> One thought would be that while one hash algorithm may exhibit a flaw
> that allows arbitrary changes to the payload without altering the
> hash, it's extremely unlikely that two hashes would be affected in the
> same way.

This is the main reason we have more than one hash: it's possible to have 
collisions, especially with weaker hashes, where a bad file can be accepted by 
MacPorts.

> I don't think MacPorts actually verifies every hash that is provided
> in the Portfile.

MacPorts checks all listed hashes.

> I think the actual reason is to provide a backup hash if the first
> algorithm isn't available. Though, I'm pretty sure rmd160 and sha256
> have been available in OS X for quite some time, via openssl, python,
> perl, etc.
> 
> Hmm, apparently a year ago sha256 support was broken in MacPorts
> anyway, I'm not sure if that's been corrected.

It was corrected in MacPorts 2.0.0.

> It'd certainly be simpler to document if only one hash algorithm was
> "blessed", with all others marked for removal by a certain date /
> version.

Attachment: smime.p7s
Description: S/MIME cryptographic signature

_______________________________________________
macports-dev mailing list
[email protected]
http://lists.macosforge.org/mailman/listinfo.cgi/macports-dev

Reply via email to