https://bugzilla.redhat.com/show_bug.cgi?id=1993670
Bug ID: 1993670 Summary: segfault with pango-view assert in cairo Product: Fedora Version: 34 Hardware: x86_64 OS: Linux Status: NEW Component: pango Severity: high Assignee: pwu@redhat.com Reporter: andre.maute@gmx.de QA Contact: extras-qa@fedoraproject.org CC: caillon+fedoraproject@gmail.com, fonts-bugs@lists.fedoraproject.org, gnome-sig@lists.fedoraproject.org, i18n-bugs@lists.fedoraproject.org, mclasen@redhat.com, pwu@redhat.com, rhughes@redhat.com, rstrode@redhat.com, sandmann@redhat.com, tagoh@redhat.com Target Milestone: --- Classification: Fedora
Description of problem:
1. This one works $ pango-view --nodisplay --text "abc"
2. This one doesn't $ pango-view --no-display --text "abc" pango-view: cairo-hash.c:217: _cairo_hash_table_destroy: Assertion `hash_table->live_entries == 0' failed. Aborted (core dumped)
So pango-view triggers an assertion in Cairo.
Version-Release number of selected component (if applicable):
$ dnf repoquery --installed pango pango-0:1.48.7-1.fc34.i686 pango-0:1.48.7-1.fc34.x86_64
How reproducible:
always
Steps to Reproduce: 1. see 1. of description above 2. see 2. of description above
Actual results: segfault
Expected results: no segfault
Additional info:
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
a3emdot andre.maute@gmx.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Comment|0 |updated
--- Comment #0 has been edited ---
Description of problem:
1. This one works $ pango-view --text "abc"
2. This one doesn't $ pango-view --no-display --text "abc" pango-view: cairo-hash.c:217: _cairo_hash_table_destroy: Assertion `hash_table->live_entries == 0' failed. Aborted (core dumped)
So pango-view triggers an assertion in Cairo.
Version-Release number of selected component (if applicable):
$ dnf repoquery --installed pango pango-0:1.48.7-1.fc34.i686 pango-0:1.48.7-1.fc34.x86_64
How reproducible:
always
Steps to Reproduce: 1. see 1. of description above 2. see 2. of description above
Actual results: segfault
Expected results: no segfault
Additional info:
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
--- Comment #1 from a3emdot andre.maute@gmx.de --- Under Fedora 34 $ pango-view --no-display --text "abc" --output=out.png works as expected
Now for Fedora 33:
Whereas under Fedora 33 one gets $ pango-view --no-display --text "abc" --output=out.png pango-view: cairo-hash.c:217: _cairo_hash_table_destroy: Assertion `hash_table->live_entries == 0' failed. Aborted (core dumped)
$ dnf repoquery --installed pango pango-0:1.48.4-1.fc33.i686 pango-0:1.48.4-1.fc33.x86_64
so somebody worked on this already, somehow. Both my Fedora 33 and Fedora 34 installations where at the moment of the bug reporting up to date
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
Peng Wu pwu@redhat.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Doc Type|--- |If docs needed, set a value Status|NEW |ASSIGNED
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
Fedora Update System updates@fedoraproject.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |MODIFIED
--- Comment #3 from Fedora Update System updates@fedoraproject.org --- FEDORA-2021-28fca597fb has been submitted as an update to Fedora 33. https://bodhi.fedoraproject.org/updates/FEDORA-2021-28fca597fb
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
Fedora Update System updates@fedoraproject.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|MODIFIED |ON_QA
--- Comment #4 from Fedora Update System updates@fedoraproject.org --- FEDORA-2021-28fca597fb has been pushed to the Fedora 33 testing repository. Soon you'll be able to install the update with the following command: `sudo dnf upgrade --enablerepo=updates-testing --advisory=FEDORA-2021-28fca597fb` You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2021-28fca597fb
See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates.
