in reply to RFC: Math::StochasticProcess

I know you were being facetious -- at least I think so -- about the testers who failed your distribution. I just want to say that failing test reports are a *blessing*. Without the reports, you just have would-be users, who will rarely communicate with you, who have failed installs. They do tend to communicate with each other though and a dist with failing tests will get a reputation of being crap even if the underlying code is gold.

Replies are listed 'Best First'.
Re^2: RFC: Math::StochasticProcess
by SilasTheMonk (Chaplain) on Dec 14, 2008 at 23:31 UTC
    Yes I was not serious bout the CPAN tests. Specifically I was trying to keep this thread away from the subject of the 2/31 failures. I will probably need to contact those testers directly as I cannot reproduce their circumstances. Also the test reports are quite clear. Rather I was just looking for general feedback.

    edit: Also as I understand it the process is completely automated so it remains true that noone but me has probably looked at beyind the name so far.

[reply]

      I had a similar dist which was failing for users and I could not figure out why (through like 4 releases of just looking stupid twiddling stuff and waiting to see if the tests passed; mine always did). It was related to prereqs not being installed even though I was declaring them. I finally added explicit versions to everything as well as some modules that the prereqs called but my code didn't and that fixed it for testers.

      Your tests seem to be crashing at the junction of Exporter + 5.6 + Darwin (and maybe Params::Validate is related though its tester report doesn't seem to agree). You could make a note of that in the Pod/Changes/README until you get it sorted.

[reply]

In Section Meditations