[Libreoffice-bugs] [Bug 135144] Saving DOCX to ODT with hiding tracking changes trashes all images
bugzilla-daemon at bugs.documentfoundation.org
bugzilla-daemon at bugs.documentfoundation.org
Tue Aug 4 07:07:58 UTC 2020
https://bugs.documentfoundation.org/show_bug.cgi?id=135144
Miklos Vajna <vmiklos at collabora.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords|regression |implementationError
CC| |michael.stahl at cib.de
--- Comment #5 from Miklos Vajna <vmiklos at collabora.com> ---
Now that bug 130707 is fixed, this is crashing in code introduced in core.git
commit 522ed1b685f5b7eb52efb9d874dfacc02bf82139 (ofz#18563 sw: fix 2 problems
with field marks, 2019-12-06). The immediate problem is:
Thread 1 "soffice.bin" received signal SIGSEGV, Segmentation fault.
0x00007fffcf0fb3fb in sw::CalcBreaks (rBreaks=std::__debug::vector of length 0,
capacity 0, rPam=SwPaM = {...}, isOnlyFieldmarks=false)
at
/home/vmiklos/git/libreoffice/master/sw/source/core/doc/DocumentContentOperationsManager.cxx:561
561
assert(std::get<0>(startedFields.top())->IsCoveringPosition(SwPosition(rTextNode,
i)));
(gdb) print startedFields
$1 = std::stack wrapping: std::__debug::deque with 1 element = {std::tuple
containing = {[1] = 0x0, [2] = false, [3] = 0, [4] = 0}}
I.e. we unconditionally dereference std::get<0>(startedFields.top()), but it's
a nullptr.
Michael Stahl: any idea what's going on here? Should I add a nullptr check
there or is there a deeper problem here?
Steps to reproduce:
1) Open attachment 163550 from this bug.
2) Save as ODT.
3) Open the ODT -> crash.
Sounds like a problem in the recently added CH_TXT_ATR_FIELDSEP feature.
--
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/libreoffice-bugs/attachments/20200804/1ea8b5cd/attachment.htm>
More information about the Libreoffice-bugs
mailing list