
> > Going to try this with gcc now
>
> Not successfull, getting same error at runtime already spotted on May
30:
On a AIX 6.1 host the same binary (compiled on a 7.1 platform) runs fine.
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Wed, 11 Jun 2014 20:26:02 GMT)
Full text and
rfc822 format available.
Message #194 received at 17598 <at> debbugs.gnu.org (full text, mbox):
On 06/11/2014 12:53 AM, Gilles PION wrote:
> Not successfull, getting same error at runtime already spotted on May 30:
>
> Remember?
I remember fixing this in the trunk. I can reproduce the bug (since it
doesn't require X), and backporting the trunk patch fixed it for me. I
installed this as emacs-24 bzr 117234. You can get the resulting tarball
here:
http://cs.ucla.edu/~eggert/emacs-24.3.91.tar.gz
Please give it a try. Eventually it should appear on Hydra, too, but
Hydra ran out of disk space so that'll have to wait until someone cleans
it out.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Thu, 12 Jun 2014 08:46:01 GMT)
Full text and
rfc822 format available.
Message #197 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
>
> I remember fixing this in the trunk. I can reproduce the bug (since it
> doesn't require X), and backporting the trunk patch fixed it for me. I
> installed this as emacs-24 bzr 117234. You can get the resulting tarball
> here:
>
> http://cs.ucla.edu/~eggert/emacs-24.3.91.tar.gz
>
> Please give it a try. Eventually it should appear on Hydra, too, but
> Hydra ran out of disk space so that'll have to wait until someone cleans
> it out.
Same error at runtime.
I would not be much surprised if it was rather the GCC compiler being the
source of this error
I've also tried compiling this tarball on an AIX 6.1 host, using with the
same configure options and gcc binary:
the emacs executable produced with setup runs fine on 6.1 and 7.1 hosts.
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Fri, 13 Jun 2014 02:25:02 GMT)
Full text and
rfc822 format available.
Message #200 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
Gilles PION wrote:
> Same error at runtime.
>
> I would not be much surprised if it was rather the GCC compiler being the
> source of this error
Which GCC compiler are you using? I am using 4.8.1, and it works for
me; see attached files.
Can you please run the following commands (or their equivalents) in a
fresh directory, and send all the shell output, config.log, lib/fcntl.h,
src/config.h, and emacs-trussout.txt files (compressed), in the same
style as the files I've attached to this email? Thanks.
wget http://cs.ucla.edu/~eggert/emacs-24.3.91.tar.gz
gunzip < emacs-24.3.91.tar.gz | tar xf -
cd emacs-24.3.91
./configure --without-all --with-x-toolkit=lucid CFLAGS="-O3 -g"
make
truss -df -o emacs-trussout.txt src/emacs
[Type "M-x shell RET C-d C-x C-c".]
You may have to run "src/emacs -nw" instead of "src/emacs", if you have
an X display.
[shellout.txt.gz (application/gzip, attachment)]
[config.log.gz (application/gzip, attachment)]
[config.h.gz (application/gzip, attachment)]
[emacs-trussout.txt.gz (application/gzip, attachment)]
[fcntl.h.gz (application/gzip, attachment)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Fri, 13 Jun 2014 08:04:02 GMT)
Full text and
rfc822 format available.
Message #203 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
>
> Which GCC compiler are you using? I am using 4.8.1, and it works for
> me; see attached files.
4.8.2, boostratpped on a AIX 6.1 Host
>
> Can you please run the following commands (or their equivalents) in a
> fresh directory, and send all the shell output, config.log, lib/fcntl.h,
> src/config.h, and emacs-trussout.txt files (compressed), in the same
> style as the files I've attached to this email? Thanks.
>
> wget http://cs.ucla.edu/~eggert/emacs-24.3.91.tar.gz
> gunzip < emacs-24.3.91.tar.gz | tar xf -
> cd emacs-24.3.91
> ./configure --without-all --with-x-toolkit=lucid CFLAGS="-O3 -g"
Also appended:
* "CC=path_to_my_gcc_bin" since gcc is not in my PATH
* --prefix=/<path_to_writeable_private_dir>
> make
> truss -df -o emacs-trussout.txt src/emacs
"make" fails to dump emacs since that's "temacs" which is crashing
That's why I have attached truss output of temacs instead
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
[anonymized-config.h.gz (application/octet-stream, attachment)]
[anonymized-config.log.gz (application/octet-stream, attachment)]
[anonymized-fcntl.h.gz (application/octet-stream, attachment)]
[anonymized-shellout.txt.gz (application/octet-stream, attachment)]
[anonymized-temacs-trussout.txt.gz (application/octet-stream, attachment)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Fri, 13 Jun 2014 14:04:02 GMT)
Full text and
rfc822 format available.
