Caolan
- PDF:
- system afdko to keep rene happy
- null deref on auto clipboard store in markdown format
- capture more javascript in coverity
- tdf#168672 Closing webdav documents is extremely slow
- fulldoc invalidations
- hack week
- experiment with cgroups for memory use
- HW: explore cgroups v2 memory tracking
- tdf#168788 fix regression
- HW: the trick with cgroups is there can’t be a
process in the cgroup when enabling subcontrollers,
but they can be in a subgroup, so create subgroup
move current pid in there, then enable controllers in
that now empty parent and things can happen- Control Group APIs and Delegation
- numbers from memory.current are wildly different from
smaps[_rollup]. anon matches exactly, while the total
is quite different. - docker/podman report memory.current - memory.stat/inactive_file
as a container’s memory usage.
Mike
-
DOCX Document that fails to convert to ODT
- Backported and merged the fixes
-
Nondeterministic result when saving ODG as PNG with Win 11 24H2
- Explained our rendering settings
-
Weird formatting of numbered lists
- Finally found a way to fix it without a huge rework
- Done
-
Upstream
- tdf#155855 - Windows clipboard issue (pasting bitmap)
Tomaz
XLSX with pivot table fails to open in Calc (regression)
- Worked around the crash but we lose pivot table formatting as it depends on the code
- will prepare a patch
- Still not completely sure why we don’t have the information - we rely on pivot table cache, which doesn’t have it.
- We do take data from the pivot table cache, but not all that is provided
- Mostly we rely that the data is available in the sheet so we trigger pivot table refresh in OOXML case
- Looked into the pivot table caching - how it works when the data is not available
- There are a few scenarios when the sheet data is not available:
- pivot table output is written to the sheet, but no pivot table is present in the sheet
- pivot table output is written and there is a pivot table there, but looking at the properties doesn’t show the members of the pivot table, refreshing destroys the output because there is no data
- pivot table output is written and there is a pivot table there, and the properties show all the fields correctly, but a refresh destroys the output because there is no data
- pivot table output is written and there is a pivot table there, and we can change the pivot table even when the sheet has no data anymore
- we mostly do 1. in some scenarios also 2. and almost 3. (should be with ODS but it is not) but we should ideally do 4.
- ODS doesn’t store the data necessary for 4., but OOXML does.
- Looked into the implementation it’s not so easy to add this - we read the data from OOXML, but we really don’t anything with the data.
- We do read some at least some of the pivot table structure, whcih could work fine when we have the data in the sheet, but if the data in the sheet isn’t available it can become a problem.
- Need to properly understand what we do and more testing… we really need to have a good and robust pivot table implementation considering how much it is used.
- There are a few scenarios when the sheet data is not available:
- I Improved the debugging for pivot table - added dumpAsXml, so we can add more information and it is structured better
- can also be used in a test
- there is code dumping that is normally not triggered and hidden behind a flag and th code was not compiling
- want to make dumping always available (like dumpAsXml is) so the code won’t become stale
- must be more clear in the code what happens when and for what reason - currently it’s a bit unclear why we need certain classes
Miklos
Worked on: