cbits/fork-exec: Don't dup2 identical fds

[bgamari]

Darwin violates POSIX by making
dup2(x,x), which should be a no-op, error. Consequently, we must take
care not to dup2 in such cases.

We had already made this change in the posix_spawnp codepath but I had
assumed that this only affected posix_spawnp, not the dup2 system
call itself.

See also: #214

[psibi]

I compiled Stack using this patch, but I'm still able to reproduce the issue when running stack's integration test. Some relevant logs:

/Users/ec2-user/.stack/programs/x86_64-osx/ghc-9.2.3/bin/ghc-9.2.3: startProcess: dup2: invalid argument (Bad file descriptor)
Main.hs: Exited with exit code: ExitFailure 1
CallStack (from HasCallStack):
  error, called at /Users/ec2-user/stack/test/integration/lib/StackTest.hs:63:34 in main:StackTest
  stack, called at /Users/ec2-user/stack/test/integration/tests/111-custom-snapshot/Main.hs:4:8 in main:Main

Unfortunately I haven't been able to find a minimal reproducible example.

[bgamari]

Hmm, this is quite mysterious. Are you certain that you are linking against the correct process? Would it be possible to collect dtruss output from this execution?

[psibi]

I mis-applied your patch, I rebased your changes against master branch and it indeed fixes the issue. Relevant logs:

End of log for 111-custom-snapshot
Successful tests:
- 111-custom-snapshot

No failures!

Although I'm still able to reproduce the posix_spawnp issue which I have described in the issue.

[psibi]

This fixes the dup2 issue that was being reproduced in the stack codebase.

[snoyberg]

Thanks @bgamari!