site_archiver@lists.apple.com Delivered-To: darwin-dev@lists.apple.com Organization: University of Michigan User-agent: Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.4) Gecko/20030624 Netscape/7.1 On Sep 12, 2004, at 6:27 AM, David N. Williams wrote: I've had a problem with a rebuild of pfe-0.33.30 with gcc 3.3 under MacOS 10.3.4, which worked fine in 2.95.2 under MacOS 10.2.x. It seems to boil down to an illegal (I think) save/restore of one or more registers declared as global variables, which I've distilled into the example below. Among several register problems, my initial scan of Bugzilla doesn't show this, but maybe I missed it (or maybe I'm just misunderstanding something). Yes you missed that GCC has its own bugzilla and it was fixed for GCC 3.4.0, I cannot remember the bug number off hand. Thanks for the reply! It encouraged me to rescan the bugzillas. I did not find the particular problem reported in my message among the closed-bug reports, but of course I might have misssed it again. There was a lot about general register problems in gcc 3.x, and the discussion in #10684 and #7871 makes me think that a fix for the problems there could well include the problem here. I did learn from those bugzillas that you have been involved with the register problems, so I trust your memory! :-) I'll wait to see if clobbering by save/restore is indeed fixed in later versions of gcc. But since I don't use any gcc beyond the Apple developer tools, that could be awhile? I work around it now by turning off all global registers in pfe. -- David _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev@lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.appl... Andrew Pinski wrote: This email sent to site_archiver@lists.apple.com