Renaming AES to AES-128
bernhard at intevation.de
Fri Dec 10 18:11:38 CET 2010
Am Donnerstag, 2. Dezember 2010 12:40:24 schrieb Werner Koch:
> On Tue, 30 Nov 2010 14:40, bernhard at intevation.de said:
> > Leaving the primary identifier like it was, is a source of confusion
> > which also has a cost. What is the cost and consequences of the ABI
> > change?
> Debian alone lists 225 packages depending on libgcrypt. And there are
> for sure many more users. All these packages would need a rebuild and
> the maintainers need to go over it to figure out whether the change may
> affect them.
Can you explain briefly to me, why they would need a rebuild
and why ABI usage is really broken?
I mean applications with use gcry_cipher_open(x,GCRY_CIPHER_AES,..) and this
will just continue to work.
Okay gcry_cipher_algo_name(GCRY_CIPHER_AES) would return "AES-128"
instead of "AES", but does this really warrant a rebuild? The index number
just stays the same.
> You don't change an ABI if there is not a serious reason to do that.
> Those who want to list "AES-128" instead of AES may easily use a wrapper
> const char *
> cipher_algo_name (int algo)
> if (algo == GCRY_CIPHER_AES)
> return "AES-128";
> return gcry_cipher_algo_name (algo);
So we could do this for gpg2 --version?
> > deprecated so it will go away after X years. Could there be software that
> > would still stumble then?
> We can't know.
The list of supported algos will evolve over time,
why not change the behaviour here?
Managing Director - Owner: www.intevation.net (Free Software Company)
Deputy Coordinator Germany: fsfe.org. Board member: www.kolabsys.com.
Intevation GmbH, Osnabrück, DE; Amtsgericht Osnabrück, HRB 18998
Geschäftsführer Frank Koormann, Bernhard Reiter, Dr. Jan-Oliver Wagner
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 198 bytes
Desc: This is a digitally signed message part.
More information about the Gnupg-devel