|
From: | Paul F. Dietz |
Subject: | Re: [Gcl-devel] (random tester) Error in FUNCALL [or a callee]: Caught fatal error [memory may be damaged] |
Date: | Mon, 01 Dec 2003 04:58:03 -0600 |
User-agent: | Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030624 |
Camm Maguire wrote:
Greetings! I'm happy to announce what I believe to be a good fix for the issue uncovered by these two examples. Committed to 2.6.1 and CVS HEAD. Will continue to test. If all goes well I'd like to push out another stable release at this point. Of course I may be delayed inthis as Debian uploads are temporarily suspended.
This appears to have fixed the problem. Thanks! I can now run the random tester without irrecoverable failures. Doing so, I see failures that are apparently due to a common bug, illustrated by this example: (let* ((fn '(lambda (b) (declare (type (integer 3352138624 13120037248) b)) (declare (optimize (speed 3) (safety 1) (space 1))) (progn (catch 'ct1 (progn (setq b 11159349376) (throw 'ct1 0))) b)))) (funcall (compile nil fn) 4108962100)) ;; should yield 11159349376, but doesn't. This bug has come up several times in an overnight run (~50K tests on terms of 'size' 200). See MISC.198 and MISC.199 in ansi-tests/misc.lsp Paul
[Prev in Thread] | Current Thread | [Next in Thread] |