imacat wrote:
On Tue, 31 Jan 2006 11:07:58 -0800
Tyler MacDonald <[EMAIL PROTECTED]> wrote:
I saw your
Class::Clone 0.01 released on Dec 25 2005, but no claim for it was made
on the [email protected] list ever before.
To the extent that one should register modules before uploading to CPAN,
the place to register them is [EMAIL PROTECTED] [email protected]
is more concerned with naming of modules, architecture of distributions,
and so forth.
(Or maybe I was wrong,
new packages should be registered somewhere else? Please correct me.)
I know CPAN allows packages files to be uploaded without registering
first.
Yes, it does. In fact, my guess is that the overwhelming majority of
modules on CPAN are *not* registered. Only two of mine are -- but that
hasn't stopped you from testing the ones that aren't registered ;-)
I think it's not a good practice, but I can do nothing on this.
But at least people should keep a good practice on this issue, by
registering package names before starting to use them.
I'm following the perl.qa version of this thread, and I looked at the
FAIL report your Linux box generated for Tyler's module. I, too, am
puzzled by why, if he had a certain version of Test::Exception specified
in his PREREQ_PM key in Makefile.PL, your CPANPLUS didn't go out to the
Net and try to install it.
I looked at some of the FAILs your setup gave to other CPAN modules
today. It seems like many are failing because they are looking for a
certain method in Pod::Man. This is a strange situation, because
Pod::Man *used to* be in the Perl core distribution -- but apparently
has *not* been core since Perl 5.8.3. Now, Pod::Man is part of a
distribution called 'podlators' -- which doesn't appear to be core in
5.8.7. So this is very tangled up -- which makes the FAILs you gave to
those *other* distributions legitimate.
Happy New Year and happy testing!
Jim Keenan