| Summary: | Kdevelop crashes when opening a document ([re]storing split views) | ||
|---|---|---|---|
| Product: | [Applications] kdevelop | Reporter: | Jonathan Verner <jonathan.verner> |
| Component: | general | Assignee: | kdevelop-bugs-null |
| Status: | RESOLVED FIXED | ||
| Severity: | crash | Keywords: | drkonqi |
| Priority: | NOR | ||
| Version First Reported In: | git master | ||
| Target Milestone: | --- | ||
| Platform: | Neon | ||
| OS: | Linux | ||
| Latest Commit: | https://invent.kde.org/kdevelop/kdevelop/commit/50eaff66056898d039047f54290b3e5d8f09594f | Version Fixed In: | |
| Sentry Crash Report: | |||
| Attachments: | New crash information added by DrKonqi | ||
Created attachment 140779 [details]
New crash information added by DrKonqi
DrKonqi auto-attaching complete backtrace.
I think the code is fine. Note how that area is trying to find the parentSplitter as a parent of the widget. So if we don't have a widget, we don't have a parent and thus cannot find the splitter. Can you please submit this patch? A possibly relevant merge request was started @ https://invent.kde.org/kdevelop/kdevelop/-/merge_requests/247 A possibly relevant merge request was started @ https://invent.kde.org/kdevelop/kdevelop/-/merge_requests/247 (In reply to Milian Wolff from comment #2) > I think the code is fine. Note how that area is trying to find the > parentSplitter as a parent of the widget. So if we don't have a widget, we > don't have a parent and thus cannot find the splitter. > > Can you please submit this patch? Milian, thanks for the quick reaction. I've created a merge request on invent.kde.org: https://invent.kde.org/kdevelop/kdevelop/-/merge_requests/247 Git commit 50eaff66056898d039047f54290b3e5d8f09594f by Jonathan L. Verner. Committed on 17/08/2021 at 11:58. Pushed by mwolff into branch 'master'. Fix crash when storing working sets area split-view config When storing working sets commit 89594aed introduced a crash, which happens when the code tries to find the parent splitter while the view doesn't have an initialized widget. In this case calling `view->widget()` method tries to create a new widget, but fails with an assert in `EditorViewWatcher::viewCreated` because no parent for the widget is given. M +1 -1 kdevplatform/shell/workingsets/workingset.cpp https://invent.kde.org/kdevelop/kdevelop/commit/50eaff66056898d039047f54290b3e5d8f09594f |
Application: kdevelop (5.6.40) Qt Version: 5.15.3 Frameworks Version: 5.84.0 Operating System: Linux 5.8.0-63-generic x86_64 Windowing System: X11 Drkonqi Version: 5.22.4 Distribution: KDE neon User Edition 5.22 -- Information about the crash: - What I was doing when the application crashed: I start kdevelop, open a project and then click on any file in the project. Kdevelop immediately crashes. The crash seems to be caused by the code which takes care of (re)storing working sets, introduced in commit 89594aed, in particular on line 278 in `kdevplatform/shell/workingsets/workingset.cpp` the call to `view->widget()` fails with an assert. The problem is that the view does not have an initialized widget, so the `Sublime::View::widget` method tries to create one. However, its not given any parent, which leads to an assert in `EditorViewWatcher::viewCreated` deeper in the callstack. The following patch seems to fix the problem for me: ```diff diff --git a/kdevplatform/shell/workingsets/workingset.cpp b/kdevplatform/shell/workingsets/workingset.cpp index 60c0946446..45b8c7671a 100644 --- a/kdevplatform/shell/workingsets/workingset.cpp +++ b/kdevplatform/shell/workingsets/workingset.cpp @@ -274,7 +274,7 @@ QSplitter* saveFromAreaPrivate(Sublime::AreaIndex *area, KConfigGroup setGroup, } ++index; - if (!parentSplitter) { + if (!parentSplitter && view->hasWidget()) { auto p = view->widget()->parentWidget(); while (p && !(parentSplitter = qobject_cast<QSplitter*>(p))) { p = p->parentWidget(); ``` however, I do not understand the code enough to know whether its not just hiding the problem. Perhaps the correct thing to do would be to provide a proper parent to the `Sublime::View::widget` method, but I don't know what that would be. If the patch is o.k., I can prepare a pull-request. The crash can be reproduced every time. -- Backtrace (Reduced): #4 __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:50 #5 0x00007f89a24ed859 in __GI_abort () at abort.c:79 #6 0x00007f89a2930bf7 in qt_message_fatal (context=..., message=<synthetic pointer>...) at global/qlogging.cpp:1914 #7 QMessageLogger::fatal(char const*, ...) const (this=this@entry=0x7fffcb0f7970, msg=msg@entry=0x7f89a2c3bea0 "ASSERT: \"%s\" in file %s, line %d") at global/qlogging.cpp:893 #8 0x00007f89a2930003 in qt_assert(char const*, char const*, int) (assertion=<optimized out>, file=<optimized out>, line=<optimized out>) at ../../include/QtCore/../../src/corelib/global/qlogging.h:90 Possible duplicates by query: bug 440777, bug 440257, bug 438433, bug 429344, bug 427384. Reported using DrKonqi