Said of every good idea that can't play out before its time, I suspect.
I hoped (hoped!) that Test::Simple was included; it is not. I have the Activestate binary, and a few remarkably archaic discussions on Perlmonks have convinced me that testing is an animal I don't quite have the chops to tangle with yet.
Setting that aside for now, I move to *actually writing the friggin' script*. Lovely. At least I know a little more about modules and such; and found a few links to stuff about how to *make* modules.
Also, I saw Acme::Pony. I had honestly thought Acme::Buffy was the zenith of wacky, but...
Oh, and I *just* discovered COY! How amusing.
Damien wrote those
Coy and Acme:Bleach modules
They are hilarious
Personally, I don't have the time not to test -- writing a Slash plugin for an article due next week, I'd have caught several bugs sooner if I'd written my test cases first. I caught them only after investing more debugging time than I would have liked.
Re:Testing isn't Esoteric
chaoticset on 2001-12-31T01:27:39
I wasn't clear, and I apologize.First, I don't quite understand *when* testing needs to be done, where it happens, etc., like knowing basically what a transmission does but not realizing how it's actually applied within the mechanism of the whole car.
Second, I am woefully bereft of any experience installing modules. (Although I'm going to the link you supply after I post this clarification to see if I can rectify the situation.)
Believe me, when I *saw* the testing explained in the advent calendar, I was all for it. If I understand it properly, I've been doing it the whole time the hard way when problems came up. "...doesn't seem to bring a value over...well, I'll print the variable name and the variable...oh, and this doesn't seem to match properly...I'll dump the keys and mark them as coming from %inventory..." That kind of thing.
If I'm off base on my concept of it, please, any clarification is welcomed. I'll be doing more research on it as I go anyway, because it looks very helpful.
Re:Testing isn't Esoteric
chromatic on 2001-12-31T18:42:44
Easily enough explained (and no apology is necessary).Test Continually. Test before you add a feature. Test when you think you've added that feature. Test after you make any change. Test when you return to the program in the morning. Test, test, test.
If you test every feature sufficiently, if you refuse to accept anything less than complete test success, and if you run the tests after sufficiently small changes, you will have more confidence in your code, spend less time debugging, and be free to make greater individual improvements.
The point is two-fold: to prove (though not in a rigorous mathematical sense) that the code does what you think it does, and to prove that any changes you make do not break your existing expectations. You get better at this as you go along.
As for installing modules, perlmodinstall has the gory details. If you can unpack a tarball and find a make utility (both very easy on your Linux box and both possible on your Win32 machine), you can install pure-Perl modules with little fuss. The Test::Simple distribution certainly applies.
Re:Testing isn't Esoteric
chaoticset on 2002-01-01T05:56:32
Very nice. Today (after I wake up) is the day I get Test::Simple installed, then.:) Thanks again.
