GNOME Bugzilla – Bug 331966
Gnome-games crashes
Last modified: 2012-01-31 23:19:18 UTC
From: acidborg <> To: submit@bugs.gnome.org X-Mailer: bug-buddy 2.13.0 Subject: Gnome-games crashes Distribution: Ubuntu 6.04 (dapper) Package: gnome-games Severity: Normal Version: GNOME2.13.91 2.13.7 Gnome-Distributor: Ubuntu Synopsis: Gnome-games crashes Bugzilla-Product: gnome-games Bugzilla-Component: aisleriot Bugzilla-Version: 2.13.7 BugBuddy-GnomeVersion: 2.0 (2.13.3) Description: Description of the crash: I just open Freecell and start playing. It just crashed Steps to reproduce the crash: 1. Open Freecell 2. Start playing 3. It Crashes Expected Results: Not to crash How often does this happen? Always Additional Information: Ubuntu Dapper updated today Debugging Information: Backtrace was generated from '/usr/games/sol' (no debugging symbols found) Using host libthread_db library "/lib/tls/i686/cmov/libthread_db.so.1". (no debugging symbols found) [Thread debugging using libthread_db enabled] [New Thread -1224489280 (LWP 3696)] [New Thread -1246401616 (LWP 3701)] [New Thread -1238008912 (LWP 3700)] 0xffffe410 in __kernel_vsyscall ()
+ Trace 66364
Thread 1 (Thread -1224489280 (LWP 3696))
------- Bug created by bug-buddy at 2006-02-21 01:01 ------- Unknown version 2.13.7 in product gnome-games. Setting version to "2.13.x".
I can't reproduce this immediately, but the stack trace is, "interesting". Is this a plain x86 machine or one with 64-bit extensions? The stack trace seems to say that the problem has a lot to do with the qthreads library that guile uses. Although the stack trace is a little odd, probably the result of the magic that qthreads does to get threading working (e.g. the qt_null call at the top is, according to the source, an empty function that calls nothing, let alone coop_abort). The problem is either in the Freecell hint code (which is the only bit in aisleriot to use threading) or a bug in guile. The coop_abort call sort of suggests the later, but I couldn't be sure.
> Is this a plain x86 machine or one with 64-bit extensions? It is a plain AMD x86 machine. It was fixed in the last update of Ubuntu Dapper. Thanks anyway.
The problem isn't exactly fixed in Dapper, they just went back to the old way of doing things since the code isn't entirely portable without upgrading the guile 1.8. I'm considering doing the same thing.
I have reverted the threaded-hint code, just as Ubuntu have done. This should fix this issue at the expense of possible responsiveness losses on slow machines.
This bug is being reassigned to the "general" component so we can close the aisleriot bugzilla component. Apologies for the mass email!