Re: [serf-dev] Re: [PATCH] serf/ra_serf: add cansend callback to stop writing requests untill authn handshake is finished

On 8/30/07, Greg Stein <gstein@gmail.com> wrote:
> Note that HTTP specifies auth is per request, not per connection.
> Holding up a pipeline based on the faulty concept of conn auth is an
> incorrect design.

Well, yes, but SSPI is fundamentally at odds with HTTP as it does
authentication at the connection level. =(

> I could very much agree with a design that incorporates a dynamic
> scaling of the pipeline fill. i.e. start with a few and quickly ramp
> to 100 or whatever we believe the max request per child is set to.
> This still assumes auth is the same across requests, but that is an
> optimization rather than designing an API that assumes a conn level
> auth.

Perhaps scaling would be good, but I wonder how that'd effect us in
the common case where SSPI isn't being used?

At least with Lieven's patches, only SSPI users would stall the
pipeline - and that's because the underlying protocol is brain-dead.
Other folks wouldn't have that issue.

*shrug* -- justin

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri Aug 31 01:14:20 2007