apr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Graham Leggett <minf...@sharp.fm>
Subject Re: svn commit: r692949 [1/2] - in /apr/apr-util/trunk: ./ build/ crypto/ test/ test/data/
Date Mon, 08 Sep 2008 09:55:48 GMT
William A. Rowe, Jr. wrote:

> Critics/Advocates,
> 
> Please, let's review this submission for objectionable characteristics
> in the API and put it to rest in New Orleans, rather than last minute
> firedrills when someone decides to champion the release of 1.4/2.0?

+1 to that.

> I presume you've reviewed and addressed all of the previous objections
> of the apr_ssl interface before moving this development from sandbox
> back into trunk?

Yes, otherwise there would have been little point in the exercise. The 
whole thing was rewritten from scratch.

A key concern was that abstraction was even possible, and thus NSS and 
OpenSSL modules were developed simultaneously to prove that yes, 
abstraction is possible, and this time round a lot of work went into the 
test suite to show that the modules could interoperate, and under what 
conditions.

The Mozilla NSS community were very helpful in getting the NSS module to 
work, and a number of NSS fixes to error code paths have resulted from 
the effort.

The SSL code has been ignored for now, but not forgotten - it is far 
easier to digest bite sized pieces of code than swallow a large module 
whole.

I hope to translate any misunderstanding over how the API works into 
better documentation as well, so if anything isn't clear, please speak up.

Regards,
Graham
--

Mime
View raw message