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
Fedora Update System updates@fedoraproject.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|ON_QA |CLOSED Resolution|--- |ERRATA Fixed In Version| |pango-1.48.4-2.fc33 Last Closed| |2021-09-10 16:35:56
--- Comment #5 from Fedora Update System updates@fedoraproject.org --- FEDORA-2021-28fca597fb has been pushed to the Fedora 33 stable repository. If problem still persists, please make note of it in this bug report.
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
--- Comment #6 from a3emdot andre.maute@gmx.de ---
pango-view still chokes on Fedora 33 AND on Fedora 34, with identical message
$ pango-view --no-display --text "abc" pango-view: cairo-hash.c:217: _cairo_hash_table_destroy: Assertion `hash_table->live_entries == 0' failed. Aborted (core dumped)
On my Fedora 33 I have ``` $ dnf list installed | grep pango pango.i686 1.48.4-2.fc33 @updates pango.x86_64 1.48.4-2.fc33 @updates pango-devel.x86_64 1.48.4-2.fc33 @updates pangomm.x86_64 2.42.2-1.fc33 @updates ```
On my Fedora 34 I have ``` $ dnf list installed | grep pango pango.i686 1.48.9-2.fc34 @updates pango.x86_64 1.48.9-2.fc34 @updates pango-devel.x86_64 1.48.9-2.fc34 @updates pangomm.x86_64 2.46.1-1.fc34 @updates ```
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
--- Comment #7 from Peng Wu pwu@redhat.com --- Could you try with the --output option?
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
--- Comment #8 from a3emdot andre.maute@gmx.de --- (In reply to Peng Wu from comment #7)
Could you try with the --output option?
using --no-display with the--output option it works, but nevertheless I the segfault without the --output option is disturbing.
Wouldn't it be possible, to simply do nothing if --no-display is specified without the --output option? Or perhaps output a small information for the user.
Segfaults are never an option because they may leak data.
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
--- Comment #9 from a3emdot andre.maute@gmx.de --- Created attachment 1823392 --> https://bugzilla.redhat.com/attachment.cgi?id=1823392&action=edit shell script showing nondeterminism
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
--- Comment #10 from a3emdot andre.maute@gmx.de --- Looks like there is some kind of nondeterminism involved, even if specifying the output option her are some runs with a simple shell script
$ ./pango-view-segfaults-20210915.sh
$ ./pango-view-segfaults-20210915.sh pango-view: cairo-hash.c:217: _cairo_hash_table_destroy: Assertion `hash_table->live_entries == 0' failed. ./pango-view-segfaults-20210915.sh: line 2: 39789 Aborted (core dumped) pango-view --no-display --dpi=72 --font="DejaVu Sans 8" --antialias=gray --backend=cairo --text="1234567890" --output="out8.png" pango-view: cairo-hash.c:217: _cairo_hash_table_destroy: Assertion `hash_table->live_entries == 0' failed. ./pango-view-segfaults-20210915.sh: line 8: 39820 Aborted (core dumped) pango-view --no-display --dpi=72 --font="DejaVu Sans 20" --antialias=gray --backend=cairo --text="1234567890" --output="out20.png"
$ ./pango-view-segfaults-20210915.sh pango-view: cairo-hash.c:217: _cairo_hash_table_destroy: Assertion `hash_table->live_entries == 0' failed. ./pango-view-segfaults-20210915.sh: line 3: 40010 Aborted (core dumped) pango-view --no-display --dpi=72 --font="DejaVu Sans 10" --antialias=gray --backend=cairo --text="1234567890" --output="out10.png" pango-view: cairo-hash.c:217: _cairo_hash_table_destroy: Assertion `hash_table->live_entries == 0' failed. ./pango-view-segfaults-20210915.sh: line 4: 40016 Aborted (core dumped) pango-view --no-display --dpi=72 --font="DejaVu Sans 12" --antialias=gray --backend=cairo --text="1234567890" --output="out12.png" pango-view: