Re: new snapshot available: grep-3.7.62-9a668

[Gary Johnson]

On 2022-04-25, Gary Johnson wrote:
> On 2022-04-24, Jim Meyering wrote:
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA512
> > 
> > We will make a new release of grep in the next week or so.
> > Please try this out and let us know about any build or
> > test failure.
> > 
> > grep snapshot:
> >   https://meyering.net/grep/grep-ss.tar.xz      1.7 MB
> >   https://meyering.net/grep/grep-ss.tar.xz.sig
> >   https://meyering.net/grep/grep-3.7.62-9a668.tar.xz
> 
> I just tried building this on an up-to-date Cygwin system running on
> Windows 10.  It built fine and seems to run fine.  However, "make
> check" reported the following failure:
> 
>     
> ============================================================================
>     Testsuite summary for GNU grep 3.7.62-9a668
>     
> ============================================================================
>     # TOTAL: 201
>     # PASS:  189
>     # SKIP:  11
>     # XFAIL: 0
>     # FAIL:  1
>     # XPASS: 0
>     # ERROR: 0
>     
> ============================================================================
>     See gnulib-tests/test-suite.log
>     Please report to bug-grep@gnu.org
>     
> ============================================================================
> 
> gnulib-tests/test-suite.log said only this about the failure:
> 
>     FAIL: test-select-out.sh
>     ========================
> 
>     FAIL test-select-out.sh (exit status: 1)
> 
> The header of that file says:
> 
>     # Test select() on file descriptors opened for writing.
> 
> The output of the script is not very informative.  It looks like it
> needs to run within some sort of framework.  It may be a while
> before I can look at this more closely.

I think I figured it out.  The test-select-out.sh script failed in
the section at lines 23 - 25, where it tries to open a pipe to cat.
I ran just that command (line 24) and subsets of it and found that
it sometimes passes and sometimes fails, depending on the program
being piped to and on the program previously piped to.  When it
fails, it saves 0 to t-select-out.tmp, which means that the select()
call at line 53 of test-select-fd.c timed out.  The timeout was set
for 10 ms.  I increased it to 100 ms and now the test passes.

Cygwin has always been a bit sluggish on this Windows 10 machine,
especially when it starts a new program, so taking longer than 10 ms
to do so doesn't surprise me.

I think this failure can be ignored.  I don't know how you feel
about changing that timeout permanently.

Regards,
Gary