mdbm library source

Chris Torek chris at umcp-cs.UUCP
Mon Aug 20 04:26:36 AEST 1984


The only ``V7 code'' that mdbm contains is in hash.c.  I figured the
tables must have been published somewhere, in articles on extensible
hashing.  If I was wrong, well, everyone who doesn't have a V7 (or
equivalent) license, go delete the tables from hash.c.

(Out of curiousity, does anyone know where the numbers really *did*
come from?  The only requirement for the hashing function is that it
generate fairly well distributed bit patterns.  It's important that the
number of 1s and 0s in any particular bit position be fairly evenly
distributed.)

The rest of the code itself is (as far as I know) my own.  I suppose
the big problem is that I read, studied, and finally decoded and
rewrote, the old dbm code about a year ago or so; therefore, anything I
wrote since then may incorporate ideas from it.  Certainly they have
the same general structure.  (There aren't many ways to implement
extensible hashing.)  But the mdbm code is NOT the same as the old dbm
code.  (For one big thing, the method of linking keys to data is
entirely my own creation.  I think.)

Anyway, I don't really care who uses mdbm for what, but if it does turn
out to be restricted under V7 license, I ask everyone to please destroy
any illegal copies.
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci (301) 454-7690
UUCP:	{seismo,allegra,brl-bmd}!umcp-cs!chris
CSNet:	chris at umcp-cs		ARPA:	chris at maryland



More information about the Comp.unix.wizards mailing list