cairo-hash.c:217: _cairo_hash_table_destroy: Assertion `hash_table->live_entries == 0' failed. ./pango-view-segfaults-20210915.sh: line 9: 40048 Aborted (core dumped) pango-view --no-display --dpi=72 --font="DejaVu Sans 22" --antialias=gray --backend=cairo --text="1234567890" --output="out22.png" pango-view: cairo-hash.c:217: _cairo_hash_table_destroy: Assertion `hash_table->live_entries == 0' failed. ./pango-view-segfaults-20210915.sh: line 10: 40058 Aborted (core dumped) pango-view --no-display --dpi=72 --font="DejaVu Sans 24" --antialias=gray --backend=cairo --text="1234567890" --output="out24.png" pango-view: cairo-hash.c:217: _cairo_hash_table_destroy: Assertion `hash_table->live_entries == 0' failed. ./pango-view-segfaults-20210915.sh: line 16: 40147 Aborted (core dumped) pango-view --no-display --dpi=72 --font="DejaVu Sans 36" --antialias=gray --backend=cairo --text="1234567890" --output="out36.png"
$ ./pango-view-segfaults-20210915.sh pango-view: cairo-hash.c:217: _cairo_hash_table_destroy: Assertion `hash_table->live_entries == 0' failed. ./pango-view-segfaults-20210915.sh: line 3: 40307 Aborted (core dumped) pango-view --no-display --dpi=72 --font="DejaVu Sans 10" --antialias=gray --backend=cairo --text="1234567890" --output="out10.png" pango-view: cairo-hash.c:217: _cairo_hash_table_destroy: Assertion `hash_table->live_entries == 0' failed. ./pango-view-segfaults-20210915.sh: line 5: 40330 Aborted (core dumped) pango-view --no-display --dpi=72 --font="DejaVu Sans 14" --antialias=gray --backend=cairo --text="1234567890" --output="out14.png" pango-view: cairo-hash.c:217: _cairo_hash_table_destroy: Assertion `hash_table->live_entries == 0' failed. ./pango-view-segfaults-20210915.sh: line 8: 40368 Aborted (core dumped) pango-view --no-display --dpi=72 --font="DejaVu Sans 20" --antialias=gray --backend=cairo --text="1234567890" --output="out20.png"
$ ./pango-view-segfaults-20210915.sh pango-view: cairo-hash.c:217: _cairo_hash_table_destroy: Assertion `hash_table->live_entries == 0' failed. ./pango-view-segfaults-20210915.sh: line 7: 40620 Aborted (core dumped) pango-view --no-display --dpi=72 --font="DejaVu Sans 18" --antialias=gray --backend=cairo --text="1234567890" --output="out18.png" pango-view: cairo-hash.c:217: _cairo_hash_table_destroy: Assertion `hash_table->live_entries == 0' failed. ./pango-view-segfaults-20210915.sh: line 9: 40641 Aborted (core dumped) pango-view --no-display --dpi=72 --font="DejaVu Sans 22" --antialias=gray --backend=cairo --text="1234567890" --output="out22.png" pango-view: cairo-hash.c:217: _cairo_hash_table_destroy: Assertion `hash_table->live_entries == 0' failed. ./pango-view-segfaults-20210915.sh: line 13: 40717 Aborted (core dumped) pango-view --no-display --dpi=72 --font="DejaVu Sans 30" --antialias=gray --backend=cairo --text="1234567890" --output="out30.png" pango-view: cairo-hash.c:217: _cairo_hash_table_destroy: Assertion `hash_table->live_entries == 0' failed. ./pango-view-segfaults-20210915.sh: line 15: 40743 Aborted (core dumped) pango-view --no-display --dpi=72 --font="DejaVu Sans 34" --antialias=gray --backend=cairo --text="1234567890" --output="out34.png"
$ ./pango-view-segfaults-20210915.sh pango-view: cairo-hash.c:217: _cairo_hash_table_destroy: Assertion `hash_table->live_entries == 0' failed. ./pango-view-segfaults-20210915.sh: line 3: 40904 Aborted (core dumped) pango-view --no-display --dpi=72 --font="DejaVu Sans 10" --antialias=gray --backend=cairo --text="1234567890" --output="out10.png"
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
--- Comment #11 from a3emdot andre.maute@gmx.de --- The nondeterminism happens on my Fedora 33 installation
