Cleanup/Update comments in liboverride code.

This commit is contained in:
Bastien Montagne 2020-11-10 15:54:38 +01:00
parent a284e559b9
commit de8d7003bc
1 changed files with 4 additions and 4 deletions

View File

@ -1881,15 +1881,15 @@ ID *BKE_lib_override_library_operations_store_start(Main *bmain,
/* This is fully disabled for now, as it generated very hard to solve issues with Collections and
* how they reference each-other in their parents/children relations.
* Core of the issue is creating and storing those copies in a separate BMain, while collection
* copy code re-assign blindly parents/children, even if they do not belong to the same BMain.
* Core of the issue is creating and storing those copies in a separate Main, while collection
* copy code re-assign blindly parents/children, even if they do not belong to the same Main.
* One solution could be to implement special flag as discussed below, and prevent any
* other-ID-reference creation/update in that case (since no differential operation is expected
* to involve those anyway). */
#if 0
/* XXX TODO We may also want a specialized handling of things here too, to avoid copying heavy
* never-overridable data (like Mesh geometry etc.)? And also maybe avoid lib
* reference-counting completely (shallow copy...). */
* reference-counting completely (shallow copy). */
/* This would imply change in handling of user-count all over RNA
* (and possibly all over Blender code).
* Not impossible to do, but would rather see first is extra useless usual user handling is
@ -1928,7 +1928,7 @@ void BKE_lib_override_library_operations_store_end(
BLI_assert(ID_IS_OVERRIDE_LIBRARY_REAL(local));
/* Nothing else to do here really, we need to keep all temp override storage data-blocks in
* memory until whole file is written anyway (otherwise we'd get mem pointers overlap...). */
* memory until whole file is written anyway (otherwise we'd get mem pointers overlap). */
local->override_library->storage = NULL;
}