[Markus == [EMAIL PROTECTED] on Mon, 20 Dec 2004 13:59:57 +0200] Markus> I was wondering if it would make sense to add the original Markus> 're_tests' file to parrot distribution, with a script which Markus> autogenerates 're_tests.t' from it.
To me, it certainly would. I would note, however, that my implementation of something like this approach was rejected prior to the release of PGE as too complicated, despite a working code base. Please sync up carefully with Patrick and Dan on your approach to this automation to avoid rework. It may be that it's still preferred to just maintain your "cooker" script outside the tree and check in only the results. (Briefly, my approach was to have a "p5_re_tests.t" script, which would generate 1 PGE test for each line in the re_tests source file, but do that testing on-the-fly, not "cooked", so no regeneration would ever be necessary.) The argument for avoiding this approach didn't really make sense to me at the time, as I saw a future with thousands of PGE tests. I believed (and still believe) that given more time, it would be realized that automating the infrastructure so people didn't have to type in or port repetitive regexp/rule tests would be beneficial and result in more, better tests. If you're on a roll, Markus, I'd like to point you to: svn://software.wordzoo.com/p6ge-test-harness/trunk where I gathered lots of additional tests (including perl 5's re_tests file) that were intended to be ported to PGE. The codebase is abortive at best, as I simply dropped it when Patrick indicated he was going a different direction. So I've since applied my open-source time in other directions, but you are welcome to browse this obsolete codebase for ideas. -- [EMAIL PROTECTED] "Better to be of a rare breed than a long line." -- TDK
