[Catalyst-dev] Smoke tests - summary of things which don't work against 5.80

[J. Shirley]

On Mon, Oct 27, 2008 at 3:50 PM, Tomas Doran <bobtfish at bobtfish.net> wrote:
> I've been smoke testing most stuff in the Catalyst repository against
> Runtime 5.80 to see what has been broken. I'm hoping to expand to a point
> where I'm testing everything in the repos soonish..
>
> http://catalyst-smokes.bobtfish.net/
>
> From my last run, there are some interesting conclusions. Most of the
> failures seem to be groupable, meaning there are two or three big things
> stopping things working at the moment - the most obvious ones being:
>
>   - Not being able to find mk_accessors
>   - Not having NEXT loaded
>
> There are also a couple of common failures in several of the HTTP engine
> modules which may indicate real bugs in 5.80 trunk, and a couple of other
> failures that look 'real' - although I haven't looked into these any deeper.
>
> I've included a useful snippet from the test output below, for each of the
> modules which don't quite work (but don't fail due to dependencies or some
> such).
>
> Catalyst-Action-REST:
> #     Error:  Could not load class (Test::Serialize) because : Could not
> load class (Catalyst::Action::Deserialize) because : Can't locate object
> method "mk_accessors" via package "Catalyst::Request::REST" at
> /home/chimps/repos/Catalyst-Action-REST/0.60/trunk/t/../lib/Catalyst/Request/REST.pm
> line 45.
>

What's the preferred fix for this?  Is mk_accessors just something to
implicitly inherit from Class::Accessor and be done with it?

I can bake a release to get this working for 5.80

-J