On 3 Sep 2007, at 09:49, Ovid wrote:
Andy (Armstrong), aside from the fact that we now know that it fails miserably on VMS, do you know of any show stoppers or potential embarrassments?
I'm just trying to reproduce a problem I had at YAPC. I installed POE via cpan with T::H 2.99 installed and although the tests all seemed to have passed it tried to send a failure mail via CPAN::Reporter.
Assuming I can either fix that quickly or it turns out to be a non- issue then I'm not aware of any show stoppers.
I guess it'd be nice if a few of us could try installing a bunch of weird modules with T::H 2.99 installed today - if we haven't already.
(The only reason I mention "potential embarrassments" is that CPANPLUS and Module::Build have left such a bad taste in many people's mouths that I would like to limit the impact of the first alpha. We screw this up and I worry that there are many programmers who will refuse to get over their initial impression. Luddites on the cutting edge. How ironic. It's a silly concern, I know :)
I don't think it is silly. If we take people who are obviously interested in testing and break their testing tools they'll be quite right to be irritated.
So one thing that someone could usefully do before we release is to add an explanation of how to uninstall T::H 2.99 and revert to 2.64 somewhere prominent in the documentation (I'd say README, Test::Harness and TAP::Parser). If there's an obvious safety net people will be less reticent about trying it.
-- Andy Armstrong, hexten.net