Re: Possible patch to Test::Pod

"David E. Wheeler" <[email protected]> writes:
> On Jun 7, 2012, at 10:20 AM, Russ Allbery wrote:

>> That seems like something that maybe we could all agree could be done,
>> no?  This seems broader than just POD tests; the test failures may be
>> due to some minor bit of functionality, or a change in the API of some
>> other module for a part of the distribution that the person isn't
>> planning on using.

> How is the user to judge whether a test failure is serious or not?

Those who don't know how to judge are less likely to take that option and
will probably still ask if it's serious.

> Currently, when a user reports a failure to me, I can then tell them if
> it's safe to `force install`. But they generally have no idea how to
> tell the difference between an unimportant failure (e.g., bad expected
> path output in the test on Windows) and an important failure (the module
> will not work on this system).

I think you're too pessimistic about the level of skill of people
installing CPAN modules.  We all hear primarily from the users who are
confused, but I think a lot of people just quietly use CPAN and don't tend
to talk to us a lot, and if the only test failure is something like
docs/pod.t, I suspect they can draw the obvious conclusions.

-- 
Russ Allbery ([email protected])             <http://www.eyrie.org/~eagle/>