[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Traces issue on DOS resolved
From: |
Akim Demaille |
Subject: |
Re: Traces issue on DOS resolved |
Date: |
15 Aug 2001 12:03:56 +0200 |
User-agent: |
Gnus/5.0808 (Gnus v5.8.8) XEmacs/21.4 (Artificial Intelligence) |
>>>>> "Tim" == Tim Van Holder <address@hidden> writes:
Tim> Hmm - apparently, the DJGPP port of Perl is not without its
Tim> problems. It is the cause of the lost traces. For whatever
Tim> reason (I'll have to delve deeper to find out exactly what) it
Tim> breaks m4's --error-output option. So when the m4 command run by
Tim> autom4te is run directly from the shell, or through system() from
Tim> a C program, it produces a traces file, but if run from Perl's
Tim> system() function, it produces an empty traces file. As a
Tim> workaround, I now have autom4te run
Tim> m4 ... 2>traces >output
Tim> instead of
Tim> m4 --error-output=traces ... >output
Beware, that's very different: errprint message are _not_ redirected
with --error-output, which means that in your case, the traces might
be polluted with error msg etc. The rest of the machinery _will_
choke some day.
Tim> and that seems to work. I'll try to find a better way; if that
Tim> is not possible, and the above change is viable on other
Tim> platforms (don't see why not) I'll submit a patch to that effect.
No, that patch cannot be accepted :(
Tim> I'll run the full testsuite later tonight to see what issues
Tim> remain.
Thanks Tim! Thanks a lot.
I would right a very small Perl script exhibiting the problem, and ask
for assistance to Perl/DJGPP-Perl gurus.