Summary: | Kontact crash while away from the computer | ||
---|---|---|---|
Product: | [Plasma] Oxygen | Reporter: | Scott Kitterman <kde> |
Component: | style | Assignee: | Hugo Pereira Da Costa <hugo.pereira.da.costa> |
Status: | RESOLVED UPSTREAM | ||
Severity: | crash | CC: | georg.wittenburg, hugo.pereira.da.costa, pragalathanm, qpmvh |
Priority: | NOR | ||
Version: | unspecified | ||
Target Milestone: | --- | ||
Platform: | Ubuntu | ||
OS: | Linux | ||
Latest Commit: | Version Fixed In: |
Description
Scott Kitterman
2011-04-14 23:36:02 UTC
*** Bug 273736 has been marked as a duplicate of this bug. *** *** Bug 274106 has been marked as a duplicate of this bug. *** mmm. I guess I'll cache oxygen's arrow rendering. well. Maybe not. Looking at the code at: #14 0xb6a217c9 in QVector<QPointF>::realloc (this=0xbfefd45c, asize=67108863, aalloc=134217727) at ../../include/QtCore/../../src/corelib/tools/qvector.h:487 Turns out the crash corresponds to a rethrown "std::bad_alloc" exception. This is likely due to the memory being full. Either dues to some memory leak somewhere (and last time it was checked with valgrind, there is zero leak in oxygen), or just the memory being used 100% by some other process (e.g. nepomuk). This is likely not an oxygen bug at all. On Wednesday 25 May 2011 15:08:00 Hugo Pereira Da Costa wrote:
> [...]
> This is likely not an oxygen bug at all.
Hmmm.... In that case, I should be seeing similar crashes if I switch to
another widget/window theme, right? I'll do that and see if the problems
persist.
Best,
Georg
yes. Likely. Though oxygen does use more (base and properly freed) memory than most other styles ... Keep me posted I haven't seen a single crash (usually I get about one per hour; also reported in bug 274096) since I switched to other widgets/styles. Specifically, I switched: - Application Appearance -> Style -> Widget Style to Plastique - Application Appearance -> GTK+ Appearance -> Widget style to Raleigh - Workspace Appearance -> Window Decorations to Plastik From this, it seems save to conclude that the crashes are related to Oxygen. Note that memory shouldn't be an issue since this machine has 3GB of RAM. I will now switch back the items above to Oxygen and see whether I can narrow things down. (Aside: I'll be traveling next week, so there'll be less feedback from my side.) @Georg Thanks for the info and the testing. To be honest I have no idea on how to adress this bug. Most likely something is corrupting the memory, creating the crashes, triggered by oxygen (but not _in_ oxygen). Now, oxygen has been run with valgrind to check memory usage (via oxygen-demo) multiple times, including recently, and no issues have been found ... So I'm unclear about what to do next. maybe running kontact through valgrind ? (I don't use kontact atm and completely fubared its install last time I attempted to give a shot to kdepim 4.6). Suggestions welcome. (and if anyone can provide a valgrind output of kontact, I volunteer to look into it). After more time with this new laptop, I'm now pretty certain that it's an X or Intel i915 driver issue. Evidence is that large batch processes crash, but only when run from an X terminal window, not if run with X disabled. Sorry for the trouble. :) No trouble taken. Thanks for closing the bug, and feel free to re-open if new clues make you think otherwise ! Feedback's always welcome. *** Bug 281447 has been marked as a duplicate of this bug. *** *** Bug 271022 has been marked as a duplicate of this bug. *** |