Nahum Shalman

Sorry, I don't think I was clear. My test is not diagnostically useful in its current form. It passes even without the proposed fix. :( For a regression test to...

> Should #263 be closed? That is what I would do. While I have your attention, any thoughts on my question: > Is there a preference for keeping the changes...

> @nshalman It may be better to build this out as a separate FEN project -- e.g. `fsnotify/fen` > > I can't see myself maintaining or even doing code reviews...

I said forever ago: > I'll look into getting this into shape to land after #289, and I'll see if I can rustle up a reviewer with an eye to...

> Especially as omnios does this already with the mattermost they package. Wow, so the code from #263 was already in use: https://github.com/omniosorg/omnios-extra/blob/687e07dc8b339613cb888b29aab36c43b91d4ed2/build/mattermost/patches-mattermost-server/fen.go.patch Good to know.

Just leaving a breadcrumb for myself based on an IRC conversation. The errors I'm seeing in CI suggest that when the VM is slowing down the code, events are getting...

In order to stress test before and after, I wrote a tweak that I think increases the chance of batching and does seem to make the original test failures occur...

With #422 incorporated, along with the changes in 04e3205 I'm no longer able to trigger test failures. I now trust this code and believe it to be superior to #263...

Again, with extensive stress testing I was able to trigger a test failure which led me to realize this issue in my x/sys/unix Event Ports implementation: https://go-review.googlesource.com/c/sys/+/380034 So, I need...

Now that https://go-review.googlesource.com/c/sys/+/380034 has been merged, I hope to pick this up again some time soon.