Re: RE: Crypt_HMAC Vs Message

--- Davey <davey@php.net> wrote: > Hi PEARs, > > It has always seemed odd that we have both Crypt_HMAC and Message. > > Message doesn't use the Crypt_HMAC package for its HMAC implementation > instead having its own. I think that Message is a) very ambiguous in > its name b) not necessary. Well, Message (other names tossed when I proposed were Digest or MessageDigest), was first released back around mid November 2002 (was discussed in pear-dev much earlier in September, see [1]), Crypt_HMAC was added in February 2003, and talking to Derick he mentioned that he did not realise that Message did all that Crypt_HMAC does [2]. At some point we talked about doing something along the lines of Cache/Cache_Lite, but this last years have been extremely eventful for me, so I did not get around to coordinate anything else with Derick. Bottomline, things could be better, and I am happy if they can be improved. > I think we need to think of deprecating the Message package in favour of > Crypt_HMAC. > > I also think that we should take Jesus' excellent Message_Hash_* > implementations and make them into a Crypt_Hash package which can then > be used by Crypt_HMAC (or anything else) to allow for more hashing algos > to be used. The HMAC algorithm in Message is also more general, being able to plug any hashing algorithm, so perhaps that could be ported too to a new Crypt_HMAC. > If anyone is in favour of this, I will do the work of making the > Crypt_Hash package if someone else with hashing algo experience will > take it over once I've done that? I'll also work with whoever maintains > Crypt_HMAC to implement Crypt_Hash in it :) Look at the scripts in CVS, most of the secondary classes are generated, just the base classes are coded by hand (mostly). Writing code that writes code is a good thing :-). One of the main problems w/ Message is the possible superabundance of objects, because of too finely grained design. So it might be a good idea to check whether such design should be reconsidered when extracting a new package (or packages) from the old one. I am still not sure that the Hashing algorithms should be in Crypt, but it just a matter of taste, they are more akin to digest calculations than encrytion algorithms. Now that I have some more time and am relaxing in Peru, I'll be more than happy to coordinate with Derick, Davey and whoever wants to pitch in the reorganization of the code bases. Cheers. [1] http://marc.theaimsgroup.com/?l=pear-dev&m=103308686415968&w=2 [2] See discussion startin at: http://marc.theaimsgroup.com/?l=pear-dev&m=104711612408965&w=2 ===== -- Jesus M. Castagnetto (jcastagnetto@yahoo.com) Web site: http://www.castagnetto.org/ Research: http://metallo.scripps.edu/ PEAR stuff: http://pear.php.net/user/jmcastagnetto __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com