On Thu, Jun 11, 2009 at 2:30 PM, Petr Rockai <[email protected]> wrote:
> Hi, > > I have flipped the mmap package to "mmap". Reasons: > > - we need writeable mmap for hashed-storage, and bytestring-mmap doesn't > - mmap, unlike bytestring-mmap, works on win32 Only because I'm curious, do you know how it works on win32? > > - it's probably better to unify on a single package and API instead of two I agree. > > > (the testsuite is still running, but I guess this should be relatively safe > as > long as it typechecks) I would be most interested in a performance and/or a stress test myself. Like you, I would hope it is relatively safe but I wonder if it helps performance. It could be that bytestring-mmap has some cleverness that speeds up the bytestring case. Eric, FWIW, I looked over the patches and they look fine at a cursory glance but I didn't try to apply them or test them. Jason
_______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
