[Touch-packages] [Bug 1812955] Re: g-ir-scanner/Shell-0.1 is crashing with SIGSEGV instead of succeeding

Fixed upstream: https://gitlab.gnome.org/GNOME/mutter/merge_requests/412 ** Changed in: gobject-introspection (Ubuntu) Status: Confirmed => Invalid -- You received this bug notification because you are a member of Ubuntu Touch seeded packages, which is subscribed to

[Touch-packages] [Bug 1812955] Re: g-ir-scanner/Shell-0.1 is crashing with SIGSEGV instead of succeeding

** Changed in: gobject-introspection (Ubuntu) Status: Invalid => Confirmed ** Description changed: + https://gitlab.gnome.org/GNOME/gobject-introspection/issues/260 + + --- + g-ir-scanner is crashing with SIGSEGV instead of succeeding. So I can't build gnome-shell any more.

[Touch-packages] [Bug 1812955] Re: g-ir-scanner/Shell-0.1 is crashing with SIGSEGV instead of succeeding

Finally bisected. It seems like a regression in upstream mutter master this week: 1bbb5c8107ab77435f97c1b037a89a76121f6e3c is the first bad commit commit 1bbb5c8107ab77435f97c1b037a89a76121f6e3c Author: Jonas Ådahl Date: Sun Jan 20 17:41:53 2019 +0100 meta: Only declare types when

[Touch-packages] [Bug 1812955] Re: g-ir-scanner/Shell-0.1 is crashing with SIGSEGV instead of succeeding

And a fresh retrace today: Core was generated by `/home/dan/git/gnome-shell/build/tmp-introspectcxwjenqm/Shell-0.1 --introspect-d'. Program terminated with signal SIGSEGV, Segmentation fault. #0 0x7f0586db711f in type_check_is_value_type_U ( type=type@entry=1694390208) at

[Touch-packages] [Bug 1812955] Re: g-ir-scanner/Shell-0.1 is crashing with SIGSEGV instead of succeeding

The same crash is happening upstream: https://gitlab.gnome.org/GNOME/gnome-build-meta/-/jobs/194604 -- You received this bug notification because you are a member of Ubuntu Touch seeded packages, which is subscribed to gobject-introspection in Ubuntu. https://bugs.launchpad.net/bugs/1812955

[Touch-packages] [Bug 1812955] Re: g-ir-scanner/Shell-0.1 is crashing with SIGSEGV instead of succeeding

Strangely, it doesn't crash under valgrind. Although valgrind does mention: ==26043== Warning: set address range perms: large range [0x2bf9a654a000, 0x2bf9e654a000) (noaccess) ==26043== Warning: set address range perms: large range [0x2bf9a654a000, 0x2bf9e654a000) (noaccess) -- You received

[Touch-packages] [Bug 1812955] Re: g-ir-scanner/Shell-0.1 is crashing with SIGSEGV instead of succeeding

Using GI_SCANNER_DEBUG=save-temps I can at least see the offending binary. So here's more useful information: Thread 1 "Shell-0.1" received signal SIGSEGV, Segmentation fault. 0x77bb911f in type_check_is_value_type_U ( type=type@entry=) at ../../../../gobject/gtype.c:4156 4156

[Touch-packages] [Bug 1812955] Re: g-ir-scanner/Shell-0.1 is crashing with SIGSEGV instead of succeeding

Annoyingly, many packages changed today. I think the regression occurred somewhere in here: Start-Date: 2019-01-23 10:59:11 Commandline: apt full-upgrade Requested-By: dan (1000) Install: libsereal-encoder-perl:amd64 (4.005+ds-1build1, automatic), libio-async-perl:amd64 (0.72-1, automatic),