On Wed, 12 Jan 2011, Gary Maxwell wrote:

>> Regarding test 560 (use of axTLS with the multi interface), it turns out
>> that axTLS does not support non-blocking read/write operations. According
>> to Cam (axTLS maintainer) this is by design to keep the binary footprint
>> and memory use to an absolute minimum.
>
> I'm wondering how much added code and complexity non-blocking support would
> actually entail? An extra 2-5%, maybe? Couldn't this be supported as a
> compile-time option when building axTLS?

In lots of areas I think axTLS suffers a bit from having been developed almost
exclusively by one person with a firm opinion on how to do things.

I've had private discussions with him about details I think should be changed
and I urged him to start the mailing list to at least start offering people
like us a possibility to particiate in the project and provide feedback,
comments and ideas.

I think the axTLS library and project will gain a lot if we provide comments
and feedback there.

(This is not criticism of axTLS, from what I've seen it seems like a great
library.)

-- 
  / daniel.haxx.se
-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-library
Etiquette:  http://curl.haxx.se/mail/etiquette.html