$ uname -a Linux localhost.localdomain 5.13.14-100.fc33.x86_64 #1 SMP Fri Sep 3 17:18:38 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux
$ dnf list installed | grep pango pango.i686 1.48.4-2.fc33 @updates pango.x86_64 1.48.4-2.fc33 @updates pango-devel.x86_64 1.48.4-2.fc33 @updates pangomm.x86_64 2.42.2-1.fc33 @updates
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
--- Comment #12 from a3emdot andre.maute@gmx.de --- Created attachment 1824034 --> https://bugzilla.redhat.com/attachment.cgi?id=1824034&action=edit shell script showing nondeterminism on Fedora 34
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
--- Comment #13 from a3emdot andre.maute@gmx.de --- Created attachment 1824035 --> https://bugzilla.redhat.com/attachment.cgi?id=1824035&action=edit shell script showing nondeterminism on Fedora 34 (output)
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
a3emdot andre.maute@gmx.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Flags| |needinfo?(pwu@redhat.com)
--- Comment #14 from a3emdot andre.maute@gmx.de --- I can finally reproduce the segfaults with the help of valgrind on Fedora 34
$ ./pango-view-segfaults-20210917.sh > pango-view-segfaults-20210917.txt 2>&1
where one can find the following in the written txt file --- pango-view: cairo-hash.c:217: _cairo_hash_table_destroy: Assertion `hash_table->live_entries == 0' failed. ==8093== ==8093== Process terminating with default action of signal 6 (SIGABRT): dumping core ==8093== at 0x4D932A2: raise (in /usr/lib64/libc-2.33.so) ==8093== by 0x4D7C8A3: abort (in /usr/lib64/libc-2.33.so) ==8093== by 0x4D7C788: __assert_fail_base.cold (in /usr/lib64/libc-2.33.so) ==8093== by 0x4D8BA15: __assert_fail (in /usr/lib64/libc-2.33.so) ==8093== by 0x4C642EF: ??? (in /usr/lib64/libcairo.so.2.11704.0) ==8093== by 0x4C64A0A: cairo_debug_reset_static_data (in /usr/lib64/libcairo.so.2.11704.0) ==8093== by 0x111E2C: ??? (in /usr/bin/pango-view) ==8093== by 0x10EAF6: ??? (in /usr/bin/pango-view) ==8093== by 0x4D7DB74: (below main) (in /usr/lib64/libc-2.33.so) ==8093== ==8093== Process terminating with default action of signal 11 (SIGSEGV) ==8093== General Protection Fault ==8093== at 0x5493C82: __pthread_once_slow (in /usr/lib64/libpthread-2.33.so) ==8093== by 0x4E8603E: __rpc_thread_variables.part.0 (in /usr/lib64/libc-2.33.so) ==8093== by 0x4EC861C: free_mem (in /usr/lib64/libc-2.33.so) ==8093== by 0x4EC8271: __libc_freeres (in /usr/lib64/libc-2.33.so) ==8093== by 0x48351E7: _vgnU_freeres (vg_preloaded.c:74) ==8093== ==8093== HEAP SUMMARY: ==8093== in use at exit: 1,226,659 bytes in 20,130 blocks ==8093== total heap usage: 37,522 allocs, 17,392 frees, 5,374,996 bytes allocated ==8093== ==8093== LEAK SUMMARY: ==8093== definitely lost: 151,552 bytes in 585 blocks ==8093== indirectly lost: 208,373 bytes in 8,421 blocks ==8093== possibly lost: 1,947 bytes in 9 blocks ==8093== still reachable: 861,435 bytes in 11,079 blocks ==8093== suppressed: 0 bytes in 0 blocks ==8093== Rerun with --leak-check=full to see details of leaked memory ==8093== ==8093== For lists of detected and suppressed errors, rerun with: -s ==8093== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 0 from 0) ---
$ uname -a Linux localhost.localdomain 5.13.15-200.fc34.x86_64 #1 SMP Wed Sep 8 15:51:46 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux
$ dnf list installed | grep pango pango.i686 1.48.10-1.fc34 @updates pango.x86_64 1.48.10-1.fc34 @updates pango-devel.x86_64 1.48.10-1.fc34 @updates pangomm.x86_64 2.46.1-1.fc34 @updates
$ dnf list installed | grep roboto google-roboto-condensed-fonts.noarch 2.138-8.fc34 @fedora google-roboto-fonts.noarch 2.138-8.fc34 @fedora
I would have liked to use the pango-view tool for a font rendering validation tool but the segfaults make it totally useless here.
