On Sun, Apr 30, 2000 at 09:36:23PM -0700, Russ Allbery wrote:
> Thanks! I did test that when I originally wrote Pod::Man, but only with
> Solaris *roff which doesn't have this problem.
This raises the question: might some person out there with spare tuits
be willing to work on some pod2xxx test-cases for the various pod2xxx
and/or Pod::Xxx modules? One could look at the existing t/pod/*.t
test scripts and see if most of the same input files could be used
but with different expected outputs for each filter. (Perhaps the
individual could even improve/replace the current "hack" I used to
compare results files modulo OS/platform differences in pathname
delimiters ;-)
Adding an '=include' to POD (either officially or unofficially) could
help with this. The current t/pod/*.t scripts actually implement an
'=include' solely for the scripts that do the testing, but this could
be made more generally useful (if p5p can actually agree on it with
only minimal bloodshed :-) IMHO the thorniest part is deciding on
rules/semantics for @PODINC and "pod-include" directories to implicitly
or explicitly look for included pod-files
Cheers!
--
Brad Appleton <[EMAIL PROTECTED]> http://www.bradapp.net/
"And miles to go before I sleep." -- Robert Frost
