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

List:       apache-test-dev
Subject:    Re: Test::More backend support
From:       Geoffrey Young <geoff () modperlcookbook ! org>
Date:       2004-04-26 19:16:19
Message-ID: 408D6003.5030205 () modperlcookbook ! org
[Download RAW message or body]


> or even better, since you can import plan() anyway, just use a forward
> declaration:
> 
> use subs qw(plan);
> 
> and then you don't need any new vars or mess with aliases. it'll be just
> there. which will make your code much simpler:

ok, I'll play with all of that.

> and one more nit -- The import() function may be called more than once,
> and every time it'll try to reimport all these symbols. So you probably
> want to have a flag so that you import them only once.

hmm, I guess so.  but I haven't seen other import() methods bothering with that.

> 
> Also what happens if -withtestmore is not the first argument to import()?

hmm, I dunno.  I just copied code from ModPerl::Const :)

> Regarding this comment:
> 
> +C<-withtestmore> tells I<Apache::Test> to use the I<Test::More> framework
> +instead of the I<Test.pm> framework behind the scenes.  Note that you are
> +not required to C<use Test::More> yourself with the C<-withtestmore>
> option.
> 
> What happens if I want to use functions from T-M, will this work?
> 
> use Apache::Test qw(-withtestmore);
> is ($foo, $bar);
> 
> I guess not. 

nope, that works just fine (see the test tarball :).  I wanted things to be
the same as they were with Test.pm - use Apache::Test and automatically get
all the Test.pm or Test::More stuff you were used to.  in the case of
Test::More, though, it's just a lot more stuff :)

> Usually you want Test::More for its functionality, not
> ok/plan/skip. So in 99.9% cases you will want to say:
> 
> use Apache::Test qw(-withtestmore);
> use Test::More;
> is ($foo, $bar);
> 
> no?
> 
> 
> ---------
> 
> 
> +        Test::Builder->failure_output(\*STDOUT);
> 
> I think Test.pm has it set to STDERR:
> 
> Test.pm:33: $TESTERR = *STDERR{IO};
> 
> so we probably want to keep things consistent, no?

ok, I guess you're right.  I didn't have it set at first and only saw stuff
in the error_log and thought it wasn't right.  maybe Test::More is putting
more/different stuff in STDERR?  I'm not exactly sure, which is why I put
that the interface was experimental in the comments - I have a feeling that
it's going to take some usage and tweaking before we get it all right, but
for the moment it's usable which is more than we had last week :)

--Geoff


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