On Thu, Nov 8, 2012 at 8:28 PM, Shmuel Fomberg <shmuelfomb...@gmail.com> wrote: > I disagree. A failure to install is a bug.
I respect your position, but this has been debated ad nauseum over the years. The general consensus is that CPAN Testers exists to see if distributions *build* and *test* correctly. Distributions that correctly specify prerequisites are providing a contract: "if these are available, then I will work". Therefore, when those prerequisites are not available, we do not consider it a valid test of the contract because the preconditions weren't met. Frankly, even a new "depfail" grade isn't terribly helpful to end users because the user has no idea if dependencies would fail for them without trying them. Did it fail because the dependency is fundamentally broken? Or only on that platform/perl? Or was an external library involved? Or did the dependency fail because it has some network test that timed out? There are so many reasons that dependencies could fail that I think a "depfail" grade winds up just being noise. David -- David Golden <x...@xdg.me> Take back your inbox! → http://www.bunchmail.com/ Twitter/IRC: @xdg
