Neil Conway <neilc(at)samurai(dot)com> writes:
> Do you think it would be worth the trouble to use both algorithms, and
> then test on the node tag of the first element to decide which one to
> use? (The assumption being lists are homogeneous).

Hard to tell.  Since I haven't seen any evidence that equal() on lists
is a particular hotspot, I'd lean against adding complexity and
maintenance burden here.

> One thing I've been wondering about is whether it would be worth ripping
> out the existing List code wholesale, and replacing it with something
> like the following:

I have already done something much like this in a few hotspots using the
FastList structure.  But notationally, it's a pain in the neck compared
to the existing List code.  If you can think of a way to implement this
without adding a lot of notational cruft, I'd sure be for it.  I'm not
for it if it imposes as much messiness as the FastList approach does...

			regards, tom lane

In response to

• Re: equal() perf tweak at 2003-11-03 22:12:19 from Neil Conway

Responses

• Re: equal() perf tweak at 2003-11-04 04:49:24 from Neil Conway

pgsql-hackers by date

Next:	From: Tom Lane	Date: 2003-11-03 23:38:01
	Subject: Re: adding support for posix_fadvise()
Previous:	From: Gaetano Mendola	Date: 2003-11-03 23:17:29
	Subject: Re: 7.4RC1 tag'd, branched and bundled ...

pgsql-patches by date

Next:	From: Tom Lane	Date: 2003-11-03 23:46:12
	Subject: Re: bufmgr code cleanup
Previous:	From: Larry Rosenman	Date: 2003-11-03 23:09:38
	Subject: Re: UW 713UP3 patch