| Summary: | kdevelop crash while editing file | ||
|---|---|---|---|
| Product: | [Applications] kdevelop | Reporter: | Matthew Woehlke <mwoehlke.floss> |
| Component: | general | Assignee: | kdevelop-bugs-null |
| Status: | RESOLVED WAITINGFORINFO | ||
| Severity: | crash | CC: | aleixpol, Ekkehard.Blanz, mwoehlke.floss |
| Priority: | NOR | ||
| Version First Reported In: | unspecified | ||
| Target Milestone: | 4.2.0 | ||
| Platform: | Fedora RPMs | ||
| OS: | Linux | ||
| Latest Commit: | Version Fixed In: | ||
| Sentry Crash Report: | |||
| Attachments: | New crash information added by DrKonqi | ||
|
Description
Matthew Woehlke
2011-04-28 19:19:35 UTC
Update: it is definitely reproducible in some cases (just had it happen again, same stack trace), doing almost exactly the same thing except using triple-click to select lines before paste (so no cursor movement except the first click). Possible method to reproduce: - Put a comment on the clipboard, e.g. '//----------------' - Open a medium-large C++ source file (say 1k+ lines) - Triple-click to select first blank line - Paste (Ctrl+V) - Repeat above two steps That said, I just went through almost a dozen files with this procedure with no hiccups. It may require larger files, or may be somehow related to the file contents (though it was two different files I've seen this happen with). Possibly it happens if katepart is still parsing one update when the second one is triggered? (...i.e. just needs a really large file, and possibly changes being made toward the beginning.) Is it still reproducible with a newer Kate Part? Hmm... can't say I've seen it happen in a while. (Also not sure I've been doing the sorts of edits that we're causing it, but gave the steps in comment #2 a whirl, and nothing...) So I guess I'll close it, and reopen / refile if it shows up again. Thanks. Created attachment 68603 [details]
New crash information added by DrKonqi
kdevelop (4.2.3) on KDE Platform 4.6.5 (4.6.5) using Qt 4.7.4
Just saw this again (same circumstances as original report; stack track appears similar).
@Aleix, is 4.6.5 sufficiently newer to be interesting? (I can't upgrade past F14 right now due to other software I need for work that does not build on >F14.)
-- Backtrace (Reduced):
#6 0x00007f1e152ccc88 in Kate::TextBuffer::rangesForLine (this=<value optimized out>, line=516, view=0x3e13ab0, rangesWithAttributeOnly=false) at /usr/src/debug/kdelibs-4.6.5/kate/buffer/katetextbuffer.cpp:752
#7 0x00007f1e153a2bfe in KateView::updateRangesIn (this=0x3e13ab0, activationType=KTextEditor::Attribute::ActivateMouseIn) at /usr/src/debug/kdelibs-4.6.5/kate/view/kateview.cpp:2767
#8 0x00007f1e153b5f80 in KateViewInternal::mouseMoveEvent (this=0x11018230, e=0x7fffe3f211e0) at /usr/src/debug/kdelibs-4.6.5/kate/view/kateviewinternal.cpp:2708
#9 0x000000333fa0b710 in QWidget::event (this=0x11018230, event=0x7fffe3f211e0) at kernel/qwidget.cpp:8280
#10 0x000000333f9b9c64 in QApplicationPrivate::notify_helper (this=0x2292f40, receiver=0x11018230, e=0x7fffe3f211e0) at kernel/qapplication.cpp:4481
you should either give us a way to reproduce it or use up2date kate to make sure tihs is really still an issue Fair enough... unfortunately I can't *reliably* reproduce it, I just know I've seen it a handful of times when doing roughly the same thing. There's a good chance it's timing dependent, also. Best I can do right now as far as 'how to reproduce' would be this: 1. Open a moderately large project in KDevelop (say, 100+ files). 2. Open a moderately large (1000+ lines) source file within said project. 3. Stick a comment (including newline) on the clipboard. 4. Replace every blank line in the file with this comment (not using search/replace). I've probably only done something like the above a half dozen or so times, but I've had KDevelop crash probably 4-5+ times when doing so. (Of course, when I'm *trying* to reproduce it, I can't :-(...) (Oh, and... *thank you* for at least looking at this, and don't lose too much sleep over it :-). This isn't a huge issue for me, and I understand your frustration being unable to reproduce it; I've been there plenty of times.) but you are sure you are using up2date kate? your last backtrace points at kde 4.6.5... or is that kate checkout from current git? The line 752 in katetextbuffer.cpp also points to a different function in git now, so I'm really wondering whether you actually use a up2date kate... *** Bug 295092 has been marked as a duplicate of this bug. *** Still in at least KDE 4.7.4 (see bug #295092) |