On Tue, 4 Feb 2003, John Peacock wrote: > > Please try them out and let me know what you think, what bugs you find, > > etc. > > I don't like the fact that I have to install DateTime::TimeZone blind (i.e. no > testing) before I can test/install DateTime. I would much rather that some way > be found to include a skeleton DT::TZ that can be included in DT just for > testing, so that it can be installed first. Then the full DT::TZ could be > installed. I'll look at your test scripts and see what I can come up with.
Patches welcome ;) The time zone stuff is quite bizarre, I admin. But I really can't think of a way to handle recurring rules (like ongoing DST rules) without being able to date math, for example. How else do you figure out the last Sunday of October? > A trivial NULL subclass works without a single failure. I am trying to think of > a clever way to rewrite your tests so that any subclass intending to be a full > replacement need only run the full DT testsuite to demonstrate compliance. > Currently it involves replacing: > > use DateTime; > > with > > eval "use $CLASS"; You could just use do something like s/DateTime([;-])/DateTime::Foo$1/g from the command line in the test directory. > where $CLASS is suitably set beforehand. I suspect there is some way to get the > Makefile.PL to automatically predefine a variable to pass to the testfiles to > automate this; I'll dig deeper into ExtUtils::MakeMaker and see what I find... Ah, the joys of customizing the build process with EU::MM. Yes, it can be done, but it's not pretty. Check out the Mason and Alzabo Makefile.PL scripts to see how I've done it in the past. -dave /*======================= House Absolute Consulting www.houseabsolute.com =======================*/
