I was quite vague about "doing something weird". Specifically, I was alluding to installing subs in the caller package in such a way that they don't look like they've been imported. This can be achieved using Sub::Util/Sub::Name or eval, but Date::Manip isn't doing that.

I really do think your test script is having problems even loading Date::Manip. Is @INC the same for this pod test as for your other tests?

OK, I seem to have been indulging in some cargo-cult programming. My pod-covarage.t looked like this:

#!perl -T

use Test::More;
eval "use Test::Pod::Coverage 1.04";
plan skip_all => "Test::Pod::Coverage 1.04 required for testing POD co
+verage" if $@;
all_pod_coverage_ok();
[download]

If I delete the shebang line, everything is OK. I'm not sure where the line came from - I had to look up the meaning of the option '-T', so I don't think it was anything I added consciously. Strangely this line crops in other copies of pod-covarage.t for other modules where it doesn't cause a problem. So there is still something odd going on.

Cheers,

loris

"-T" is for taint mode. I don't think many people use it these days. It adds a little metadata to all strings to track whether they come from the outside world or not; the idea being that string data from the outside world is potentially untrustworthy, so shouldn't be interpolated directly into SQL queries, etc.

Date::Manip does check an environment variable (which counts as tainted data) as part of its initialization, so that could be it?

toby döt ink