Rosio Pavoris a blog

I gets presents! All of them came from ThinkGeek, because apparently nobody else ships to Belgium in less than three months.
I got an Einstein action figure, Tetris magnets (weakest magnets I’ve ever seen, though they’ll hold up a single piece of paper adequately), an RFID-blocking wallet thing, two of these (brain cell (which isn’t technically a microbe) and sleeping sickness), and these, which look really good on my copy of Origin but smell like death.

I was also supposed to get Real World Haskell, but apparently Barnes & Noble sucks and it hasn’t arrived “yet”. This only goes to confirm my suspicion that Haskell isn’t a real programming language, but rather a very elaborate running joke. You nearly had me with GHC, guys.

Anyway, I updated progscrape to address two nitpicking complaints and one marginally more serious one (though that post is mine), as well as fix the fact that it didn’t extract e-mails at all (despite claiming to).
The thread-parsing code could still do with a desucking, but I’m not going to do it.

The new code is here (diff), and since running that on your old database will cause all kinds of interesting errors, a new database is here (15.0 MB, 59.9 MB uncompressed). I could have written a script to just siphon data from the old database into the new one, but instead I rebuilt it from scratch so I could time how long it took:

real	123m19.681s
user	4m45.306s
sys	0m22.145s

That’s for 6,826 threads and 173,390 posts.

The main advantage is that it’s now possible to do things like

select * from posts where thread = 1228980536 and id = 50;

which used to require

select * from posts where thrid = 1228980536 order by time limit 1 offset 49;

which is considerably less secksy.

Ceterum censeo Rasmus Lerdorf is to be beaten to death with a hardcover copy of SICP.