It also looks like the segfaulting behavior depends on the font to be used. Using "DejaVu Sans" instead of "Roboto Condensed" I don't get the segfaults with the newer reproducing shell script.
So could we please reopen this issue?
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
Peng Wu pwu@redhat.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|CLOSED |NEW Resolution|ERRATA |--- Keywords| |Reopened
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
Peng Wu pwu@redhat.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo?(pwu@redhat.com) |
--- Comment #15 from Peng Wu pwu@redhat.com --- Sorry, it is still hard to reproduce it in Fedora 34.
I write a small patch for pango-view, not sure whether it works or not.
Could you try the pango build?
Koji URL: https://koji.fedoraproject.org/koji/taskinfo?taskID=76145896
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
--- Comment #16 from a3emdot andre.maute@gmx.de --- I installed the NEW rpm with
# rpm -iv --force --replacepkgs /tmp/pango-1.48.10-1.fc34.x86_64.rpm
the segfaults are still there
$ ./pango-view-segfaults-20210923.sh > pango-view-segfaults-20210923.txt 2>&1 $ cat pango-view-segfaults-20210923.txt | grep assert ==6949== by 0x4D7C788: __assert_fail_base.cold (in /usr/lib64/libc-2.33.so) ==6949== by 0x4D8BA15: __assert_fail (in /usr/lib64/libc-2.33.so) ==7010== by 0x4D7C788: __assert_fail_base.cold (in /usr/lib64/libc-2.33.so) ==7010== by 0x4D8BA15: __assert_fail (in /usr/lib64/libc-2.33.so) ==7081== by 0x4D7C788: __assert_fail_base.cold (in /usr/lib64/libc-2.33.so) ==7081== by 0x4D8BA15: __assert_fail (in /usr/lib64/libc-2.33.so) ==7099== by 0x4D7C788: __assert_fail_base.cold (in /usr/lib64/libc-2.33.so) ==7099== by 0x4D8BA15: __assert_fail (in /usr/lib64/libc-2.33.so)
I added some more pango-view lines to the reproducer script
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
--- Comment #17 from a3emdot andre.maute@gmx.de --- Created attachment 1825534 --> https://bugzilla.redhat.com/attachment.cgi?id=1825534&action=edit shell script showing nondeterminism on Fedora 34 2021-09-23
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
--- Comment #18 from a3emdot andre.maute@gmx.de --- Created attachment 1825535 --> https://bugzilla.redhat.com/attachment.cgi?id=1825535&action=edit shell script showing nondeterminism on Fedora 34 2021-09-23 (output)
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
--- Comment #19 from a3emdot andre.maute@gmx.de --- because I have both pango packages i686 and x64_86 installed
# rpm -q --whatprovides $(which pango-view) pango-1.48.10-1.fc34.x86_64
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
--- Comment #20 from Peng Wu pwu@redhat.com --- Does this issue still happen with Fedora 34 or 35?
If it still happens, maybe you can report this issue to upstream.
URL: https://gitlab.gnome.org/GNOME/pango
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
a3emdot andre.maute@gmx.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Flags| |needinfo?(pwu@redhat.com)
--- Comment #21 from a3emdot andre.maute@gmx.de --- (In reply to Peng Wu from comment #20)
Does this issue still happen with Fedora 34 or 35?
