httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From TOKI...@aol.com
Subject Re: HTTP Protocol Question
Date Thu, 17 Apr 2003 13:52:20 GMT

> Greg Ames just pointed out that Roy said in an earlier on-list 
> conversation that RFC2616 did not disallow full duplex HTTP.  I must 
> have missed it even though I do recall starting a similar thread a while 
> back on a full duplex interface to mod_cgi, which is essentially the 
> same thing.
>
> Bill

Roger that... but I will see Roy's MAY and raise you a MUST.

As I just pointed out in last message... if you REALLY read
RFC2616 carefully then 'full-duplex' is non-optional to fulfill all the 
things
that MIGHT happen ( Unexpected 100 Continue's showing up,
etc. etc. ).

Full-duplex is actually REQUIRED and all HTTP transaction
processors really are SUPPOSED to be coded just like 
an SSL CONNECT passthrough where the handshake is
never gonna work unless you are feeding whatever one
side has to side IMMEDIATELY back to the other side.
( full-duplex ).

Later...
Kevin



In a message dated 4/17/2003 12:42:51 PM Central Daylight Time, 
bill@wstoddard.com writes:


> Greg Ames just pointed out that Roy said in an earlier on-list 
> conversation that RFC2616 did not disallow full duplex HTTP.  I must 
> have missed it even though I do recall starting a similar thread a while 
> back on a full duplex interface to mod_cgi, which is essentially the 
> same thing.
> 
> Bill
> 


Mime
View raw message