On Tue, 17 Apr 2001 22:53:24 CDT, "Craig A. Berry" wrote:
>A late change (later than foolperl) broke this test for us, though
>given the mess we've made in there it's not too surprising something
>like this would happen. I think (hope) bleadperl has a much
>cleaned-up version of this test.
Ironically, I only made the changes to pod/*.t because I knew neither
t/TEST nor t/harness actually ran any of those tests, and I ass_u_med
VMS wouldn't run them either because I saw some VMS code in t/TEST.
Note that these tests still aren't run in 5.6.1 on any platform that
uses TEST or harness, so I felt pretty safe in making the change.
The change in question is this one:
http://public.ActiveState.com/gsar/APC/5.6.1/diffs/9624.gz
Now, it's still not clear to me how the tests passed in foolperl on VMS
when @INC wasn't set correctly in the pod/*.t tests. Are the tests
being run on VMS with the moral equivalent of "perl -I../lib ..."?
I guess the moral here is to integrate the VMS test harness into
t/TEST or t/harness, or something.
Sarathy
[EMAIL PROTECTED]
