Bug 1209787
| Summary: | Rebase libical to 1.0.1 | ||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | Red Hat Enterprise Linux 7 | Reporter: | David Jaša <djasa> | ||||||||||||||||||||||||||
| Component: | libical | Assignee: | Matthew Barnes <mbarnes> | ||||||||||||||||||||||||||
| Status: | CLOSED ERRATA | QA Contact: | Desktop QE <desktop-qa-list> | ||||||||||||||||||||||||||
| Severity: | high | Docs Contact: | |||||||||||||||||||||||||||
| Priority: | medium | ||||||||||||||||||||||||||||
| Version: | 7.1 | CC: | djasa, jkoten, lmiksik, mbarnes, mcrha, ToddAndMargo, tpelka, vrutkovs | ||||||||||||||||||||||||||
| Target Milestone: | rc | Keywords: | Rebase | ||||||||||||||||||||||||||
| Target Release: | --- | ||||||||||||||||||||||||||||
| Hardware: | x86_64 | ||||||||||||||||||||||||||||
| OS: | Unspecified | ||||||||||||||||||||||||||||
| URL: | http://faf-report.itos.redhat.com/reports/bthash/55c947579e3e2de9ad353e011624f09815e8423d | ||||||||||||||||||||||||||||
| Whiteboard: | abrt_hash:2023742d58ab4ceb254b740d5969487db8c722e2 | ||||||||||||||||||||||||||||
| Fixed In Version: | libical-1.0.1-1.el7 | Doc Type: | Rebase: Bug Fixes and Enhancements | ||||||||||||||||||||||||||
| Doc Text: | Story Points: | --- | |||||||||||||||||||||||||||
| Clone Of: | Environment: | ||||||||||||||||||||||||||||
| Last Closed: | 2015-11-19 07:57:56 UTC | Type: | --- | ||||||||||||||||||||||||||
| Regression: | --- | Mount Type: | --- | ||||||||||||||||||||||||||
| Documentation: | --- | CRM: | |||||||||||||||||||||||||||
| Verified Versions: | Category: | --- | |||||||||||||||||||||||||||
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||||||||||||||||||||||||
| Cloudforms Team: | --- | Target Upstream Version: | |||||||||||||||||||||||||||
| Embargoed: | |||||||||||||||||||||||||||||
| Attachments: |
|
||||||||||||||||||||||||||||
|
Description
David Jaša
2015-04-08 08:34:52 UTC
Created attachment 1012083 [details]
File: backtrace
Created attachment 1012084 [details]
File: cgroup
Created attachment 1012085 [details]
File: core_backtrace
Created attachment 1012086 [details]
File: dso_list
Created attachment 1012087 [details]
File: environ
Created attachment 1012088 [details]
File: limits
Created attachment 1012089 [details]
File: maps
Created attachment 1012090 [details]
File: open_fds
Created attachment 1012091 [details]
File: proc_pid_status
Created attachment 1012092 [details]
File: var_log_messages
Created attachment 1012093 [details]
File: binary
Created attachment 1012109 [details]
invite that makes evolution crash
attaching the "guilty" invite
I added the invitation to the calendar twice: after first addition, gnome-shell was also crashing with first four frames being the same. ABRT found a report on retrace server for that crash: http://faf-report.itos.redhat.com/reports/6070/ A note for reproducers: I needed to remove the appointment in web UI and restart evolution-calendar-factory in order to stop evolution crashing... *** Bug 1186744 has been marked as a duplicate of this bug. *** David was able to reproduce this reliably, the trick might be to have a calendar with events which have set different timezones. As a valgrind log shows: Thread 1: Invalid read of size 8 at 0x6BD33F0: icaltimezone_get_utc_offset_of_utc_time (icaltimezone.c:965) by 0x6BD35F4: icaltimezone_convert_time (icaltimezone.c:778) by 0x6BCEA2D: icaltime_from_timet_with_zone (icaltime.c:226) by 0x36D1551B: e_day_view_add_event (e-day-view.c:5417) by 0x36D15947: process_component (e-day-view.c:2399) by 0x36D15B45: model_rows_inserted_cb (e-day-view.c:2504) by 0x1329DDAB: ffi_call_unix64 (unix64.S:76) by 0x1329D6D4: ffi_call (ffi64.c:522) by 0xEF2D627: g_cclosure_marshal_generic (gclosure.c:1448) by 0xEF2CDE7: g_closure_invoke (gclosure.c:768) by 0xEF3E70C: signal_emit_unlocked_R (gsignal.c:3553) by 0xEF46470: g_signal_emit_valist (gsignal.c:3309) Address 0x7c3ceea8 is 24,168 bytes inside a block of size 29,952 free'd at 0x4C2AD17: free (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) by 0x6BC1B8C: icalarray_expand (icalarray.c:183) by 0x6BC1B8C: icalarray_append (icalarray.c:113) by 0x6BD3A68: icaltimezone_get_builtin_timezone (icaltimezone.c:1395) by 0x697CF98: e_cal_match_location (e-cal-check-timezones.c:46) by 0x697CF98: e_cal_match_tzid (e-cal-check-timezones.c:120) by 0x6962257: