- From: Jeffrey Mogul <mogul@pa.dec.com>
- Date: Wed, 24 Apr 96 14:09:09 MDT
- To: Paul Leach <paulle@microsoft.com>
- Cc: "'dmk@allegra.att.com'" <dmk@allegra.att.com>, "'jg@w3.org'" <jg@w3.org>, "'rst@ai.mit.edu'" <rst@ai.mit.edu>, "'http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com'" <http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com>
Paul writes:
I think this is good. This is easier than Jeffs suggestion, which allows
extra LWS in lots of places.
To effect this:
Change the first line of the definition of Full_Request in section 4.1
(Message Types) and section 5 (why are they duplicated?) from
> Full-Request = Request-Line
>to
> Full-Request = *( CRLF ) Request-Line
I agree that this would be a good way to make an explicit but
not particularly kludgey change for HTTP/1.1 implementations,
if they were the only implementations we need to worry about.
However, one would have to be quite careful to say that there
are different grammars for senders and receivers, because senders
have to avoid sending extra CRLFs to HTTP/1.0 servers. And I
think that having something like
Clients MUST generate
Full-Request = Request-Line
and servers MUST accept
Full-Request = *( CRLF ) Request-Line
might be more confusing than simply stating a specific
application of the robustness principle.
-Jeff
Received on Wednesday, 24 April 1996 14:16:16 UTC