GNOME Bugzilla – Bug 104936
Frequent AisleRiot early exit when winning
Last modified: 2012-01-31 23:28:10 UTC
Klondike FREQUENTLY crashes when approaching a winning hand, usually before completeing the game, but occasionally after punching button to start a new subsequent game. I have NEVER seen this when losing. It used to work SOLID with Red Hat 7.0, so its discouraging for an update to Red Hat 8.0. On rare occasions I can complete a win and successfully start a new game. Sounds related to bug 104444. Might be related to double clicking, and very slow double clicking at end seems to "sneak up on a win", ... sometimes.
Doesn't sound like bug 104444 to me, since that bug report says there is no crash while you say there is one. Since there's a crash, can you get us a stack trace? (See http://bugzilla.gnome.org/getting-traces.cgi for information on how to do this) Setting severity->critical (crashes are critical, not blocker unless a maintainer or someone else says so) & priority->high (again, it's a crasher). Marking as NEEDINFO until we can get a stack trace; also adding bugsquad keyword.
OK, maybe this is 2 bugs each (bug 104936 and bug 104444 ). I got to this bug page from the Error Message Pop Up window which said --------------------------- Application "sol" (process 1603) has crashed due to fatal error (Segmentation fault). --------------------------- Since the bug buddy was still up I attempted to complete it and get a stack trace as requested, but got none, probably because I was NOT running as root as required of a game. This full crash happens very infrequently, maybe once a month, so it may be a while to reproduce. So I opened a terminal window, did "su", then "gdb /usr/bin/sol", then "run". I played a couple losing games normally, then finally got close to winning one. The behavior is the more frequently observered behavior of bug 104444. At the point of ALMOST winning, the window disappeared and I got the following on "gdb": --------------------------- ERROR: In procedure car: ERROR: Wrong type argument in position 1: () Program exited with code 02. (gdb) bt No stack. (gdb) --------------------------- I was able to get this same end message 3 times, and in addition was able to win 2 and restart! Of course, many normal loses in between (I'm not that good). This is the more annoying of the 2 behaviors because it happens on almost every NEAR WIN. Well, 3 out of 5 this go around. Anyway, I'll keep trying to get a stack trace, and will post it if I get one in the "SU" mode with "gdb on".
Thanks for the quick follow up. If you can keep trying occasionally and let us know when you get a stack trace, that would be great. In the mean time, I'll reopen this bug so the developers can have a chance to look at it in case the information you've provided so far might be helpful. Also, I'm going to mark bug 104444 as a duplicate of this one, since this bug now contains all the information in that one plus more.
...and I'll mark as NEW to put this bug on the radar.
*** Bug 104444 has been marked as a duplicate of this bug. ***
Hi, this is Tom from bug 104444. I started sol in my own shell. It crashed twice in a row again. Both times I was winning :) There's no core because it seems that this is a scheme script error with klondike or something. So yeah, definately mine is a dup of this. tom> sol ERROR: In procedure car: ERROR: Wrong type argument in position 1: ()
Infrequent gives the WRONG impression. It's frequent, usually 60% to 90% of the NEAR WINS. So, maybe the word "crash" was the word to change, if you limit "crash" to clinical "segmentation faults". "Psychologically crashing" when NEAR a win is FREQUENT. Apparently losing is "bullet proof".
One more tidbit of information on the relationship of the "two" apparent bugs ("early exit" vs."segmentation fault:). I now got 2 segmentation faults in the last 24 hours, but unfortunately I was not "super user" for either one so the Bug Buddy didn't give a stack dump. But I noticed that both "segmentation faults" happened when trying to restart a new game after completing a winning game! There are 3 possible outcomes of restarting a new game after winning a game BUT ONLY IF THE FIRST DOESN"T TAKE AN EARLY EXIT, which is the MOST FREQUENT OUTCOME! 1) The signature ERROR messages and exit recorded above. 2) A segmentation fault. 3) A successful new game It is REALLY hard to get past the FIRST early exit, to even try one of these! And #3 is REALLY rare! Again, losing is "bullet proof". You can go on forever. And it seems that I have proven that many times.
THE CRASH STACK DUMP !!! This is longer than one comment pane. It's continued on next comment. ------------------------------------------- Backtrace was generated from '/usr/bin/sol' (no debugging symbols found)...[New Thread 8192 (LWP 1059)] 0x420ae169 in wait4 () from /lib/i686/libc.so.6
+ Trace 33421
Stack dump continued from previous comment ---------------------------------------
+ Trace 33423
Thread 1 (Thread 8192 (LWP 1059))
This is bug #89822 (the stack trace and the scheme code match). *** This bug has been marked as a duplicate of 89822 ***
This bug is being reassigned to the "general" component so we can close the aisleriot bugzilla component. Apologies for the mass email!