This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH]: random global function name
rth@redhat.com (Richard Henderson) wrote on 08.07.03 in <20030709000930.GI481@redhat.com>:
> On Tue, Jul 08, 2003 at 04:24:09PM -0700, Zack Weinberg wrote:
> > A base64 encoded MD5 sum is only 22 characters; that might be a good
> > choice (it would of course be necessary to replace / with _ in the
> > character set)
>
> That's not really any better, IMO. Just because a collision
> is less likely doesn't mean it can't happen.
I'm pretty certain a md5 collision is far less likely than just about
every reasonable expectation you can come up with. This fear is entirely
spurious.
http://www.squid-cache.org/mail-archive/squid-users/200010/0725.html
Real-World testing in Squid. No collision found.
http://www.theneoproject.com/main.asp?PageRequest=ABOUTMD5
"Should our server ever detect two identical distinguished points, we
will begin searching from the previously submitted distinguished point
of each of the colliding points. Nobody has ever found a collision (two
input which produce the same output) in the MD5 hash . So finding a
collision would be very significant discovery."
http://sources.redhat.com/ml/gdb-patches/2001-07/msg00745.html
"Considering it was thought to be good enough for global duplicate
elimination by the DWARF2.1 committee, i think it's good enough for us.
Besides, you'd need more than one collision in a single file to have
any real effect on the debug info."
It's also what rsync relies on.
MfG Kai