Status report: consultancy team, week 7, 2024


  • regression on invaliding cell area
  • merge “don’t invalidate sheet on switching tabs” to 23.05, etc
  • disable printer queue for online, like we do for tests/fuzzing etc
  • ponder the oss code coverage stuff for a while, there are at least
    4 modules with 0% use, so presumably we shouldn’t need to build them
    at all if things work without these getting used.
    → drop reg, store, bib, cui from fuzzing builds
  • have a look at what compiler hardening flags are used by rh at the
  • sync input handler and zoom, works for one user now, but
    still seeing issues for multiple user case.
  • make check failure with emoji glyph with OpenDyslexic font installed
  • painful grind to build a test case for calc editing vs
    viewing mismatch
  • cell editing background regression in calc dark/light modes
  • debug multi view mismatch
  • misplaced text on editing shape
  • claimed regression, but seems to be like this for ages, fixed now
  • getComputedStyle slow for comments
  • improve calc note font categorization caching


  • CTRL + A not working with hidden sections
    • The current problem is a section with a page break with page style
      after a hidden section in the beginning of the document. In this case,
      the page style is ignored
    • Turns out that it’s old problem, but didn’t happen with sections -
      but a hidden bare paragraph in the beginning always caused the same problem.
    • A new 2nd page is created for the break after the hidden frame; then
      the paragraph gets moved to the first empty page …
    • When the break is in the very first document’s paragraph, the first
      page gets created initially with the correct style → need to search for
      the first visible paragraph.
    • SwLayHelper::CheckInsertPage
    • SwFrame::CheckPageDescs
    • Fixed and created unit tests for the issues not caught by existing
      tests (thanks to fdo114306_2.odt, which allowed to see the breakages
      visually, even though its unit test somehow passed).
    • Submitted the final (?) version of pre-requisite “0-height hidden
      section frames” to gerrit
    • Have a partial fix to selection problem
    • Turns out, that the bugdoc has additional factor (a frame in the
      hidden section)
    • debugging
    • Fixed; submitted
    • Fails sw_tiledrendering - testSearchTextFrameWrapAround
    • Luckily the tiledrendering test caught another breakage - find in
      frames (caused by a change that prevented passing boundaries of special
      document segments - redlines, footnotes, headers/footers, inserts, main
      body). The search crosses these boundaries freely. Created a flag.
    • The updated change passes all the tests
  • LibreOffice stops working after cleaning temp directory
    • Started
  • LO hangs while callling createTextRangeByPixelPosition([…]
    • Trying to repro to provide an estimation


pivot table:

  • Pivot table fix merged to master and backported to LO 24.2 branch

XLS document with DRM add-on:

  • Researched the issue and estimated, then implemented
  • Found the issue in import that prevents us to open the un-encrypted sheet
  • Added an infobar message, that the encrypted content couldn’t be opened
  • Pushed to master, LO 24.2 branch

Calc: add currency dropdown to the notebookbar:

  • Reworked the currency dropdown, so it is activated with single-click
    (no double click needed anymore)
  • Changed the “OK” button to “More Formats…” button that opens the
    Options dialog
  • Make the currency list nicer (WIP)


Worked on:

  • tdf#159483 Clipboard: Implement permissions for pasting in Chrome-based browsers
  • cool#8270 Multi-user text editing status bar confusion
  • cool#8292 Paste special in notebookbar