On Thu, 25 Nov 2004 11:32:40 +0100, Oliver Zeigermann
<[EMAIL PROTECTED]> wrote:
> Thanks for the patch, I really like it and would like to see it in 2.1
> final. The idea to limit path size to 128 is a nice idea, but you know
> there is the risk of  two different paths to be encooded to the same
> MD5 hash. Although this is rather unlikely...

Yes, there is that risk, but it is very small.  I keep the first 95
characters of the path, add an underscore, then add the encoded MD5
hash as the final 32 characters.  So not only do the paths need to
have the same MD5 hash, but they also need to share the same first 95
characters for a collision to occur.

You could add a unique ID like ~N to the end of the path to
differentiate between collisions, but the code is very localized and
seems to require a simple 1-1 mapping.

--Andy

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to