Created attachment 198975 [details] docx/tdf127118-6.docx_mso.pdf: how it looks in Word 2019 The text of a rotated table, split across pages, is missing on the first page. This started with 25.2 commit d1c5ed30d1128c0500d3fab72bb5505f81a1bdc1 Author: Mike Kaganski on Sat Aug 24 23:51:25 2024 +0500 tdf#162614: report the reason why a frame couldn't grow ... and use it in SwTextFrameBreak::IsBreakNow to decide if the content needs to go to a follow or not. Steps to reproduce: 1.) open tdf127118.docx (attachment 165255 [details] from bug 127118) Notice that there is no text on page 1 (only numbered cells). At the time of the commit, there was no text on page 2 either. That came back with commit d5fcc8ea25e416a4d2bea5e0f92c0e403997d845 Author: Michael Stahl on Wed Oct 2 18:51:44 2024 +0200 tdf#163032 sw: fix painting of row frame with rowspan > 1 cell Found by Collabora's mso-test
Hello Justin L, Thank you for reporting the bug. Can you see if bug 127118 is similar to your problem so that I can mark your report as a duplicate of it? (Later bugs are generally marked as duplicates of earlier bugs). *** This bug has been marked as a duplicate of bug 127118 ***
(In reply to mikhail.machine from comment #1) > Can you see if bug 127118 is similar to > your problem so that I can mark your report as a duplicate of it? (Later > bugs are generally marked as duplicates of earlier bugs). Later bugs are only marked as dupes of earlier, when: 1. The later bugs don't provide substantially better data (reproduction steps, analysis, etc.) over the older ones. 2. The later bugs didn't start *later* than the older ones (even if similar). In this case, there is explicit "This started with 25.2 commit d1c5ed30d1128c0500d3fab72bb5505f81a1bdc1", which rules out any bug appeared before 25.2. 3. You paid attention to wait for response, which you asked for ("Can you see if...?"), especially when the author is Justin, one of the code revelopers.
Confirm with Version: 25.8.0.0.alpha0+ (X86_64) / LibreOffice Community Build ID: 7da1497aa462e2b719aa9b308a749caf7b9a19b1 CPU threads: 4; OS: Linux 6.8; UI render: default; VCL: gtk3 Locale: cs-CZ (cs_CZ.UTF-8); UI: en-US Calc: threaded
I am unsure if it's correct to mark this as a regression. IIUC, we rendered it wrong earlier, too - it didn't match Word; and a change from one incorrect render to another is not a regression :) But more importantly, the previous rendering just happened to show stuff on page 1, because the layout loop happened to be forcefully terminated on that layout stage. Let me remove the "regression" field. The commits that changed the behavior are indeed important code pointers. But please of course revert my changes, if I'm wrong.