Message #206 received at 17598 <at> debbugs.gnu.org (full text, mbox):
Gilles PION wrote:
>> >Which GCC compiler are you using? I am using 4.8.1, and it works for
>> >me; see attached files.
> 4.8.2, boostratpped on a AIX 6.1 Host
I'm afraid that does not work. If you are using GCC on AIX 7.1, your
GCC must be built on AIX 7.1. A GCC bootstrapped on AIX 6.1 will run on
AIX 7.1 but will generate incorrect code sometimes. That appears to be
what's happening in your case, as the truss output contains lines like this:
18809032: 42074209: 0.2175: faccessat(-3041965,
"<<HOMEDIR>>/src/emacs/emacs-gcc-crashtest/emacs-24.3.91/src/lib-src",
0, 4) Err#22 EINVAL
which means Emacs thinks AT_FDCWD equals -3041965, but this is wrong for
AIX 7.1: /usr/include/fcntl.h #defines AT_FDCWD to be -2. Apparently
your GCC is incorrectly using fixincluded files from AIX 6.1 (which do
not define AT_FDCWD) and Emacs is therefore incorrectly falling back on
gnulib's definition of AT_FDCWD.
Bad fixincluded files would explain a lot of problems, including the
subshell problem you reported earlier, as well as the Xpm problems.
Possibly we were too hasty in suppressing Emacs's use of Xpm on AIX?
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Fri, 13 Jun 2014 15:57:01 GMT)
Full text and
rfc822 format available.
Message #209 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
> >> >Which GCC compiler are you using? I am using 4.8.1, and it works
for
> >> >me; see attached files.
> > 4.8.2, boostratpped on a AIX 6.1 Host
>
> I'm afraid that does not work. If you are using GCC on AIX 7.1, your
> GCC must be built on AIX 7.1. A GCC bootstrapped on AIX 6.1 will run on
> AIX 7.1 but will generate incorrect code sometimes. That appears to be
> what's happening in your case, as the truss output contains lines like
this:
>
> 18809032: 42074209: 0.2175: faccessat(-3041965,
> "<<HOMEDIR>>/src/emacs/emacs-gcc-crashtest/emacs-24.3.91/src/lib-src",
> 0, 4) Err#22 EINVAL
>
> which means Emacs thinks AT_FDCWD equals -3041965, but this is wrong for
> AIX 7.1: /usr/include/fcntl.h #defines AT_FDCWD to be -2. Apparently
> your GCC is incorrectly using fixincluded files from AIX 6.1 (which do
> not define AT_FDCWD) and Emacs is therefore incorrectly falling back on
> gnulib's definition of AT_FDCWD.
>
> Bad fixincluded files would explain a lot of problems, including the
> subshell problem you reported earlier, as well as the Xpm problems.
I should be now able to bootstrap a new instance of GCC for aix7.1 since
my colleague which administrate this machine have found some time to
reinstall from scratch a working version of AIX xlc compiler.
Will try this next week.
> Possibly we were too hasty in suppressing Emacs's use of Xpm on AIX?
Maybe.
I'll also try do make a test build adding "-with-xpm" to "configure" when
the new version of the GCC compiler will be set up, and keep you informed
of the results.
Regards, and have a nice week-end,
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Mon, 16 Jun 2014 09:52:01 GMT)
Full text and
rfc822 format available.
Message #212 received at 17598 <at> debbugs.gnu.org (full text, mbox):
On 06/13/2014 07:56 PM, Gilles PION wrote:
> > I'm afraid that does not work. If you are using GCC on AIX 7.1, your
> > GCC must be built on AIX 7.1. A GCC bootstrapped on AIX 6.1 will run on
> > AIX 7.1 but will generate incorrect code sometimes. That appears to be
> > what's happening in your case, as the truss output contains lines like this:
BTW, this GCC seems to be compiled on AIX 7.1:
http://www.oss4aix.org/download/RPMS/gcc/gcc-4.8.2-1.aix7.1.ppc.rpm
Dmitry
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Mon, 16 Jun 2014 12:46:01 GMT)
Full text and
rfc822 format available.
