Busy Busy!

chromatic on 2001-10-11T05:48:26

What's this oddly-named guy doing lately? Well, the book editor preferred chapters one and two as they were last Thursday, as one humongous chapter one. I spent Thursday afternoon and Friday splitting them apart! Mine is evidently not to reason why, mine is just to type in vi.

After earning certain fame, I went on to a few other things. First came Devel::Constants, which endured a name change. I think it's a solid idea. perl-modules does not exactly agree. Hey, if Damian can get away with evil ideas, why not let me let people localize constants? (Of course, Damian could probably remember the name of that otherwise forgettable ST:TNG episode where Q loses his powers and thinks that changing the gravitational constant of the universe is an acceptable solution to the problem of a moon crashing into a planet.)

A gentleman named Doug sent feedback today on my DBI is OK article. He noted that there's no way to call finish() when using my bind_hash() subroutine. Oops. Change the return to:

return (\%results, sub {
    return $sth->finish() if @_;
    $sth->fetch();
});

Then you can just say $fetch->('finish') or $fetch->(1) or whichever argument you prefer instead of fetching the entire result set.

(Gosh, pudge, the CODE tag works elsewhere. Don't tell me I'll be trawling through the Journal plugin tomorrow to send you guys another patch just to make my life easier!)

It's nice to know people are reading it (Tim Bunce sent some very kind words when it first came out), and nicer to know that the ideas are useful. Next time maybe I'll think through all of the possibilities first.

Yeah, right. Then I wouldn't be a programmer. I'd be God.

Next up is Regexp::English, which needs a bit of a rethink for capturing matches and a bit of documentation before public release. It's almost there.

I'd also like to bolt a nicer interface on to my rewrite of File::Find. Blame it on gnat. It is 50% shorter than the previous version, and passes all tests on my x86 Linux box.

Then there's an idea for a Parrot disassembler, which would be the basis for a bytecode analyzer like the Stanford checker...

But I should leave the house tomorrow, at least for a few minutes. They don't deliver groceries here. It's probably for the best.


CODE

pudge on 2001-10-11T11:47:38

CODE is problematic. We cannot allow it to do what CODE actually does (and I am too tired right now to recall all the reasons). Right now I have a hacked-in hack of a hack for a LITERAL tag.


    use D'oh;
        if ($DEBUG) {
       slap_forehead();
    }

It works fairly well. There's surely a better way, but it isn't just "adding CODE support".