Bug 120631

Summary: crashtesting assert
Product: LibreOffice Reporter: Caolán McNamara <caolan.mcnamara>
Component: WriterAssignee: Not Assigned <libreoffice-bugs>
Status: RESOLVED WORKSFORME    
Severity: normal CC: caolan.mcnamara, michael.stahl, nemeth, serval2412, xiscofauli
Priority: medium Keywords: haveBacktrace
Version: 6.2.0.0.alpha0+   
Hardware: All   
OS: All   
Whiteboard:
Crash report or crash signature: Regression By:
Bug Depends on:    
Bug Blocks: 105537    
Attachments: bt with debug symbols

Description Caolán McNamara 2018-10-15 19:07:04 UTC 
Description:
with the attachment https://bugs.documentfoundation.org/attachment.cgi?id=126429 from bug https://bugs.documentfoundation.org/show_bug.cgi?id=101145

./instdir/program/soffice.bin --headless --convert-to pdf ./tdf101145-1.odt

asserts with...

soffice.bin: /home/caolan/LibreOffice/core/sw/source/core/txtnode/thints.cxx:2759: bool SwpHints::MergePortions(SwTextNode&): Assertion `aIter1 != aRange1.second && aIter2 != aRange2.second' failed.
Application Error


since...

commit 1bbbe57dfc0b43d6b5444798d77dcdf5e4e76e49
Author: László Németh <nemeth@numbertext.org>
Date:   Tue Aug 28 21:13:09 2018 +0200

    tdf#119571 change tracking: show layout changes at paragraph join


Steps to Reproduce:
./instdir/program/soffice.bin --headless --convert-to pdf ./tdf101145-1.odt

Actual Results:
assert

Expected Results:
noassert


Reproducible: Always


User Profile Reset: No



Additional Info:
Comment 1 Julien Nabet 2018-10-17 19:48:45 UTC 
Created attachment 145789 [details]
bt with debug symbols

On pc Debian x86-64 with master sources updated today, I could reproduce this.
Thought a bt with symbols may help here (+ some errors on traces just before the assert)
Comment 2 Julien Nabet 2018-10-17 19:51:07 UTC 
2759 assert(aIter1 != aRange1.second && aIter2 != aRange2.second);
(gdb) p aIter1
$1 = {first = 3, second = {first = 0x55555aa6afa0, second = false}}
(gdb) p aRange1.second
$2 = {first = 3, second = {first = 0x55555aa6afa0, second = false}}
(gdb) p aIter2
$3 = {first = 3, second = {first = 0x55555aa6afa0, second = false}}
(gdb) p aRange2.second
$4 = {first = 4, second = {first = 0x55555adb6e80, second = false}}
Comment 3 Caolán McNamara 2018-12-19 16:36:23 UTC 
this seems to have healed
Comment 4 Xisco Faulí 2018-12-20 10:54:50 UTC 
(In reply to Caolán McNamara from comment #3)
> this seems to have healed

Setting to RESOLVED WORKSFORME unless we know the commit fixing it...