GNOME Bugzilla – Bug 755347
evolution-source-registry crashed with SIGABRT in g_assertion_message() (assertion hash_table->ref_count > 0 failed)
Last modified: 2019-11-12 17:37:55 UTC
Created attachment 311750 [details] Stacktrace with source We've got some reports of this crash in Launchpad and errors.ubuntu.com from 3.16 (this trace is from 3.16.5 but there are some reports from 3.16.4 - this looks like the first affected version AFAICS). I don't know what the cause is and I haven't yet reproduced a crash but I asked the reporter over there - if he comes back then I'll update this bug. In the meantime maybe the trace is useful to you. There's a comment in ghash.c around the assertion we're hitting: /* If this happens, then the application is probably doing too much work * from a destroy notifier. The alternative would be to crash any second * (as keys, etc. will be NULL). * Applications need to either use g_hash_table_destroy, or ensure the hash * table is empty prior to removing the last reference using g_hash_table_unref(). */ g_assert (hash_table->ref_count > 0);
Thanks for a bug report. The code in the backtrace had been added within bug #749135, which is in 3.16.3 and beyond. I do not understand why this is the first instance of the crash, when it is out for so long (I would guess it'll strike much sooner, but it can be it has only something to do with Ubuntu Online Accounts?). Could you ask the reporter: * what accounts he/she has configured, whether any are from Ubuntu Online Accounts or GNOME Online Accounts * whether it crashes also when he/she runs the evolution-source-registry from a terminal manually * whether running the source registry under valgrind will also reproduce the crash (it can happen that other timing, due to all slowness caused by the memory checking, will not trigger the crash, but also that the valgrind will log about the issue and will keep the application running). The command might look like (I do not know exact location of the executable, thus it can differ in your distribution): $ G_SLICE=always-malloc valgrind /usr/libexec/evolution-source-registry &>log please? I'm pasting the bcaktrace for better searching: #0 0x00007f00a3f5f267 in __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:55 [Error: raise.c was not found in source tree]
+ Trace 235483
Ah, I just looked and .4 was the first release in the 3.16 series that we shipped (starting on July 20). We have "only" 50 reports of this so I didn't notice until now, sorry about that. I proxied your questions.
Unsure if that's the same issue bug I reported and invalid read error triggered by turning ubuntu online accounts off/on, see bug #755410
Closing this bug report as no further information has been provided. Please feel free to reopen this bug report if you can provide the information that was asked for in a previous comment. Thanks!