Geoff Hutchison (ghutchis@wso.williams.edu)
Thu, 11 Mar 1999 21:08:32 -0500 (EST)
On Thu, 11 Mar 1999, Hans-Peter Nilsson wrote:
First off, thanks very much Hans-Peter! I'll outline more
database-breaking changes this weekend. ;-) In particular, I'm hoping to
check in changes to make Gilles happy. The DocHead fields will optionally
be stored in a separate database, accessed through the DocHead() methods
and not Deserialize. This should be a win in indexing since updating
backlinks and anchors won't require reading in the whole DocHead.
> There's a quirk with the ResultList class; it handles DocIDs as
> strings, not int:s. I did not change this, but performance
> (speed as well as memory consumption) would probably be further
> improved if it was changed to handle int:s.
Yes. This will involve changing the Dictionary class too. Right now
dictionaries are hash tables on chars. So this means that integer keys
have to be converted to strings. :-( I checked in a minor improvement on
this a few weeks ago--when calculating the hash function, it realizes it
has an integer and uses that as the hash value. Fixing this will speed up
several areas of code.
-Geoff Hutchison
Williams Students Online
http://wso.williams.edu/
------------------------------------
To unsubscribe from the htdig3-dev mailing list, send a message to
htdig3-dev@htdig.org containing the single word "unsubscribe" in
the SUBJECT of the message.
This archive was generated by hypermail 2.0b3 on Thu Mar 11 1999 - 18:24:12 PST