[Libreoffice-bugs] [Bug 112539] Index TOC 'eats' next paragraph (may also be a page crossing) after saving as docx
bugzilla-daemon at bugs.documentfoundation.org
bugzilla-daemon at bugs.documentfoundation.org
Thu Sep 21 08:36:35 UTC 2017
https://bugs.documentfoundation.org/show_bug.cgi?id=112539
Xisco FaulĂ <xiscofauli at libreoffice.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords|bibisectRequest |bibisected, bisected
Status|UNCONFIRMED |NEW
CC| |mstahl at redhat.com,
| |xiscofauli at libreoffice.org
Version|6.0.0.0.alpha0+ Master |5.4.0.0.alpha0+
Ever confirmed|0 |1
--- Comment #3 from Xisco FaulĂ <xiscofauli at libreoffice.org> ---
Regression introduced by:
author Michael Stahl <mstahl at redhat.com> 2017-03-02 21:55:27 (GMT)
committer Michael Stahl <mstahl at redhat.com> 2017-03-02 22:26:38
(GMT)
commit 06f9ad262742ea939bf23e82530b7166ca4ce456 (patch)
tree 9e2186a67301b7d9ecfdae661d07568075b96b49
parent 100da4fadbab795bf470a2e46790d36a7c46d944 (diff)
sw: Fix STL assert on DOCX export of ooo29679-42.odt
The problem was that the StartField_Impl() was called again and again on
the same field (actually 0-length ToXmark with dummy char), or in other
words, EndField_Impl() wasn't called and it wasn't removed from
m_Fields.
So tweak the south-pointing chariot DocxAttributeOutput::EndRun() again
in the hope it will go another km or two before it starts pointing east.
This doesn't actually produce the elements in the ideal order in some
cases, but given that this code has clearly passed the complexity event
horizon that is too much to ask for.
Bisected with bibisect-linux-64-5.4
Adding Cc: to Michael Stahl
--
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/20170921/0357d652/attachment.html>
More information about the Libreoffice-bugs
mailing list