Jeffrey Hutzelman wrote:
> More generally, it is not reasonable to assert that code which uses the
> DES API's (in this case, des_string_to_key) without calling rx_Init is
> broken. First, the DES API, which is not unique to AFS, does not have
> this requirement.  Second, rx_Init has significant side-effects, which
> should not be incurred for things that just want to use the DES library.
> 
> The brokenness here is in the DES library, and its inappropriate
> dependencies on rx and rxkad.  Fix that, and the other problems will go
> away.
> 
> -- Jeff

I didn't realize this was a side-effect of the thread local storage
changes.   I agree we must fix the DES library.

Jeffrey Altman

Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

Reply via email to