Re: [BUG]? Strange return values of close()

[Eli Zaretskii]

> Date: Wed, 9 Feb 2022 16:01:28 -0500
> From: "Andrew J. Schorr" <aschorr@telemetry-investments.com>
> Cc: bug-gawk@gnu.org
> 
> Thanks for the bug report. The problem is that the exit status of a previously
> reaped process is not being handled properly. The attached patch seems to fix
> the issue for me, but I'm not able to test the Mingw32 piece of it.

It should work, although the assumption that exit status values are 8
bits is incorrect for MS-Windows: the exit status there can be any
unsigned 32-bit value.

(Of course the test program returns zero status for both "empty" and
"non-empty" cases on MS-Windows.  Can you tell me where 'status' value
comes from in the Unix case when running this script, so I could see
if this is due to Gawk's code or to some Windows idiosyncratic
behavior?)

> We should probably add a test case and check that it works on all
> platforms.

Agreed.