[prev in list] [next in list] [prev in thread] [next in thread] 

List:       perl-qa
Subject:    Re: Strange interaction between new Test::More and Test::Builder::Tester
From:       Buddy Burden <barefootcoder () gmail ! com>
Date:       2012-02-17 18:03:32
Message-ID: CAL0mxoTV2-pDG0qhdKnU8qLg-FrZkYEk0Ddw-Q__uQH0pGbsew () mail ! gmail ! com
[Download RAW message or body]

Ovid,

> I'm not sure what's going on here. You've mentioned the Test::More, \
> Test::Builder::Tester, Test::Tester and Test::File.

Sorry; perhaps I overexplained.  This is a problem between Test::More
and Test::Builder, like the subject says.  The other two are
irrelevant.

> I don't know exactly what is causing the problem you have with your tests, but if \
> you check the TAP::Parser::Grammar \
> (https://metacpan.org/module/TAP::Parser::Grammar) you'll see that the SKIP (and \
> TODO) directives are case-insensitive. Thus, both SKIP and skip should be fine. If \
> something is marking that as a failure, it's probably ignoring case-sensitivity for \
> directives or it's expecting an exact text match.

Okay, that puts the blame squarely on Test::Builder::Tester then.  It
definitely is expecting an exact text match.  Going through the code a
bit, it looks like it would accept a regex (although this is not
documented in the POD) so I suppose I could code defensively enough to
work around it, but it seems like if the standard says case
insensitive, it might be better for Test::Builder::Tester to ignore
case altogether.  Perhaps there are subtleties that would break other
stuff though ...

> In any event, I can't tell how to reproduce the issue from the plethora of modules \
> you've listed. Can you send a small code example of a test failure?

This code:

{
local $^O = 'Win32';

my @subs = qw(
    file_mode_is file_mode_isnt
    file_executable_ok file_not_executable_ok
    );

foreach my $sub ( @subs )
    {
    no strict 'refs';

    test_out("ok 1 # skip $sub doesn't work on Windows!");
    &{$sub}();
    test_test();
    }

}

works on Test::More (say) 0.98, but not on Test::More 1.005000_002.

I think I'll raise an issue on Test::Builder::Tester's github and see
what schwern has to say.


            -- Buddy


[prev in list] [next in list] [prev in thread] [next in thread]