BigPtrArray::Index2Block undo crash
Steps to Reproduce:
1. Open attachment 94362 [details] from bug 75208,
User Profile Reset: No
but not in
Bibisected with Linux 44max to https://git.libreoffice.org/core/+/b4ccde72b8e2e45e7276d5b08b182495a1b1a617%5E!/
Copy first-header-footer test from ww8 to ooxml
Adding Cc: to Luke Deller
Repro step between 2 and 3 should be "CTRL+C"
Created attachment 162150 [details]
GDB trace of crash
Arch Linux 64-bit
Build ID: ad0351b84926075297fb74abbe9b31a0455782af
CPU threads: 8; OS: Linux 5.7; UI render: default; VCL: kf5
Locale: fi-FI (fi_FI.UTF-8); UI: en-US
Built on 17 June 2020
Created attachment 162153 [details]
For the record, I can't reproduce it in
Build ID: c98b1f1cd43b3e109bcaf6324ef2d1f449b34099
CPU threads: 4; OS: Linux 4.19; UI render: default; VCL: gtk3;
Locale: en-US (en_US.UTF-8); UI-Language: en-US
so I do believe the issue was fixed at some point between the commit in commit 1 and c98b1f1cd43b3e109bcaf6324ef2d1f449b34099. Later, this issue was reintroduced, likely related to bug 134107 or bug 129805
@Telesto, when checking with older versions, please test with other versions in between. LibreOffice 4.4.7 was released more than 5 years ago, which is a huge window of time
(In reply to Xisco Faulí from comment #4)
> @Telesto, when checking with older versions, please test with other versions
> in between. LibreOffice 4.4.7 was released more than 5 years ago, which is a
> huge window of time
Indeed, apologies to Luke for the noise.
The first hint of badness in Linux 6.4 line is a crash upon quitting. Blamed commit: https://git.libreoffice.org/core/+/5ba30f588d6e41a13d68b1461345fca7a7ca61ac%5E!/
Later in Linux 6.4 repo it morphed to crashing already in the copying step. Appeared in range: https://git.libreoffice.org/core/+log/4dc1615c80e8e66d339dc86fa95bbc76e884d988..d9030ad6298e2f49ee63489d6158ea6ad23c0111
Finally, it became what it is today in Linux 6.5 repo with the commit https://git.libreoffice.org/core/+/cf226535f9903a048b1c105b180ae3a50a776e68%5E!/
tdf#129805 writerfilter: fix import of nested generic field
Adding Cc: to Michael Stahl
this crashes because:
during pasting, SwUndoInserts is created first, then SwUndoInsLayFormat is created for flys anchored in a footer from a page style referenced by a page break on some text node, then SwUndoInserts::SetInsertRange() is called.
on Undo, first SwUndoInsLayFormat deletes the flys, then SwUndoInserts has wrong node indexes.
this is not a regression, can repro it in LO 5.2 and OOo 3.3.
what changed in LO 6.4 is that apparently the flys anchored in the footer are imported now, when previously they weren't.
Created attachment 162590 [details]
*** Bug 134434 has been marked as a duplicate of this bug. ***
Changing to Inherit from OOo based on comment 6