----- Original Message -----
From: "David Oswald"
But here's one that somehow didn't get the message:
http://www.cpantesters.org/cpan/report/d4bf9379-6d64-1014-8708-fa3e1267b5eb
I notice that we get the messages about the missing P::RD *twice*.
Once for each of the 2 times that it goes through the loop that Patrick's
patch introduced ?
And, having gone through that process twice it then finds that
$config{languages}->{$o->{API}{language_id}} is still undefined, and rightly
emits the M19 error message.
I'm hoping that it's the unlocatability of P::RD that's leading to
$config{languages}->{$o->{API}{language_id}} being undefined ... but haven't
actually verified that yet.
It also says that it's failed to find Parse::RecDescent, but lower
down in the report mentions that Parse::RecDescent is installed. If
anyone has any thoughts I'm scratching my head on this one.
I'm quietly confident that once P::RD gets found, things will run correctly.
In any case, that's the first thing that needs to be fixed.
Note that the tester's @INC specifies:
C:\strawberry\cpan\build\Parse-RecDescent-1.965001-bhkLjV/blib/lib
I'm guessing that's where P::RD is being found (on those occasions when it
*is* found).
But then, that location is not even in the search path when P::RD is *not*
found - and I'm guessing it's not being found for that very reason.
I'd start by contacting the tester, and see if he has any ideas as to why
his @INC changes when the tests are being run. I would think it's something
to do with his setup - but perhaps one of the Inline modules (or some other
module) is also doing something it oughtn't.
Cheers,
Rob
