https://bugs.kde.org/show_bug.cgi?id=518581

--- Comment #38 from Altarius <[email protected]> ---
UPDATE — the root cause is NOT the operating system.

In my earlier report I attributed severe UI slowness to macOS 27.
That diagnosis was wrong. After reinstalling macOS Sequoia 15.8.1
from scratch, the problem reappeared. Further testing isolated the
actual trigger: the number of grouped items in the current album.

ENVIRONMENT
- digiKam 9.1.0
- macOS Sequoia 15.8.1
- MacBook Pro 16" 2021, Apple M1 Max, 64 GB RAM, 4 TB SSD
- Database: MariaDB (Homebrew) on 127.0.0.1:3306
- exiftool 13.58
- Collection: 3206 files, flat folders, RAW+JPG pairs (RAF/ORF + JPG)

STEPS TO REPRODUCE
1. Open an album containing 622 items (309 RAW+JPG pairs).
2. Browse thumbnails and open Preview — everything is fast.
3. Select all, right-click, Group > Group Selected By Filename.
4. Thumbnails and Preview become extremely slow; the whole
   application stops responding ("digikam (non risponde)" in
   Activity Monitor).
5. In the same session, without changing album or restarting,
   ungroup the same items. Speed returns to normal immediately.

KEY OBSERVATION — the slowdown scales with the NUMBER of groups,
not with the number of files:
- album with 26 groups: no slowdown at all
- album with 309 groups: application unresponsive

Both albums are in the same collection, same session, same settings.

ADDITIONAL NOTES
- CPU and memory usage stay normal during the freeze. No swap
  activity. This looks like a blocked UI thread, not resource
  exhaustion.
- Light Table is never affected, which fits: it does not display
  grouping.
- Relevant settings: Miscellaneous > Grouping has Metadata = Yes
  and Renaming = Yes; "Draw frames around grouped items" is enabled.
- Earlier I migrated from SQLite to MariaDB with no improvement,
  which rules out the database backend.

This fully explains the behaviour I previously blamed on the OS:
grouping was present whenever the application was slow, and it was
lost on every database migration — which is why the problem seemed
to disappear and come back on its own.

Happy to provide further test data or run specific builds.

-- 
You are receiving this mail because:
You are watching all bug changes.

Reply via email to