Re: Darwin vs. cat

[Jim Meyering]

Eric Blake wrote:
> When running test-closein.sh, I'm getting spurious output on Darwin:
>
> cat: standard output: Bad file descriptor
> PASS: test-closein.sh
>
> $ cat --version | head -n 1
> cat (GNU coreutils) 7.1
> $ uname -v
> Darwin Kernel Version 8.11.0: Wed Oct 10 18:26:00 PDT 2007;
> root:xnu-792.24.17~1/RELEASE_PPC
>
> I ran ktrace/kdump, and it looks like the problem occurs any time a pipe
> is created, but the reader exits without reading anything prior to the
> writer attempting to write.  For example:
>
> $ echo hi > foo
> $ cat foo | :
> cat: standard output: Bad file descriptor
> $ cat foo | { :; sleep 0.01; }
> $
>
> It seems like the Darwin kernel is failing to give SIGPIPE/EPIPE failures,
> and instead gives EBADF when the read end of the pipe is gone.  Is this
> anything that coreutils should try to work around?

Personally, I haven't been too concerned with coreutils "problems"
arising from flaws in Darwin, since I rarely use that type of system.
In this case, unless you want to work on it and find a clean fix,
I'd be inclined to defer working on it indefinitely.