Ways to waste time: remember to grab various modules. Remember to grab copy of db and remember to grab copy of websites.
Forget a few little modules here and there.
So I forgot to download Tie::Cycle and couldn't be bothered downloading it since I figured it would take next to no time to write. And true, it didn't.
Then I figured: hmm. I appear to be missing some other modules, that aren't quite so easy to duplicate. So I'll have a play with Tie::Cycle. Ended up making a lovely test suite for it (that makes use of Test::Inline). It handles errors better than brian d foy's does and handles assignation better (I ended up downloading his Tie::Cycle and comparing features - even found a bug in his test suite while I was testing my module against his suite).
So then I thought: let's make it relatively easy on people - rather than have my Makefile.PL call pod2test, better off having an automated system. So I ploughed in and rewrote most of my makemaker program.
It works better than ever. It'll find all the pms. Extract their tests, if any. Write an appropriate MANIFEST, Makefile.PL and README. And I'm just finishing off the reformatting of cvs log output (some people disagree that cvs log output should be a Changes file - I tend to structure my log messages, so they tend to make a bit more sense than others. I think it's fine). Even have a Pod::Parser thing going to have brian d foy's Changes file as part of the Cycle.pm pod (perldoc won't show it, but makemaker will use it and include it if it is there).
So all my CVS repository has for the module is Cycle.pm, makemaker (not actually tied to Cycle.pm, so I should really make it its own project), t/briandfoy.t and t/bench.pl (benchmarking of some alternative methods of cycling arrays). Everything else a normal distribution requires is generated by makemaker from various information sources.
Now I just need to send it to brian d foy and see if he laughs. It is, after all, one of the most useless things to do. Remember kids: if it ain't broke, don't fix it.
