I uploaded Inline::CPP 0.33_004 to CPAN a few days ago, and right out of the chute got four failures, all with the same issues:
1: "Can't locate Parse::RecDescent in @INC..." 2: "I currently only know about .... C, Foo, foo..." (M18 error) The reports were all from the same individual, and I'm CC'ing him as well on this to see if he can shed any light. However, I do expect to see the same issue turn up with at least a couple more testers. The CPAN regression statistical analysis tool ( http://analysis.cpantesters.org/solved?distv=Inline-CPP-0.33_004 ) is interesting, but the data-set for this version is still too small. However, the small data set may actually be working to our advantage at the moment, in that it is not cluttered by other forms of failure; only the one. That gives a 1.0 R coefficient for any element that is present in all failed tests. Nevertheless, I'm still unable to decipher why P::RD isn't found, when the metadata lower down in the reports indicates it's present. Furthermore, P::RD must have passed its own test suite, or we wouldn't be getting to this point (when a module dependency fails to install that doesn't show up as a primary module smoke test failure). I suppose what I should do is write a test such as the following: # 00prereqs.t use Test::More; use_ok( 'Parse::RecDescent' ); require_ok( 'Inline::C' ); done_testing(); ...so that I can better isolate the point of failure. If anyone has some additional insight I'm sitting on the edge of my seat in anticipation. ;) Thanks again! Dave -- David Oswald daosw...@gmail.com
