On 8 Nov 2007, at 07:29, Andreas J. Koenig wrote:
Since 3.00 I have the following test output in my Image::Magick test
t/wmf/read..........
1..2
ok 1
ok 2
ok
You already have a parser for (t/wmf/read.t) at /home/src/perl/
repoperls/installed-perls/perl/pe1S7WD/[EMAIL PROTECTED]/lib/5.10.0/
TAP/Harness.pm line 412
On IRC AndyA and I already had a short exchange about this and Andy
cannot reproduce it. I can confirm that I see the other broken test
output from I:M in t/setattribute.t. But when I fix this in their
test, the problem with T:H output remains for me.
I've sent them a patch for t/setattribute.t. Bob Friesenhahn <[EMAIL PROTECTED]
> has accepted it and seems to be a good contact.
% make test
/bin/sh ../magick.sh PERL_DL_NONLAZY=1 /home/src/perl/repoperls/
installed-perls/perl/pe1S7WD/[EMAIL PROTECTED]/bin/perl "-
MExtUtils::Command::MM" "-e" "test_harness(0, 'blib/lib', 'blib/
arch')" t/*.t t/bzlib/*.t t/jpeg/*.t t/jp2/*.t t/png/*.t t/tiff/*.t
t/wmf/*.t t/wmf/*.t t/zlib/*.t
Because in the Makefile.PL it says:
foreach $delegate (qw/bzlib fontconfig freetype jpeg jp2 lcms png
tiff wmf x11 xml wmf zlib/) {
if( -d "t/$delegate" ) {
$delegate_tests .= " t/$delegate/*.t";
}
}
I see the "wmf" twice there. Everything beyond this point is not of
interest for Test::Harness. I'll take that bug over to Image-Magick.
Maybe an upgrade has this fixed already.
I think I'm not seeing that bug because IM is building without WMF
support on my machine.
So this is a new (to me) incompatibility: with old T:H it was OK to
run a test twice, with the new one it isn't. Maybe this restriction
could be lifted, I have no strong opinion. In any case the error
message could be improved, something like. "Did you call this test
twice?" might really help to nail such a bug quicker.
Definitely +1 to a better error message. I've added the text
"Perhaps you have run the same test twice."
after the existing error message.
The reason we error at all is part of a general fail-early philosophy
- it's a testing after all - so anything unexpected might indicate
trouble.
--
Andy Armstrong, Hexten