cherokee could accept empty POST

[GoogleCodeExporter]

What steps will reproduce the problem?
1. POST with no key/value pairs

What is the expected output? What do you see instead?
expected output: whatever the target script does
see instead: HTTP Error 411: Content-Length required (but Content-Length
looks set)

I'm not sure if this is completely legal according to the spec, but apache
accepts it as such

Original issue reported on code.google.com by johnnyop...@gmail.com on 3 Dec 2008 at 6:22

[GoogleCodeExporter]

Original comment by alobbs on 3 Dec 2008 at 7:11

• Changed state: Accepted
• Added labels: OpSys-All, Priority-Medium, Type-Defect

[GoogleCodeExporter]

I have just realized that there is a QA test covering that case (it successes):

  http://svn.cherokee-project.com/browser/cherokee/trunk/qa/030-Post%2Burlencoded.py

Johnny, how did you get it to fail? I cannot reproduce the issue.

Original comment by alobbs on 5 Dec 2008 at 9:46

[GoogleCodeExporter]

I have just added another QA test to check whether empty POSTs works with PHP 
(FastCGI). They do:

  http://svn.cherokee-project.com/browser/cherokee/trunk/qa/194-PHP-zero-len-post.py

Johnny, I'm gonna need a little help to reproduce this one. Everything works at 
my end. ':-)

Original comment by alobbs on 5 Dec 2008 at 11:08

[GoogleCodeExporter]

perhaps the issue only happens through a redirection

$(document).ready(function(){
    window.setInterval(
        function(){
            $.post('$ajaxurl',{ foo: "bar"});
        },
        1000 * 60 * 5 // 5 minutes
    );
});

where ajax url is http://localmomentum.net/admin/keep_login

which is handled by redirection on Default ruleset with internal redirection 
handler 

^.*$ -> /index.php

Original comment by johnnyop...@gmail.com on 5 Dec 2008 at 7:55

[GoogleCodeExporter]

Tested with redirections and works at my end too... :-?

Original comment by skar...@gmail.com on 14 Jan 2009 at 7:03

Attachments:

• 196-PHP-zero-len-post+Redir.py

[GoogleCodeExporter]

Skarcha, I'd rather use this QA test file (find it attached).
It seems that the issue is that POST information is lost when an internal 
redirection is performed.

The test successes. I wonder whether this bug is still a real issue.
Johnny, could you please confirm it?

Original comment by alobbs on 15 Jan 2009 at 7:53

Attachments:

• 196-PHP-zero-len-post+Redir.py

[GoogleCodeExporter]

Johnny, we need some feedback on this issue.. 

Original comment by alobbs on 2 Mar 2009 at 4:12

[GoogleCodeExporter]

maybe this one had to do with the ignored Location Header Bug I reported last 
month
where some additional headers were missing...?

Original comment by michael....@gmail.com on 24 Jun 2009 at 7:32

[GoogleCodeExporter]

I'm closing this bug: The QA test is in place, and the behavior is alright now 
(and besides, the bug has been 
inactive for more than 6 months).

Original comment by alobbs on 2 Nov 2009 at 6:50

• Changed state: Fixed