[Python-Dev] is this obselete?: [Patch #100888] Fix UCNs mach ine with >= 32bit longs

Bill Tutt billtut@microsoft.com
Tue, 25 Jul 2000 17:27:09 -0700

> From: 	Fredrik Lundh [mailto:effbot@telia.com] 

> note that if #100899 goes in instead, Bill's tweak of Andrew's
> original code won't be needed for anything related to unicode.

> the perfect hash module is still a useful tool (not that I ever
> got it to do a decent job on the unicode database.  Bill's either
> very lucky, or he's got a really fast computer... ;-)

The secret to getting it to work is to start with a high step rate, find
your initial solution so you can test code. Then you set the code up to
search for a smaller solution while you're off doing weekend things. Given
the number of Unicode character names there wasn't any other way to do it.
Smaller key spaces may not care about finding a smaller multiplier solution.

I do like my work computer though. Dual 733Mhz with 512MB of ram. Have to
have lots of ram to test a web app that caches 30 MB of data out of the
database, while still running DevStudio, VSS, etc.... :)