I can't speak for the official position. The primary interoperability problem is due to CGI 1.1 (and earlier). CGI REQUIRES a CONTENT-LENGTH. Some servlet engines also have this restriction. Another possible problem is due to proxies: a proxy server is allowed to coalesce all the chunks into a monolithic request. In practice, you never see this, because it could require ulimited buffer space on the proxy server, and because it has bad effects on response time.
-Carl
"Mark Nottingham" <mnot@...>
06/10/2002 10:14 PM
Please respond to http-compliance
To: <http-compliance@yahoogroups.com>
cc:
Subject: [http-compliance] Chunked POST
Can someone remind me what the state of the art is re: chunked POST? From
what I can see, it's allowed, but there's not a good interoperability
story with 1.0, so it's mostly avoided.
What's the official position, and what can a client reasonably expect if
they chunk a POST and there's a 1.1 server somewhere in the message path?
Thanks,
--
Mark Nottingham
------------------------ Yahoo! Groups Sponsor ---------------------~-->
Tied to your PC? Cut Loose and
Stay connected with Yahoo! Mobile
http://us.click.yahoo.com/QBCcSD/o1CEAA/Zr0HAA/CefplB/TM
---------------------------------------------------------------------~->
To unsubscribe from this group, send an email to:
http-compliance-unsubscribe@egroups.com
Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
Can someone remind me what the state of the art is re: chunked POST? From what I can see, it's allowed, but there's not a good interoperability story with 1.0,...
Mark- I can't speak for the official position. The primary interoperability problem is due to CGI 1.1 (and earlier). CGI REQUIRES a CONTENT-LENGTH. Some...
Mark, FYI: Since supporting chunked encoded _messages_ is a MUST, our HTTP compliance test suite has test cases that use chunked _requests_. So far, none of...
Alex Rousskov
rousskov@...
Jun 11, 2002 4:17 pm
Thanks. So, from a client perspective, they might expect any of the following behaviours when sending a chunked POST; a) success b) 411 Length Required c) 5xx...
... e) a malformed/invalid message reaches the server (proxy corrupts message body and/or headers because it does not understand what is going on); the server...
Alex Rousskov
rousskov@...
Jun 11, 2002 5:07 pm
[Alex Rousskov: Tue, Jun 11, 2002 at 11:07:10AM -0600] ... f) 100 continue -Patrick...
Patrick McManus
mcmanus@...
Jun 11, 2002 5:14 pm
... g) nothing (i.e., the recipient [erroneously] waits for more data, without sending 100 continue) Alex....
Alex Rousskov
rousskov@...
Jun 11, 2002 5:24 pm
h) 413 Request Entity Too Large. -Carl Alex Rousskov <rousskov@...> 06/11/2002 11:24 AM Please respond to http-compliance To:...