Description of problem: https://bugzilla.gnome.org/show_bug.cgi?id=795740#c9 contains an HTML reproducer to make Epiphany (web browser) crash. When running 'bt full' on the resulting core dump, gdb itself crashes. Version-Release number of selected component (if applicable): 8.1-15.fc28 How reproducible: Always Steps to Reproduce: 1. Build epiphany git master with -O0 -g 2. Open the crash testcase in epiphany 3. 'coredumpctl gdb' 4. 'bt full' Umm... is there a way I can help you reproduce this that doesn't involve compiling epiphany? I have uploaded the core file to https://people.gnome.org/~mcatanzaro/epiphany-core (for a few days only) but I don't think this will actually be useful without debuginfo? Anything I should run to get that for you? Actual results: gdb segfaults Expected results: No segfaults, except from Epiphany Additional info:
Actually, no need to do anything. We thought we had a fix in the pipeline, but it turns out not to be sufficient. However, thanks to your report of 1574015, I have a reproducer now on Fedora 28, and I am currently tracking down the root of the problem. I see that 1560010 and all its dups (including 1574015) have been closed. I am leaving this one open to track. However, Sergio may choose to re-open 1560010 and dup this bug.
(In reply to Keith Seitz from comment #1) > Actually, no need to do anything. We thought we had a fix in the pipeline, > but it turns out not to be sufficient. > > However, thanks to your report of 1574015, I have a reproducer now on Fedora > 28, and I am currently tracking down the root of the problem. Note that, in that bug, I saw an error message that I don't see here, so I assumed it was a different issue: ../../gdb/dictionary.c:690: internal-error: void insert_symbol_hashed(dictionary*, symbol*): Assertion `SYMBOL_LANGUAGE (sym) == DICT_LANGUAGE (dict)->la_language' failed. Anyway, I guess I can delete the core file from the server now? You don't seem to need it?
I've grabbed the core file. I will add this bug to my TODO list.
Keith, do you think we can mark this bug as duplicate of bug 1560010? Or the other way around? I don't really have a preference, but I'm trying to clean things up a little bit. Thanks!
(In reply to Sergio Durigan Junior from comment #4) > Keith, do you think we can mark this bug as duplicate of bug 1560010? Or > the other way around? I don't really have a preference, but I'm trying to > clean things up a little bit. Thanks! Yes, this is a duplicate to 1560010 and the (numerous) other similar bugs. I'm sure you will be as happy as I to see he upstream patch approved!
I'll hesitantly mark it as a duplicate, noting: (In reply to Michael Catanzaro from comment #2) > Note that, in that bug, I saw an error message that I don't see here, so I > assumed it was a different issue: > > ../../gdb/dictionary.c:690: internal-error: void > insert_symbol_hashed(dictionary*, symbol*): Assertion `SYMBOL_LANGUAGE (sym) > == DICT_LANGUAGE (dict)->la_language' failed. *** This bug has been marked as a duplicate of bug 1560010 ***
Also note the upstream Bugzilla comment: "With the fix applied, I still get the error, so this does not seem solved." https://sourceware.org/bugzilla/show_bug.cgi?id=23010#c5
(In reply to Michael Catanzaro from comment #7) > Also note the upstream Bugzilla comment: > > "With the fix applied, I still get the error, so this does not seem solved." > > https://sourceware.org/bugzilla/show_bug.cgi?id=23010#c5 That's a different solution to a different, but related, problem. See the upstream discussion: https://sourceware.org/ml/gdb-patches/2018-05/msg00703.html