> I (finally!) took a look at this patch, and it looks fine to me; any > bugs with it are worth the risk to have a real implementation of > scatter/gather IO, in my opinion.
Thanks, there are a couple tests checked into mauve that add further some coverage on this stuff > > One nit is that gnu_java_nio_VMChannel.c doesn't compile with -Werror > for me (Darwin/x86, gcc 4.0.1), due to some unused variables and > missing prototypes. I'll attach a patch that fixes these warnings. What's the next step? I don't have CVS access. Should I reissue a patch with your changes or would someone be able to apply both patches to CVS. Regards, Michael Barker.
