Philip M. Gollucci wrote: > Stas Bekman wrote: > >> Philip M. Gollucci wrote: >> [...] >> >>> I was going to add some Apache-Test tests using Apache::TestMM (per >>> ask's requests), but I couldn't figure out how not to kill the >>> existing test suites if Apache-Test isn't installed (for mp1 users >>> this will usually be the case as its not in the core mp1 core) >> >> >> >> Just require Apache::Test. In the short time to come most Apache >> modules will use Apache::Test to do the testing anyway. >> > Excellent! I'll see what I can come up with.
BAH! alright, call me a pain, but I really dislike the idea of lumping test modules in prerequisites - Makefile.PL prerequisites are for modules that are required for the module to operate correctly. the test suite is something else entirely :) see one of my recent Makefile.PL for an idea on how to go about this without requiring Apache-Test, for example http://search.cpan.org/src/GEOFF/Apache-HEADRegistry-0.01/Makefile.PL of course, this is really being academic and there's no reason you can't use PREREQ_PM. but don't let PREREQ_PM lure you into a false sense of security. for example, just finding Apache-Test isn't sufficient to know that mod_perl is available within httpd - you'll want to check for mod_perl.c in your *.t tests and make sure that httpd will always start by using appropriate <IfModule> calls in your config. anyway, not to rant. this is just one of my little nits, is all. but I figure if I can convince people why the runtime environment and testing environment are different in enough different ways people will start to think a bit about why they do what they do and everyone will be better off :) --Geoff
