Re: A few questions on regtest/

On Mon, 21 Aug 2000 16:26:08 -0600 (MDT), Joey Smith wrote: >Hi, > 1) What is the purpose of runtests? It does the exact smae thing on my >Linux box as maketests... runtests is redundant. > > 2) Is there any particular reason that actuals/ is not in a .cvsignore >file? That would seem more logical to me. Sounds like a good idea. Or at least, I can't think of any reason why not... In fact, the master and compare scripts should not be there either, they can be recreated on users platform. > > 3) Would anyone be opposed to a top-level Makefile? It'd be sooo nice, >IMHO, to just say "make tests", and have it do all the steps (add new >tests, run them all, and then compare output). Only possible problem would be platform compatibility, Win32's make (or nmake) can be a bit limited. But a make file would be good because check for platform, some tests would need to be skipped, and maybe we could check which modules are installed. Which brings me to another question, how do we skip tests for which the modules are not installed? Ideally, the maketests script should 'know' which modules are not installed and then not create the tests for them, but I'm not sure how it would 'know'. Any ideas? > 4) Is there a preferred space for temporary files? Normally I'd use $TEMP... > 5) Has any of this been tested using perl+php on Win32? When I first set it up, I tested it on Win32, but then carried on in Linux. I'll do some tests under Win32 later today. Gonna think about how to guess installed modules.... -- jalal --