CPAN::Reporter

[Andy Armstrong]

On 3 Sep 2007, at 17:39, Eric Wilhelm wrote:
> # from Andy Armstrong
> # on Monday 03 September 2007 07:50 am:
>
>> CPAN::Reporter fails under 2.99_01 - I think because it screen
>> scrapes the test results and some of our text differs.
>
> It's the summary lines that throw it off:
>
>     for my $i ( reverse 0 .. $#{$output} ) {
>         if ( $output->[$i] =~ m{^All tests successful}ms ) {
>
> Admittedly, they throw me off as well.  Anything that doesn't say "All
> tests successful." very near to the end looks like a failure.  With  
> the
> summary lines, there's a lot of noise about skips, so possibly we
> should omit those in Test::Harness mode until we come up with  
> something
> better.

I'm see how close I can get with minimal patches to Test::Harness  
first. I may clean up our output in legacy mode too.

> Of course, we have no 'protocol' for exchanging test failure/success
> with things like CPAN::Reporter -- but maybe an archive or etc  
> would do
> the trick.

That's a whole different area :)

I guess we should offer a YAML summary option...

-- 
Andy Armstrong, hexten.net