On Thu 09 Aug 2001 11:30, Nicholas Clark <[EMAIL PROTECTED]> wrote:
> On Thu, Aug 09, 2001 at 09:52:16AM +0200, H.Merijn Brand wrote:
> > On Thu 09 Aug 2001 05:15, Michael G Schwern <[EMAIL PROTECTED]> wrote:
> > > I'm getting false negatives because for some reason @nok contains some extra 
>junk:
> > > 
> > >   DB<5> p @nok
> > > sh mv-if-diff configpm.tmp lib/Config.pm
> > > File lib/Config.pm not changed.
> > >   make TEST_ARGS= TESTFILE=TEST _test_tty   ; \
> > >   make TEST_ARGS= TESTFILE=TEST _test_notty ; \
> > 
> > These should be filtered out in the mktest script. Haven't done that (yet)
> 
> Stuck together cut and paste may not apply patch for FreeBSD and Debian
> (And probably other BSDs and Linuxes):

Locally applied.

Also changed the $nok[0] =~ to grep as of Michaels remark. The log should
still catch the rest.

> --- mktest.pl~  Fri Jan  5 11:27:36 2001
> +++ mktest.pl   Wed Aug  8 17:10:58 2001
> @@ -128,6 +151,13 @@
>             m,^[-a-zA-Z0-9_/]+\.*(ok|skipping test on this platform)$, ||
>             m,^(xlc|cc_r) -c , ||
>             m,^\s+$testdir/, ||
> +           m,^sh mv-if-diff configpm\.tmp lib/Config\.pm, ||
> +           m,File lib/Config\.pm not changed, ||
> +           m,make TEST_ARGS= TESTFILE=TEST, ||
> +           # Don't know why BSD's make does this
> +           m,^Extracting .*with variable substitutions, ||
> +           # Or this:
> +           
>m,cc\s+-o\s+perl.*perlmain.o\s+lib/auto/DynaLoader/DynaLoader\.a\s+libperl\.a, ||
>             m,^   ### , and next;
>             if (m/^u=.*tests=/) {
>                 s/(\d\.\d*) /sprintf "%.2f ", $1/ge;
> 
> 
> I think Andy Doherty's recent patch to p5p make cure Makefile of its desire
> to relink perl and re-extract the pods each make test.
> 
> Nick


-- 
H.Merijn Brand    Amsterdam Perl Mongers (http://www.amsterdam.pm.org/)
using perl-5.6.1, 5.7.1 & 623 on HP-UX 10.20 & 11.00, AIX 4.2, AIX 4.3,
     WinNT 4, Win2K pro & WinCE 2.11 often with Tk800.022 &/| DBD-Unify
ftp://ftp.funet.fi/pub/languages/perl/CPAN/authors/id/H/HM/HMBRAND/

Reply via email to