Mateusz Kowalczyk

Results 150 comments of Mateusz Kowalczyk

> I don't know the opinion of the other maintainers, but I think we don't have much to lose since it's behind a feature gate. Well, there _is_ additional maintenance...

I addressed the couple of comments.

Is there some specific tracking issue we have to wait for? There's https://github.com/imgui-rs/imgui-rs/issues/462 but that's not looking like it's going to happen any time soon. Technically I think they could...

Also this is without zifter; I tried to run it but ``` diff --git a/zift.hs b/zift.hs index 4793fe2..9b844d8 100755 --- a/zift.hs +++ b/zift.hs @@ -1,5 +1,7 @@ #!/usr/bin/env stack {-...

I have not tried with imports nor didn't try running the tests. I will try in coming days. I don't see why imports should become broken as far as GHC...

GHC tells us the module name. It's either in the file or Main.

Is this not just: ```haskell createDirectoryUnder top = withCurrentDirectory top . createDirectoryIfMissing True ``` I'm unsure what kind of new guarantees/races could `createDirectoryUnder` provide that the simple user-land code above...

@shym Ah, right, completely forgot. I don't think what you're asking for can be reasonably atomic, just like `mkdir -p` is not. However, we can still implement what OP asked...

I'm making an uneducated guess and going to say here:https://github.com/haskell/process/blob/17b8a15ce174fe7fa1d7572441b40cccb4a906dd/cbits/posix/runProcess.c#L292 Edit: notably, `childFailed` function. As the message comes from the OS, I think probably you want to examine the error...