* Eric Wilhelm <[EMAIL PROTECTED]> [2008-02-25T12:53:03]
> # from Ricardo SIGNES
> # on Monday 25 February 2008 04:59:
>
> >Basically, it lets you say "author_tests('xt')" in your Makefile.PL
> > (using Module::Install) to have a directory (or directory tree) of
> > tests run only by the module's authors.
>
> Now we only have to determine 'author'ness. I was thinking something
> along the lines of a capabilities/willingness registry for testers
> would be the way to handle it. That doesn't address the on/off switch
> for extra so much as which bits to test (do I have a chroot, ramdisk,
> etc. available.)
>
> But: I dislike the 'authortests' naming. These are simply "extra"
> tests, so whether you're the author only matters in terms of your 'make
> dist' default -- that is: a volunteer tester should be able to run as
> many of the extra tests as desired, while an author must run them all.
The name AuthorTests makes sense and should not be objectionable given the
mechanism used. It uses the "is the current directory an author's development
workspace, or a dist from the CPAN?" mechanism built into Module::Install. I
know that there was talk about PERL_TESTER_RUN_EXTRA_TESTS=1 and so on, but I
don't care about that. In fact, I don't ever want anyone who isn't hacking on
the module from a git checkout to run those tests, so it works very well for
me. Obviously, YMMV.
--
rjbs
