GeezerVison

Feel free to kill the workflows for now. Something doesn't seem right at the moment.

I have successfully gotten uvloop to work on windows now. Now we just need to kill the remaining workflows up until merge more of winloop into uvloop commit. Turns out returning 0 when the socket is AF_UNIX makes windows behave normally. I would assume it has to do with the way sockets on windows work, all I know is that the job is done and that I am finished with merging all of the tests from winloop on over to uvloop and that I have gotten my tests to run successfully on my laptop. on python 3.10.20.

@fantix I know I'm asking for quite a bit today but could you go ahead and kill all workflows up until #790 for me?

Oh wow, this is EPIC, @Vizonex! Thank you so much for all the hard work over the years! Looks like the workflows are stopped now (sorry for missing the request!), but I would suggest running CI debugging on your fork before PR is ready, so that you'll have more timely control to kill/restart them.

@fantix I managed to get every version from 3.8-3.12 to behave, I did however have to silence a few unittests for example: test_dns since it seems example.com might have some new quirks with it that are causing for IP addresses to not come out the same.

Another thing to point out is that I can't seem to get windows to compile and be tested in debug-mode onto the workflows. I'm sure someone or another reviewer will figure it out for me eventually but I haven't figured that out myself yet.

@Vizonex yeah I think we just need to find time and/or people to review this, as well as adding you to the uvloop maintainers, but that's up to @1st1.

There are a few fixes in master branch, which may unblock a few failing CI here.

Seems I had a couple conflicts to resolve. @fantix I found out that we might be able to get debug libraries for windows and others using some form of the uv package manager should I look into trying that some more?

@fantix What other changes should I look into before finalizing the merge?

All of the bugs and buggy tests that I could fix as of now are complete, was there anything else that I need to accomplish here?

I had a quick look at this, and it feels quite solid to me! I'll spend some time later to get the CI run again, and get a good-to-go from @1st1.

I had a quick look at this, and it feels quite solid to me! I'll spend some time later to get the CI run again, and get a good-to-go from @1st1.

That will be wonderful. Thank you for your patience.

@fantix Seems a new bug emerged since I was last here. everything within range of 3.10 or lower seems to fail. Vizonex#13

@fantix I seem to be failing at keeping caught up with all the changes going on. I'm wondering if I should close and repoen the PR or try utilizing a different approach to get this added in

Gonna reconvert to a draft for the time being. Seems there's not enough motivation yet.

@fantix I think I'm going to close and reopen as a new PR since there are now too many branch conflicts and many new things to take account for:

• example: playwright support which needed new flags for subprocesses.
• various bug fixes
• various improvmenets that would take a lot longer to review and merge.
• pyinstaller fixes on my end which added lazy hooks to compile exe applications
• improvements to DNS resolving which involved a cython version of a Future object
• improvements to socket selectors using a cdef class extension of a Future object

Do you have any ideas on what should be done before I reattempt this or if the improvements on winloop's side need mirroring on uvloop's end first?