If it still happens, maybe you can report this issue to upstream.
I haven't upgraded to Fedora 35 yet. So I do still have Fedora 34. My Fedora was updated yesterday. And yes it still segfaults with the script pango-view-segfaults-20210923.sh
We still have # rpm -q --whatprovides $(which pango-view) pango-1.48.10-1.fc34.x86_64
The same as in september, so what is your rationale here that the segfault might have disappeared?
Most of the bugs in the issue tracker of pango upstream are quite old, do you think there is at least a chance to get this segfault fixed?
I would like to use pango-view as a tool in a script for diagnosing font problems, but at the moment the segfaults are really annoying.
Peng Wu, did you even manage to reproduce the segfaults?
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
Jens Petersen petersen@redhat.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Version|34 |35 Flags|needinfo?(pwu@redhat.com) |
--- Comment #22 from Jens Petersen petersen@redhat.com --- Thanks for the script.
It is pretty easy to reproduce, even by hand:
$ sudo dnf install google-roboto-condensed-fonts valgrind $ sudo dnf debuginfo-install pango $ valgrind pango-view --no-display --font="Roboto Condensed" --text="ABCDEFGHIJKLMNOPQRSTUVWXYZ" ==7889== Memcheck, a memory error detector ==7889== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al. ==7889== Using Valgrind-3.18.1 and LibVEX; rerun with -h for copyright info ==7889== Command: pango-view --no-display --font=Roboto\ Condensed --text=ABCDEFGHIJKLMNOPQRSTUVWXYZ ==7889== ==7889== ==7889== HEAP SUMMARY: ==7889== in use at exit: 298,288 bytes in 8,224 blocks ==7889== total heap usage: 24,814 allocs, 16,590 frees, 3,951,037 bytes allocated ==7889== ==7889== LEAK SUMMARY: ==7889== definitely lost: 0 bytes in 0 blocks ==7889== indirectly lost: 0 bytes in 0 blocks ==7889== possibly lost: 36 bytes in 1 blocks ==7889== still reachable: 295,452 bytes in 8,191 blocks ==7889== suppressed: 0 bytes in 0 blocks ==7889== Rerun with --leak-check=full to see details of leaked memory ==7889== ==7889== For lists of detected and suppressed errors, rerun with: -s ==7889== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 0 from 0) [liveuser@localhost-live ~]$ valgrind pango-view --no-display --font="Roboto Condensed" --text="ABCDEFGHIJKLMNOPQRSTUVWXYZ" ==7970== Memcheck, a memory error detector ==7970== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al. ==7970== Using Valgrind-3.18.1 and LibVEX; rerun with -h for copyright info ==7970== Command: pango-view --no-display --font=Roboto\ Condensed --text=ABCDEFGHIJKLMNOPQRSTUVWXYZ ==7970== pango-view: cairo-hash.c:217: _cairo_hash_table_destroy: Assertion `hash_table->live_entries == 0' failed. ==7970== ==7970== Process terminating with default action of signal 6 (SIGABRT): dumping core ==7970== at 0x4DE685C: __pthread_kill_implementation (pthread_kill.c:44) ==7970== by 0x4D996B5: raise (raise.c:26) ==7970== by 0x4D837D2: abort (abort.c:79) ==7970== by 0x4D836FA: __assert_fail_base.cold (assert.c:92) ==7970== by 0x4D923A5: __assert_fail (assert.c:101) ==7970== by 0x4C652DF: _cairo_hash_table_destroy (cairo-hash.c:217) ==7970== by 0x4C659FA: UnknownInlinedFun (cairo-scaled-font.c:441) ==7970== by 0x4C659FA: cairo_debug_reset_static_data (cairo-debug.c:67) ==7970== by 0x1120DC: pangocairo_view_destroy (viewer-pangocairo.c:93) ==7970== by 0x10ED76: main (viewer-main.c:176) ==7970== ==7970== HEAP SUMMARY: ==7970== in use at exit: 809,565 bytes in 9,884 blocks ==7970== total heap usage: 24,093 allocs, 14,209 frees, 3,862,588 bytes allocated ==7970== ==7970== LEAK SUMMARY: ==7970== definitely lost: 3,072 bytes in 5 blocks ==7970== indirectly lost: 24,353 bytes in 1,032 blocks ==7970== possibly lost: 1,996 bytes in 10 blocks ==7970== still reachable: 776,872 bytes in 8,803 blocks ==7970== suppressed: 0 bytes in 0 blocks ==7970== Rerun with --leak-check=full to see details of leaked memory ==7970== ==7970== For lists of detected and suppressed errors, rerun with: -s ==7970== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 0 from 0) Aborted (core dumped)
Should this be moved to cairo then?
