https://bugs.documentfoundation.org/show_bug.cgi?id=134203
--- Comment #10 from Jan-Marek Glogowski <[email protected]> --- (In reply to Buovjaga from comment #9) > (In reply to Jan-Marek Glogowski from comment #8) > > FYI: the bibisect commit was reverted to fix bug 141556 in > > https://git.libreoffice.org/core/+/ > > 0fedac18214a6025401c4c426466a5166553e8ec%5E%21 > > > > I tried to find an alternative fix (dropping quite some code), but > > everything I came up with resulted in broken layout: > > https://gerrit.libreoffice.org/c/core/+/114105 > > > > Main problem is that I simply don't understand the condition, when it's > > actually fine to break the layout, so it can be resumed correctly. 2nd > > problem is, that the idle layout is currently stateless, but for my current > > approach it would need some state to handle removing of pages. > > > > So I'm removing the bibisect* keywords; maybe it can be bibisected again > > hoping for a better result. > > I don't see how it could be bibisected again as > 0fedac18214a6025401c4c426466a5166553e8ec makes no difference to the result. > Do you mean bibisecting earlier than 6.4? What would be the good/bad states? No. I thought it may have been fixed by 0fedac18214a6025401c4c426466a5166553e8ec and then broken again after it and you just unluckily skipped that in the bibisect. My assumption is that 0fedac18214a6025401c4c426466a5166553e8ec is ok, as it's a revert. OTOH, while it technically reverts my commit, it also includes new code to fix the original problem. So you can test 0fedac18214a6025401c4c426466a5166553e8ec, but if that is still broken, you need to rule out, that the new code introduces this, which would mean to build LO at that commit and just revert the new part; that's a lot of work... Or maybe it's enough to just revert that new code part on master. -- You are receiving this mail because: You are the assignee for the bug.
