Subject: Re: How to avoid GC in tight numeric test loop? (CMUCL)
From: rpw3@rpw3.org (Rob Warnock)
Date: Thu, 26 Jun 2008 21:39:01 -0500
Newsgroups: comp.lang.lisp
Message-ID: <0bKdnYzU3uZYz_nVnZ2dnUVZ_tzinZ2d@speakeasy.net>
Raymond Toy (RT/EUS) <raymond.toy@ericsson.com> wrote:
+---------------
| Actually I made a few mistakes.  The following one is better:
...
| Also, if you use this deftransform, you probably also want to modify
| RANDOM itself so that it computes random numbers in the same way.
| Otherwise, you get the weird result that RANDOM from the repl with the
| same arg and state may produce different answers than a compiled
| version with the same arg and state.
+---------------

Based on a reading of CLHS RANDOM, *RANDOM-STATE*, and MAKE-RANDOM-STATE,
I'd say that "weird" is not quite a strong enough adjective here;
"wrong" or even "badly broken" would be better.  ;-}


-Rob

-----
Rob Warnock			<rpw3@rpw3.org>
627 26th Avenue			<URL:http://rpw3.org/>
San Mateo, CA 94403		(650)572-2607