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
Jens Petersen petersen@redhat.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Fixed In Version|pango-1.48.4-2.fc33 |
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
--- Comment #23 from Jens Petersen petersen@redhat.com --- (oh I tested that in a Fedora Live image, both F34 and F35 (pango-1.49.3-1.fc35))
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
--- Comment #24 from Peng Wu pwu@redhat.com --- It seems it is possible to reproduce this issue in VM with only one CPU assigned and without valgrind in Fedora 35.
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
--- Comment #25 from a3emdot andre.maute@gmx.de --- FYI, I used valgrind as a means to get more segmentation faults and stacktraces and to not be forced to resort to a debugger.
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
--- Comment #26 from Peng Wu pwu@redhat.com --- If we only assign one CPU to the pango-view process, the bug can be reproduced.
Here are the command line.
$ taskset -c 0 pango-view --no-display --dpi=72 --font="Roboto Condensed, 8" --antialias=none --backend=cairo --output="out8.png" --text="ABCDEFGHIJKLMNOPQRSTUVWXYZ"
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
Peng Wu pwu@redhat.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Link ID| |GNOME Gitlab | |GNOME/pango/-/merge_request | |s/521 Status|NEW |POST
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
Peng Wu pwu@redhat.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Link ID| |GNOME Gitlab | |GNOME/pango/-/merge_request | |s/528
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
--- Comment #27 from a3emdot andre.maute@gmx.de --- So if there is a patch for Fedora 35, will there be a a backport on top of pango-1.48.10 which is used in Fedora 34?
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
--- Comment #28 from Peng Wu pwu@redhat.com --- I think it depends on the bug.
For this crash bug, maybe I can backport it for Fedora 34 later...
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
--- Comment #29 from a3emdot andre.maute@gmx.de --- I can't reproduce this Bug anymore on Fedora 35, with a recent pango-1.50 rpm. Though an updated rpm for Fedora 34 would be highly appreciated on my side ;-)
Thank you very much Peng Wu for helping me out Andre
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
Fedora Update System updates@fedoraproject.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|POST |MODIFIED
--- Comment #30 from Fedora Update System updates@fedoraproject.org --- FEDORA-2021-32419016af has been submitted as an update to Fedora 34. https://bodhi.fedoraproject.org/updates/FEDORA-2021-32419016af
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
Fedora Update System updates@fedoraproject.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|MODIFIED |ON_QA
--- Comment #31 from Fedora Update System updates@fedoraproject.org --- FEDORA-2021-32419016af has been pushed to the Fedora 34 testing repository. Soon you'll be able to install the update with the following command: `sudo dnf upgrade --enablerepo=updates-testing --advisory=FEDORA-2021-32419016af` You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2021-32419016af
See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates.
https://bugzilla.redhat.com/show_bug.cgi?id=1993670
Fedora Update System updates@fedoraproject.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Fixed In Version| |pango-1.48.10-2.fc34 Status|ON_QA |CLOSED Resolution|--- |ERRATA Last Closed|2021-09-10 16:35:56 |2021-12-31 01:34:23
--- Comment #32 from Fedora Update System updates@fedoraproject.org --- FEDORA-2021-32419016af has been pushed to the Fedora 34 stable repository. If problem still persists, please make note of it in this bug report.
i18n-bugs@lists.fedoraproject.org