- From: Laurent Carcone <carcone@w3.org>
- Date: Tue, 10 Mar 2009 12:54:09 +0100
- To: Hugh Sasse <hgs@dmu.ac.uk>
- CC: Amaya Mailing List <www-amaya@w3.org>
Hugh Sasse wrote: > On Wed, 4 Mar 2009, Laurent Carcone wrote: > > >> Hello Hugh, >> >> Amaya is based on the libwww for HTTP requests and there is indeed a maximum >> of redirection allowed in the library (I guess to avoid infinite loop). By >> default this maximum is set to 5 and Amaya set it to 7 at the initialization >> (so more than the spec suggests). >> I haven't see any difference between HTTP 1.0 and HTTP 1.1 concerning this >> aspect in the library. >> > > The part of the spec that I quoted suggests to me that there should be. > Who maintains the library? I seem to recall that the linemode browser > which was part of the library is no longer maintained. > Indeed, the libwww is no longer maintained > >> So , there is no way in the current version to dynamically change this >> maximum. >> > > What about future versions of Amaya? == Are you giving this the status > of WONTFIX? > I guess we won't change the maximum of redirects defined in Amaya initialization. Can you give us a uri so that we can test the number of redirections needed in this case ? > >> I guess it's the first report on this problem, the mail you pointed out >> concerns a redirection problem for the PUT method and I think it has been >> fixed. >> > > Would redirect code be so different for put and post that it would not be > shared? What causes you to think it has been fixed -- can you look it up, > or is this from memory? > The problem was in Amaya code (when Amaya treats the result of the request), not in the libwww, and it has been fixed. Thanks, Laurent Carcone >> Regards, >> Laurent Carcone >> >> > Thank you, > Hugh > >> Hugh Sasse wrote: >> >>> I have been attempting to debug a problem (with colours not showing >>> up) on an internal server. I've tried a number of browsers, >>> (Firefox, internet explorer and Opera) but when I remembered Amaya >>> is more rigorous, I tried that. It detected bugs in the HTML, but >>> also it gave me a message about too many redirects. >>> >>> It turns out that HTTP 1.1 mentions this: >>> >>> Q> 10.3 Redirection 3xx >>> Q> [...] >>> Q> Note: previous versions of this specification recommended a >>> Q> maximum of five redirections. Content developers should be aware >>> Q> that there might be clients that implement such a fixed >>> Q> limitation. >>> >>> so therefore it no longer applies in HTTP 1.1, only 1.0. I can't find >>> any means of turning this error detection off. I agree that it is a >>> useful warning for 1.0 compliance, but if the client has requested 1.1 >>> then it should not matter, should it? >>> >>> Searching the archives for previous reports of this problem I found this: >>> >>> http://lists.w3.org/Archives/Public/www-amaya/2006JulSep/0019.html >>> >>> which is about 2 and half years old, but doesn't seem to have been >>> implemented. I had six redirect questions to respond to before this >>> other error message came up, and that was for one submit button. >>> >>> Thank you, >>> Hugh >>> >>> >>> > > > >
Received on Tuesday, 10 March 2009 11:54:46 UTC