HTTP Methods and Redirect Status Codes

re: HTTP Methods and Redirect Status Codes

Bikram — Tue, 11 Jun 2013 10:31:24 GMT

By Fiddler I have verified. it is getting the 307 status code. I have both the request i.e. original POST request and the Redirected POST request, both are having the same header except the request method. Because of this I am losing the POST parameters. the same thing is working on Firefox and Chrome.

EricLaw: Please email me a repro URL. Use Help > About in Fiddler.

re: HTTP Methods and Redirect Status Codes

Bikram — Fri, 31 May 2013 18:20:48 GMT

My application initiates a POST request from JavaScript and the server sends a redirect new url and code 307. IE even though receieves 307, it initiates a GET request. Why?? Where I could have gone wrong?

EricLaw: Please provide a repro URL? Did you verify using Fiddler that it's really returning a 307?

re: HTTP Methods and Redirect Status Codes

Merc — Sat, 12 Jan 2013 11:45:50 GMT

Hi, I am using a 307 redirect in a htaccess file, like this:

RewriteRule ^add/(.*)/$ /_ucp/moduls/$1/add/add.php [L,R=307]

your browser now sends a post request, recives a 307 and then starts a GET request with the new URL. I tracked it in the F12 developer tools, any tipps, explanations ?

EricLaw: This seems more likely to either be a DevTools bug (a POST is really sent) or a bug in your server (you're not really sending a 307). Please email me (using Help > Send Feedback in Fiddler) a SAZ file captured using Fiddler that demonstrates this problem.

re: HTTP Methods and Redirect Status Codes

Julian Reschke — Fri, 02 Sep 2011 08:28:30 GMT

HTTPbis draft 17 will *allow* rewriting of POST (and only POST) to GET for the status codes 301 and 302. See trac.tools.ietf.org/.../1428

re: HTTP Methods and Redirect Status Codes

Julian Reschke — Sat, 20 Aug 2011 14:08:22 GMT

HTTPbis WG issue: trac.tools.ietf.org/.../160

The current plan is to keep the definitions of 301/302, except by noting that rewriting POST-to-GET is ok (ut not for other methods).

re: HTTP Methods and Redirect Status Codes

EricLaw [ex-MSFT] — Sat, 20 Aug 2011 00:00:44 GMT

@Aseem: I'm interested to see what HTTPBIS specifies here. The current behavior of preserving the method is spec-compliant and it has been this way for over a decade.

While matching behavior is appealing, the notion that a 301 or 302 response should turn a HEAD into a GET doesn't make much sense to me.

re: HTTP Methods and Redirect Status Codes

Aseem — Fri, 19 Aug 2011 23:55:53 GMT

Thanks for the clarification. Just out of curiosity, was this a candidate for changing in IE9 or IE10 simply for compatibility w/ all other modern browsers? (E.g. maybe just in IE9/IE10 standards modes?)

re: HTTP Methods and Redirect Status Codes

Pebe — Fri, 19 Aug 2011 19:57:04 GMT

Mozilla bug tracker entry for this: bugzilla.mozilla.org/show_bug.cgi

re: HTTP Methods and Redirect Status Codes

Michael Edgar — Fri, 19 Aug 2011 19:51:18 GMT

Hi Eric, you've pulled out the "Note" on the definition of 302 to explain IE9's behavior, but you left out the immediately preceding sentence:

"If the 302 status code is received in response to a request other than GET or HEAD, the user agent MUST NOT automatically redirect the request unless it can be confirmed by the user, since this might change the conditions under which the request was issued."

While the developer was wrong in relying on DELETE's conversion to GET on a 302, IE9 is still wrong in not requiring confirmation. Cheers, Mike

[EricLaw: Hi, Mike! You should probably go ahead and read the entire post. There's an entire section on that topic. Thanks! ]