On Mon, Jul 09, 2007 at 01:51:37AM +0200, Eric Y. Kow wrote: > Anybody want to comment on this patch and the associated bug?
The patch looks fine and it does what the curl documentation says is needed. Besides making the code a bit more complicated it doesn't change anything (if curl would have copied the string), so it should be quite safe. Speaking against accepting it is we have only one report of this problem, and the patch, which in theory should have solved the problem, is reported to not fully have done so in practice. I still think we should accept the patch. The memory in question is a small chunk allocated in a gc-environment and maybe used for only a short time by curl. One is probably "lucky" if one's setup hits such a bug often enough to track it down. -- Tommy Pettersson <[EMAIL PROTECTED]> _______________________________________________ darcs-devel mailing list darcs-devel@darcs.net http://lists.osuosl.org/mailman/listinfo/darcs-devel
