W3C home > Mailing lists > Public > ietf-http-wg@w3.org > October to December 2011

Re: issue 325: When are Location's semantics triggered?, was: Protocols/APIs and redirects

From: Cameron Heavon-Jones <cmhjones@gmail.com>
Date: Thu, 15 Dec 2011 15:23:48 +0000
Cc: Julian Reschke <julian.reschke@gmx.de>, Mark Nottingham <mnot@mnot.net>, Willy Tarreau <w@1wt.eu>, "Roy T. Fielding" <fielding@gbiv.com>, Amos Jeffries <squid3@treenet.co.nz>, ietf-http-wg@w3.org
Message-Id: <2EF0A9BA-E3E6-4183-B381-4B70FDB8A783@gmail.com>
To: Daniel Stenberg <daniel@haxx.se>
On 14/12/2011, at 10:11 PM, Daniel Stenberg wrote:

> On Wed, 14 Dec 2011, Cameron Heavon-Jones wrote:
> 
>> I'm more that willing to accept if there is some fallibility in the tests and its great that there is validation. I'm seriously unsure where such error may be since i've been tracing http request\responses through in-browser network tools. i'm going to repeat capturing whats on the wire instead and if i still can't find the same results i'll setup a public test case which everyone can use.
> 
> Pick your browser of choice and click on this link for a GET test:
> 
> 	http://daniel.haxx.se/test/redir.cgi
> 
> And you can POST to the same URL by using this form:
> 
> 	http://daniel.haxx.se/test/redir.html
> 
> (It returns a 302 redirect with 282 bytes of contetns. It redirects to http://example.com/funny/page.html)
> 
> -- 
> 
> / daniel.haxx.se


I have found the source of the errors in the test results which was due to a bug in the test harness affecting >10% cases which were testing the unique combination of redirection codes, location headers AND content. the bug arose through recent expansion of the test coverage which included testing some non-standard responses. the erroneous results can be regarded as non-conformance tests over redirection codes without a location header.

i've since fixed the bug and confirmed the same results that were expected by expertise field knowledge. its a good catch and i'm really glad that it was found as incorrect results are far worse than no results. the tests were setup precisely to determine field behaviour where such preexisting knowledge did not exist. 

the unexpected nature of the results was the very reason that i thought pertinent to bring to the discussion, since the results are incorrect, both them and my resulting observations and analysis should be disregarded.

i've repeated the affected tests and updated the matrix with the correct results. 

Thanks,
Cameron Jones
Received on Thursday, 15 December 2011 19:17:09 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:51:51 GMT