Re: Fw: Fixing SKIP:

2006-07-16 Thread Michael G Schwern
On 7/16/06, Jonathan Rockway <[EMAIL PROTECTED]> wrote: I retract my previous comment. My prototype was, in fact, completely bogus :) Let's make that array a reference to a list and everyone will be happy: can_ok($object, [qw(foo bar baz)], 'reason'); Hopefully that works and makes sense. If

Re: Fw: Fixing SKIP:

2006-07-16 Thread Jonathan Rockway
I retract my previous comment. My prototype was, in fact, completely bogus :) Let's make that array a reference to a list and everyone will be happy: can_ok($object, [qw(foo bar baz)], 'reason'); Hopefully that works and makes sense. If !ref $_[1], then just treat it as a single method name.

Re: Fw: Fixing SKIP:

2006-07-16 Thread Jonathan Rockway
> I do not think that prototype means what you think it means. It means what I think it means. Same syntax as we have currently, no. You would have to do this: my @methods = qw(foo bar baz) can_ok($object, @methods, 'reason') Instead of: can_ok($object, qw(foo bar baz), 'reason') Note that i

Re: Fw: Fixing SKIP:

2006-07-16 Thread Michael G Schwern
On 7/15/06, Jonathan Rockway <[EMAIL PROTECTED]> wrote: Worse, it's inconsistent with the rest of the API: ok$foo == $bar, $REASON; is$foo, $bar, $REASON; is_deeply $foo, $bar, $REASON; And then this: skip $REASON, $num; Sadly, it would be hard to change that since so many

Re: Fw: Fixing SKIP:

2006-07-16 Thread Michael G Schwern
On 7/15/06, Jonathan Rockway <[EMAIL PROTECTED]> wrote: What's the reasoning behind accepting an array, anyway? Convenience. You almost always use can_ok() with a list of methods. It also makes calculating the plan a little easier for it to be one test. I recall waffling around on the interfa

Re: Fixing SKIP:

2006-07-16 Thread Michael G Schwern
On 7/15/06, Ovid <[EMAIL PROTECTED]> wrote: I don't particularly like that this is a silent failure, but I'm not sure of a robust way to fix that. In any event, I reread the docs a couple of times before I realized I was being stupid. That suggests to me that this little nit could be improve

Re: Fw: Fixing SKIP:

2006-07-15 Thread A. Pagaltzis
* Jonathan Rockway <[EMAIL PROTECTED]> [2006-07-16 00:45]: > > can_ok cannot accept a description because it accepts a list. > > This could be repaired by re-prototyping can_ok as ([EMAIL PROTECTED]). I do not think that prototype means what you think it means. Regards, -- Aristotle Pagaltzis /

Re: Fw: Fixing SKIP:

2006-07-15 Thread Jonathan Rockway
>> Re-reading the man page, it looks like isa_ok and can_ok can't even >> accept a test description? > > can_ok cannot accept a description because it accepts a list. This could be repaired by re-prototyping can_ok as ([EMAIL PROTECTED]). What's the reasoning behind accepting an array, anyway?

Re: Fw: Fixing SKIP:

2006-07-15 Thread Ovid
- Original Message From: Jonathan Rockway <[EMAIL PROTECTED]> To: Perl QA List Sent: Saturday, July 15, 2006 9:33:42 PM Subject: Re: Fw: Fixing SKIP: > Worse, it's inconsistent with the rest of the API: > > ok$foo == $bar, $REASON; > is$foo, $ba

Re: Fw: Fixing SKIP:

2006-07-15 Thread Jonathan Rockway
> That's incorrect, even though saying "skip X tests" reads naturally > to me. Since "skip this many for tis reason" is how I mentally think > of SKIP: blocks, I keep writing them like that, even though it's > wrong. As mentioned, it fails silently. Worse, it's inconsistent with the rest of the

Re: Fixing SKIP:

2006-07-15 Thread chromatic
On Saturday 15 July 2006 11:35, Ovid wrote: > If the docs are updated to indicate that the skip message must not consist > solely of a positive integer, then we're OK.  Will that break anything out > there? Perhaps if you check both arguments, and then issue a warning if the first looks solely n

Re: Fixing SKIP:

2006-07-15 Thread David Wheeler
On Jul 15, 2006, at 11:35, Ovid wrote: That was my initial thought, but there's nothing explicitly wrong with having a numeric skip message. No, I said make sure that the *second* argument is numeric. It must always indicate the number of tests to be skipped. Best, David

Re: Fixing SKIP:

2006-07-15 Thread Ovid
- Original Message From: David Wheeler <[EMAIL PROTECTED]> > Perhaps skip() should be updated to ensure that the second argument > =~ /^\d+$/; That was my initial thought, but there's nothing explicitly wrong with having a numeric skip message. If the docs are updated to indicate tha

Fw: Fixing SKIP:

2006-07-15 Thread Ovid
- Forwarded Message From: Ovid <[EMAIL PROTECTED]> > I have returned to working on the tap parser and stumbled across this > irritation: > > #!/usr/bin/perl > > use Test::More tests => 1; > SKIP: { >skip "I'm lazy and don't wanna run no tests", 1 if 1; Well, naturally I had to

Re: Fixing SKIP:

2006-07-15 Thread David Wheeler
On Jul 15, 2006, at 10:52, Ovid wrote: That's incorrect, even though saying "skip X tests" reads naturally to me. Since "skip this many for tis reason" is how I mentally think of SKIP: blocks, I keep writing them like that, even though it's wrong. As mentioned, it fails silently. Perha

Fixing SKIP:

2006-07-15 Thread Ovid
I have returned to working on the tap parser and stumbled across this irritation: #!/usr/bin/perl use Test::More tests => 1; SKIP: { skip "I'm lazy and don't wanna run no tests", 1 if 1; ok 1, 'We don't get to here'; } That produces the output: 1..1 # No tests run! I don't