Re: [Perlunit-users] Solution for the problem

[Adam S. <ad...@sp...>]

I'm back!  And this time I really am going to get some stuff done,
because my manager has scheduled some time for me to get our test
suite up to scratch.

So, time to catch up.  Where were we ...

Regarding broken overloading in Error.pm,
Matthew Astley (mc...@us...) wrote a long long time ago:
> On Wed, May 16, 2001 at 05:55:57PM +0200, Christian Lemburg wrote:
> > I think the other approach (diff files and requiring users to patch)
> > could be a major hurdle in the install process. People just don't
> > want to think about such issues.
> 
> Absolutely. Being deliberately mean isn't my favourite either.
> 
> Can we be helpful? Is it possible to shove an overload into Error.pm
> (with big flashing warnings, of course) from inside our code, if we
> detect an old version? I've not played with overload much.
> 
> The pragmatic solution is to patch as Andrew suggests, then write a
> perlunit selftest which a) checks the functionality and b) fails with
> a suggestion to revert the patch if it detects a newer Error.pm

I notice that Error.pm now has a new maintainer, and the latest CPAN
version 0.15 fixes this problem.  Therefore it's probably best just to
add a self-test which checks for brokenness.