Yep, expanded the comment to note this isn't simply a paranoid null check.
/* The object may be set to NULL if the ID is being cleared from #collection_foreach_id,
* generally `cob->ob`…With you on the first two issues, but don't think BLI_ghash_ensure_p helps here as the old pointer used as as a key needs to be removed, the new one needs to be inserted again.
Sorry, misclicked.
- After this change it's no longer possible to symlink
release/into{BUILD_DIR}/bin/{MAJOR_VER}.{MINOR_VER}/to bypass theinstalltarget. While I don't see…
Some notes:
- After this change it's no longer possible to symlink
release/into{BUILD_DIR}/bin/{MAJOR_VER}.{MINOR_VER}/to bypassmake install. While I don't see this as an argument to…
Generally seems fine, tests/python/bl_keymap_completeness.py:8 needs "release", "scripts" to be replaced with "scripts".
@PratikPB2123 it's a known issue, I opened it to keep track of any problems related to the inability to know absolute positions of windows.
All issues should be addressed now.
@blender-bot build
Agree, at the time of writing I was eager to flag up any issues early, but this adds too much overhead, even for debug builds, removing the BLI_assert.
It's awkward either way, Sergey requested to move this into a utility function, while it's nice to split out in some ways - it's rather spesific functionality which is only used in one place,…
Makes sense, moved into collection_gobject_hash_ensure.
Following up on discussion on D17016, where I noted.
The hypothetical worst-case this would avoid is building and throwing away the object hash on frame-change (for e.g.),
Sergey…