Message #215 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
>
> BTW, this GCC seems to be compiled on AIX 7.1:
>
> http://www.oss4aix.org/download/RPMS/gcc/gcc-4.8.2-1.aix7.1.ppc.rpm
Thanks, this helped since I had issues bootstraping my own gcc build with
xlc
I've now recompiled emacs (tarkit from latest url given by Paul:
http://cs.ucla.edu/~eggert/emacs-24.3.91.tar.gz) with this gcc and it
*works*!!!
After that, also tried with "--with-xpm", but getting SIGSEV in that case.
Maybe I'll try to compile my own libXpm from sources and link emacs with
it.
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Mon, 16 Jun 2014 13:27:02 GMT)
Full text and
rfc822 format available.
Message #218 received at 17598 <at> debbugs.gnu.org (full text, mbox):
On 06/16/2014 04:44 PM, Gilles PION wrote:
> Maybe I'll try to compile my own libXpm from sources and link emacs with it.
Try libXpm from here first:
http://www.oss4aix.org/download/RPMS/libXpm/libXpm-3.5.10-2.aix6.1.ppc.rpm
http://www.oss4aix.org/download/RPMS/libXpm/libXpm-devel-3.5.10-2.aix6.1.ppc.rpm
BTW, linking 3rd-party libXpm with vendor's X libraries may be problematic.
Anyway, it's worth trying.
Dmitry
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#17598; Package
emacs.
(Mon, 16 Jun 2014 14:53:02 GMT)
Full text and
rfc822 format available.
Message #221 received at 17598 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
> Try libXpm from here first:
>
>
http://www.oss4aix.org/download/RPMS/libXpm/libXpm-3.5.10-2.aix6.1.ppc.rpm
> http://www.oss4aix.org/download/RPMS/libXpm/libXpm-devel-3.5.
> 10-2.aix6.1.ppc.rpm
Seem to be working when liked against this package: no crashes encountered
so far.
>
> BTW, linking 3rd-party libXpm with vendor's X libraries may be
problematic.
Youre probably right,
That's why I think I'll better stick using "--without-xpm": not sure
adding those kind dependencies just for having a colored icon is
worthwhile.
--
Gilles
. Si vous n'êtes pas destinataires de ce message, merci d'avertir
l'expéditeur de l'erreur de distribution et de le détruire
immédiatement.
Ce message contient des informations confidentielles ou appartenant
a La Française des Jeux. Il est établi a l'intention exclusive de
ses destinataires. Toute divulgation, utilisation, diffusion ou
reproduction (totale ou partielle) de ce message ou des
informations qu'il contient, doit être préalablement autorisée.
Tout message électronique est susceptible d'altération et son
intégrité ne peut etre assurée. La Française des Jeux décline toute
responsabilité au titre de ce message s'il a été modifie ou
falsifie.
[Message part 2 (text/html, inline)]
Reply sent
to
Paul Eggert <eggert <at> cs.ucla.edu>:
You have taken responsibility.
(Fri, 20 Jun 2014 04:20:02 GMT)
Full text and
rfc822 format available.
Notification sent
to
Gilles Pion <gpion <at> lfdj.com>:
bug acknowledged by developer.
(Fri, 20 Jun 2014 04:20:04 GMT)
Full text and
rfc822 format available.
Message #226 received at 17598-done <at> debbugs.gnu.org (full text, mbox):
As the issues in this bug report seem to be resolved, I'm closing it.
If further AIX-related issues come up we can open a new bug report.
bug archived.
Request was from
Debbugs Internal Request <help-debbugs <at> gnu.org>
to
internal_control <at> debbugs.gnu.org.
(Fri, 18 Jul 2014 11:24:04 GMT)
Full text and
rfc822 format available.
This bug report was last modified 10 years and 277 days ago.
Previous Next
GNU bug tracking system
Copyright (C) 1999 Darren O. Benham,
1997,2003 nCipher Corporation Ltd,
1994-97 Ian Jackson.