On Sun Nov 5 14:06:36 2023 +0000, Loïc Rebmeister wrote:
Not really an "exhaustion" problem, more like conflicts for the same number twice, also ip is often shared across containers. Also I checked about FIFO, and windows doesn't have an api that would match with unix implementation and that could be used for communication. (Source: https://github.com/python/cpython/issues/103510) While I agree using IP is an alternative for unix socket, sockets are also exposed to local network and other users. Anyway, I'm offering money, even if `AF_UNIX` is not used I'm not going back on my word for the money. I guess I'm getting overhyped about everything on wine cause of my health isn't allowing me to make code for wine. I really want to help wine move forward, and @zfigura I'm really grateful of all the work you put into wine. I know I may look like someone that doesn't know anything, but I watch commit made on wine almost every day. I have health issues for many years, even before I made any commit on wine, my health is just getting too bad I guess. And @julliard I see the work you put into wine too, I think you are putting too much work on yourself. I know you take your weekends off, as I watch activity on wine, but you should delegate some of the work you do to other peoples. I think adding gitlab labels that trusted peoples can add like `approved`/`need more work`/`need review` would help. I know @zfigura you worked on wine server on similar stuff before. I know I'm getting a bit off-topic, but I thought it was important for me to tell how I am grateful for the work of everyone involved here.
Zebediah is right; there are other ways to cross the Wine/Unix barrier that work on master. Anyone that does have a pressing need/desire for AF_UNIX sockets in a program under Wine can apply this patch to their local build, which is what I'm currently doing. I'm keeping it free of merge conflicts (this is what the last two updates were) but otherwise it's complete as far as I'm aware.
While it would be nice to have this merged at some point, there's no rush in my opinion, and I'm certain there are more important merge requests that impact existing/commonly-used programs. Winsock code is touched rarely, so it's also not much work on my end to resolve the few merge conflicts that do crop up.