Bug 128738

Summary:	Edit Macros crashes (kf5)
Product:	LibreOffice	Reporter:	Buovjaga <ilmari.lauhakangas>
Component:	UI	Assignee:	Not Assigned <libreoffice-bugs>
Status:	RESOLVED DUPLICATE
Severity:	critical	CC:	m.weghorn
Priority:	high	Keywords:	haveBacktrace
Version:	6.4.0.0.alpha0+
Hardware:	x86-64 (AMD64)
OS:	Linux (All)
See Also:	https://bugs.documentfoundation.org/show_bug.cgi?id=123851
Whiteboard:
Crash report or crash signature:		Regression By:
Bug Depends on:
Bug Blocks:	102495
Attachments:	GDB trace of crash

Description Buovjaga 2019-11-12 10:17:47 UTC

Created attachment 155736 [details]
GDB trace of crash

1. Use the kf5 backend
2. Tools - Macros - Edit Macros

Macro security setting has no effect. This might be related to the Qt version because I am seeing this with all commits of 6.4 bibisect repo (I have Qt 5.13.2). Also, at some point with my master the problem went away. It is not related to user profile.

I am not seeing this with 6.3.

Arch Linux 64-bit
Version: 6.4.0.0.alpha1+
Build ID: bcc701e7c544331f0b5992df532919232b5ae5a0
CPU threads: 8; OS: Linux 5.3; UI render: default; VCL: kf5; 
Locale: fi-FI (fi_FI.UTF-8); UI-Language: en-US
Calc: threaded
Built on 12 November 2019

Comment 1 Michael Weghorn 2019-11-12 12:25:38 UTC

Works OK for me (no crash) on Debian testing with Qt 5.12.5 and

Version: 6.4.0.0.alpha1+
Build ID: a0c30af2ac83eb0f7e76c336d60a31872db19938
CPU threads: 4; OS: Linux 5.2; UI render: default; VCL: kf5; 
Locale: en-GB (en_GB.UTF-8); UI-Language: en-US
Calc: threaded

(which is ~20 commits older than the commit you built)

Comment 2 Michael Weghorn 2019-11-12 12:33:51 UTC

Looking at the backtrace, this might be the same cause as bug 123851. We (I think it was bubli, jmux and me) quickly discussed the latter one a while ago, that seemed to be some "division by zero" problem inside the Qt styling code (maybe caused by values passed by LO), but we didn't dig deeper so far.

Comment 3 Jan-Marek Glogowski 2019-11-12 19:18:18 UTC

(In reply to Michael Weghorn from comment #2)
> Looking at the backtrace, this might be the same cause as bug 123851. We (I
> think it was bubli, jmux and me) quickly discussed the latter one a while
> ago, that seemed to be some "division by zero" problem inside the Qt styling
> code (maybe caused by values passed by LO), but we didn't dig deeper so far.

Both backtraces have a SIGFPE, Arithmetic exception, which can basically just happen by div/0 with normal math, or sqrt(-x) or somesuch. As with bug 123851 I can't reproduce. And both want to draw QStyle::CC_ScrollBar. I'm quite sure it's the same.

Comment 4 Buovjaga 2019-11-12 19:21:08 UTC

Duping is fine by me, so let's do it.

*** This bug has been marked as a duplicate of bug 123851 ***