11 Feb
2025
11 Feb
'25
9:15 a.m.
On Mon Feb 10 21:43:53 2025 +0000, Elizabeth Figura wrote:
That's a valid concern, but I'm more than a little hesitant to *not* lazily initialize just because of that. I'd rather just add a comment explaining why we lazily create and that the state is identical to what it is at creation time.
I don't think we should preemptively guard against issues like that at the cost of making the code so much difficult to follow. If wineserver / kernel state usages are exclusive, it should IMO be explicit in the code. If we still need and can make things lazier later then we can consider it, but maybe it's not even going to be needed. If it really is, it should probably be hidden in the inproc sync logic (like only create the fd lazily).