[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Compilation of methods per class.
- To: gregor@parc.xerox.com
- Subject: Compilation of methods per class.
- From: Jon L White <jonl@lucid.com>
- Date: Tue, 25 Sep 90 23:05:08 PDT
- Cc: lange@CS.UCLA.EDU, rivieres@parc.xerox.com, faieta@parc.xerox.com, gerson@parc.xerox.com, brotsky@parc.xerox.com, bobrow@parc.xerox.com, kimber@parc.xerox.com, cutting@parc.xerox.com, robertso@parc.xerox.com, rodrigue@parc.xerox.com, rao@parc.xerox.com, haarslev@parc.xerox.com, cloops-arisia@arisia.Xerox.COM
- In-reply-to: Gregor Kiczales's message of Tue, 25 Sep 1990 21:11:58 PDT <90Sep25.211206pdt.269@spade.parc.xerox.com>
re: There was a horrible bug in Victoria Day PCL which had symptoms like
what you are mentioning --- an entry could be in the cache, but there
would still be a cache miss. I recall that bug being fixed, but I don't
know if the patch was widely distributed.
The behaviour I'm referring to was present in previous versions too
(at least one -- maybe more). It wasn't really a bug in that it would
still correctly find the cached entry; but it was a performance pitfall
in that it had to go out of line at unpredictable times to do it.
Lucid's CLOS uses a different approach to generic function "caching",
which doesn't follow any of the PCL versions nor of that referred to in
the L&FP90 paper. Rather, it is more akin to TICLOS's. Method-functions
and other data are being "cached" (in a general sense) from the lookup
phase; but the data-structure in which the cached entries are stored is a
hash-table rather than a cache.
-- JonL --