[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

base64 (was: Re: tools location?)

From: Greg Stein <gstein_at_lyra.org>
Date: 2000-12-12 00:57:38 CET

Let's not discuss what is right, or what isn't. This is a simple problem of
somebody sitting down and writing a bit of code to bring them together.

Base64 encoding was not in APR before (it will be in APRUTIL). Therefore,
GregH needed *something*. I pointed him at the code in Apache (at the time)
and the interface wasn't quite right. There was never a chance to *share*
code, but merely to copy what had been done before.

Now that we have a chance to share code, then we can. 'nuf said. Any one of
us can solve the sharing / duplication problem with little effort.

[ and yes: neon could use APR(UTIL)'s base64 and md5 functions if someone
  were so inclined, and Joe was willing to pick up the dependency; he has
  indicated that he's okay with the choice of using APR sockets, but I'd
  guess he meant a config time option, rather than a hard dependency ]

Cheers,
-g

On Tue, Dec 12, 2000 at 12:42:54AM +0100, Branko Cibej wrote:
> Greg Hudson wrote:
>
> >>> GregH built one that operates in a streamy fashion. The APR version
> >>> does not (yet).
> >>
> >> Then I like GregH's (yet). :-)
> >
> >
> > base64 encoding is one of those things where the implementation is so
> > small that the interface dominates the code required to do it. I
> > don't know if we'll ever get convergence unless we put the concept of
> > "generic streams" (svn_{read,write}_fn_t) into aprutil.
>
> What for? O.K., you'll still need a base64 editor for filtering the
> delta stream, but it'll just pass chunks of data to the en/decoder in APR.
>
> (Yes, I know that base64 is a ridiculously small bit of code. But I'd
> rather avoid duplication, anyway.)
>
> The same can be done for hashing, by the way (md5, SHA-1, RC5, ...),
> except that the en/decoder must accept a hash seed so that it starts
> each chunk off of the previously computed hash value. Now that's
> definitely one thing we don't want to duplicate.
>
> --
> Brane Cibej
> home: <brane_at_xbc.nu> http://www.xbc.nu/brane/
> work: <branko.cibej_at_hermes.si> http://www.hermes-softlab.com/
> ACM: <brane_at_acm.org> http://www.acm.org/
>

-- 
Greg Stein, http://www.lyra.org/
Received on Sat Oct 21 14:36:16 2006

This is an archived mail posted to the Subversion Dev mailing list.