On Thu Feb 6 15:44:00 2025 +0000, Jinoh Kang wrote:
There are more:
- concrt140: https://gitlab.winehq.org/wine/wine/-/blob/7b57598dcfc9198aa9cf6ac4ce17e2328...
- Seems like the same for msvcirt, msvcp140_1, msvcp60, and msvcp90,
although I haven't verified.
- mapi32:
https://gitlab.winehq.org/wine/wine/-/blob/7b57598dcfc9198aa9cf6ac4ce17e2328... calls https://gitlab.winehq.org/wine/wine/-/blob/7b57598dcfc9198aa9cf6ac4ce17e2328...
- secur32:
https://gitlab.winehq.org/wine/wine/-/blob/7b57598dcfc9198aa9cf6ac4ce17e2328... calls https://gitlab.winehq.org/wine/wine/-/blob/7b57598dcfc9198aa9cf6ac4ce17e2328... calls https://gitlab.winehq.org/wine/wine/-/blob/7b57598dcfc9198aa9cf6ac4ce17e2328...
- Same for mpr:
https://gitlab.winehq.org/wine/wine/-/blob/7b57598dcfc9198aa9cf6ac4ce17e2328... calls https://gitlab.winehq.org/wine/wine/-/blob/7b57598dcfc9198aa9cf6ac4ce17e2328... calls https://gitlab.winehq.org/wine/wine/-/blob/7b57598dcfc9198aa9cf6ac4ce17e2328... At this point it's plausible that I may be missing even more.
Besides, since you mentioned uncertainty around apisets, I wonder if such refactoring would merit a separate MR. I'm uncomfortable with consolidating too much possibility of regression into one MR, I think it's best spread over time.
Maybe go only the `!current_modref->is_dynamic` part applied, although I'm not sure if I'm still breaking someone else's workflow with any of the DLLs above.