# from Dave Rolsky
# on Tuesday 10 June 2008 15:07:
>> If they can't be bothered to run `./Build
>> testpod` and `./Build testpodcoverage`, that's not such a huge deal
>> because it won't ship until those pass. Certainly it isn't worth me
>> maintaining ~100 files which might have the same content.
>
>Well, I can say that _I'm_ not going to bother running those two
> targets!
Fortunately, my pre-release ritual is automated and runs them for me.
If you were me, you would do it the way I do.
> My pre-release ritual does include running "prove -lr", so
> having pod.t and pod-coverage.t among my tests is quite important for
> me.
And *not* having them among my tests is quite important to me.
>I've gone to the trouble of disabling them for normal installs just to
>shut up false failures.
I don't want to maintain redundant, duplicate instances of those two
files in dozens of distributions.
Including them as test files has the advantage of being simple, but the
disadvantage of being simplistic.
The CPANTS metric assuming the names of these two files or whatever
heuristic about the contents has the disadvantage of being incorrect.
While it *might* indicate that I tested the pod/podcoverage, it doesn't
actually say that I did. It also might mean that these tests will
generate spurious failures on the client. While in the absence of
these (redundant, duplicate) files, it completely misses the fact that
*I* already tested the pod/podcoverage and further: gives me absolutely
no way to indicate that.
And further still: the metric is based on finding the use of the
"Test::Pod" and "Test::Pod::Coverage" modules, which provides zero room
for improvement even if you want to engage in the questionable practice
of duplicating information by including a couple of "blah.t" files in
every distribution.
--Eric
--
I arise in the morning torn between a desire to improve the world and a
desire to enjoy the world. This makes it hard to plan the day.
--E.B. White
---------------------------------------------------
http://scratchcomputing.com
---------------------------------------------------
