Created attachment 151901 [details] The document that freezes LibreOffice Writer. LibreOffice Writer freezes when opening the attached DOCX file with LibreOffice Writer. Even a slight modification in Microsoft Word (like deleting some text from a cell or adding an empty paragraph before the first table) can “fix” the document, so this seems to be a corner-case rendering issue. Steps to reproduce: 1. Download the attached file. 2. Open the file with LibreOffice Writer. Actual results: LibreOffice Writer freezes when opening this particular document. Expected results: LibreOffice Writer shouldn't freeze. LibreOffice details: Version: 6.4.0.0.alpha0+ (x64) Build ID: 02bee5f5d846f40f244c5acab044177576a6d869 CPU threads: 4; OS: Windows 6.3; UI render: GL; VCL: win; TinderBox: Win-x86_64@42, Branch:master, Time: 2019-06-02_21:57:08 Locale: hu-HU (hu_HU); UI-Language: en-US Calc: threaded Bibisected in bibisect-win32-6.3 https://cgit.freedesktop.org/libreoffice/core/commit/?id=e502463fa9a601963aa9f5a8783eb1318de36c13 author Miklos Vajna <vmiklos@collabora.com> 2019-03-11 21:40:44 +0100 committer Miklos Vajna <vmiklos@collabora.com> 2019-03-12 08:43:22 +0100 tdf#123104 DOCX import: fix lack of vertical merge due to rounding Technically this commit doesn't cause the freezing itself, just changes the height of the tables, which accidentally causes LibreOffice to freeze. Adding Cc: to Miklos Vajna, please take a look sometimes.
Created attachment 151902 [details] A screenshot showcasing the issue.
Reproduced in Version: 6.3.0.0.beta1+ Build ID: 4abdaf4afb2245d404f6709124b3c627b07b8a3c CPU threads: 4; OS: Linux 4.15; UI render: default; VCL: gtk3; Locale: ca-ES (ca_ES.UTF-8); UI-Language: en-US Calc: threaded Adding Cc: to Miklos Vajna Setting priority to high since the patch was backported to libreoffice-6-2 branch
We have a trade-off here: depending on if you round or not round the table column sizes, either you get bug 123104 or the bug described in commit 29cbbad64088354425c606f9eb6c267bdf7731dc. Given the original commit was only a "nice to have" change and the regression was reported by users, I just reverted my change, since I expected sorting this out would require considerable resources (I'm not short on problems right now). I think it's reasonable to say that if we had a baseline, then a commit introduced a regression, and now the revert (fixing the regression) causes further problems, then yes, this is a bug, but it's not a regression, since it never really worked without serious side-effects. Updating keywords accordingly.
I guess we can set the importance to normal since this is not a real regression... we can increase it later if we see many documents affected...
Issue fixed by https://cgit.freedesktop.org/libreoffice/core/commit/?id=44e44239de35c1548809c96e13bfa9d64c7ca441 Closing as duplicate of bug *** This bug has been marked as a duplicate of bug 120315 ***