W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > January to April 1997

Re: mid-course errors

From: Jeffrey Mogul <mogul@pa.dec.com>
Date: Fri, 18 Apr 97 15:25:01 MDT
Message-Id: <9704182225.AA24482@acetes.pa.dec.com>
To: http-wg@cuckoo.hpl.hp.com
"Scott Lawrence" <lawrence@agranat.com> writes:

  Our present implementation just closes the connection without
  sending the terminating (zero-length) chunk.  Browsers all seem to
  display some sort of error when you cut off a response in the middle
  if you've passed Content-Length in the header; we assume that they
  would do the same for this case.

This makes sense as long as the content that *is* received is
non-corrupt.  I guess the worrisome case is if the failing
content-generation mechanism generates bogus content, not just
truncated content.

The browsers I've used, when they see a premature close for a
non-chunked response, display the data received so far, and a "Transfer
Interrupted!" warning.  This doesn't necessarily warn the user "and by
the way, what you already have on your screen might be bogus".

But this is probably just "advice to client implementors", not
something we need to put into the spec.  (And I've seldom seen
any content-generator realize when it has generated bogus data,
although I have certainly seen enough cases where the output was
clearly broken.)

-Jeff
Received on Friday, 18 April 1997 15:33:45 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:32:35 EDT