cal_client_get_cached_timezone (e-cal-client.c:1352) by 0x6969C2C: e_cal_client_get_timezone_sync (e-cal-client.c:6658) by 0x696AE59: e_cal_client_generate_instances_for_object (e-cal-client.c:2981) by 0x36D5B7C1: tag_calendar_by_comp (tag-calendar.c:258) by 0x36D4B906: gnome_cal_objects_added_cb (gnome-cal.c:919) by 0xEF2FF7C: g_cclosure_marshal_VOID__POINTERv (gmarshal.c:1236) by 0xEF2D016: _g_closure_invoke_va (gclosure.c:831) by 0xEF45AD6: g_signal_emit_valist (gsignal.c:3218) After an investigation, libical 0.48 suffers of this issue, while libical 1.0.1 doesn't (the one in Fedora 22). I'd prefer to rebase libical to 1.0.1. Just in case, the rebase also means a soname version bump of libical libraries, thus it requires rebuild of all dependencies against that new version. Created attachment 1049192 [details] libical-0.48-increase-timezone-array-expand.patch An alternative workaround for the crash is to preallocate the timezone array sufficiently, which does this patch. This doesn't change any API, only internal things, thus instead of rebase, an update would be sufficient and no rebuilds of other packages needed. Here [1] is built a test libical package with the patch applied. [1] https://brewweb.devel.redhat.com/taskinfo?taskID=9476335 Built packages: libical-1.0.1-1.el7 bluez-5.23-2.el7 kdepimlibs-4.10.5-4.el7 openchange-2.0-8.el7 evolution-data-server-3.12.11-19.el7 There is missing gnome-shell, evolution, evolution-ews and evolution-mapi, because the build of evolution brings in gnome-shell (somehow through dependencies), and the gnome-shell cannot be currently built, because it requires gnome-shell (I do not understand the error there and why it was brought in). This is part of the root.log from the gnome-shell rebuild attempt: DEBUG util.py:257: --> 1:gnome-bluetooth-libs-devel-3.14.1-1.el7.x86_64 DEBUG util.py:257: --> 1:control-center-3.14.5-6.el7.x86_64 DEBUG util.py:257: --> gtk-doc-1.19-3.el7.noarch DEBUG util.py:257: --> gnome-common-3.7.4-4.el7.noarch DEBUG util.py:257: Error: Package: gnome-shell-3.14.4-13.el7.x86_64 (build) DEBUG util.py:257: Requires: libicalss.so.0()(64bit) DEBUG util.py:257: Error: Package: gnome-shell-3.14.4-13.el7.x86_64 (build) DEBUG util.py:257: Requires: libicalvcal.so.0()(64bit) DEBUG util.py:257: Error: Package: gnome-shell-3.14.4-13.el7.x86_64 (build) DEBUG util.py:257: Requires: libical.so.0()(64bit) DEBUG util.py:257: You could try using --skip-broken to work around the problem DEBUG util.py:257: You could try running: rpm -Va --nofiles --nodigest I think I know what the problem is. When the gnome-shell is built, the evolution-data-server brings in rebased libical, but the already installed gnome-shell needs older version of libical, which fails here. Thus, I think, the problem is that the brew has installed gnome-shell, when it builds gnome-shell. Kalev pointer out that it could be libnotify package bringing in gnome-shell, and he was right. I tagget the updated libnotify for build and the gnome-shell can be built now. The other built packages are: gnome-shell-3.14.4-15.el7 evolution-3.12.11-10.el7 evolution-mapi-3.12.10-3.el7 evolution-ews-3.12.11-5.el7 Scientific Linux 7.1 They old version of libical is affecting both Osmo and Krusader: https://sourceforge.net/p/osmo-pim/bugs/99/ https://bugs.kde.org/show_bug.cgi?id=332537 It has to do with a dependency for libical.so.1 (in 1.0.1) The solution was to rebuild Fedora Core 22's SRPM. And for some odd reason, "rpm" allows me to install (rpm -i) both side by side, but will not allow me to do and upgrade (rpm -U): $ rpm -qa libical libical-0.48-6.el7.x86_64 libical-1.0.1-1.el7.x86_64 (In reply to Todd from comment #31) > It has to do with a dependency for libical.so.1 (in 1.0.1) > > The solution was to rebuild Fedora Core 22's SRPM. Right, the rebase bumped soname versions, thus anything what depends on it requires rebuild. Standard RHEL packages had been rebuild. 3-rd party packages need rebuild "on demand". > And for some odd reason, "rpm" allows me to install (rpm -i) both side by > side, but will not allow me to do and upgrade (rpm -U): Sounds like a bug in the rpm. I'm not able to install older libical when I have installed the newer (and downgrade of the whole system isn't feasible for me). Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://rhn.redhat.com/errata/RHBA-2015-2226.html |