This is what SCGI reports ( apache 2.2.4, mod_scgi 1.12, with the
default main() invocation of scgi/scgi_server.py ).
In this implementation at least, REQUEST_URI does not include the
proctal or hostname.
Essentially:
REQUEST_URI == SCRIPT_NAME + PATH_INFO + QUERY_STRING
HTTP_REFERER: 'http://
This borks query-string handling. ['REQUEST_URI'] includes the query
string.
instead:
self.path = environ['SCRIPT_NAME'] + environ['PATH_INFO']
On Feb 13, 5:41 pm, "mulc" <[EMAIL PROTECTED]> wrote:
> OK.
>
> I hacked django/core/handlers/wsgi.py. ( line #56 )
>
> In the __init__ function of
On Tue, 13 Feb 2007 17:41:16 -0500, mulc <[EMAIL PROTECTED]> wrote:
>
>
> OK.
>
> I hacked django/core/handlers/wsgi.py. ( line #56 )
>
> In the __init__ function of WSGIRequest I changed:
>
> self.path = environ['PATH_INFO']
>
> to
>
> self.path = environ['REQUEST_URI']
>
> I then chang
OK.
I hacked django/core/handlers/wsgi.py. ( line #56 )
In the __init__ function of WSGIRequest I changed:
self.path = environ['PATH_INFO']
to
self.path = environ['REQUEST_URI']
I then changed up my urls.py
(r'^admin/', include('django.contrib.admin.urls')),
becomes
Is there any mechanism for making Django 'aware' of it's SCGI
mountpoint?
I have been experimenting with SCGI/WSGI bridges ( both flup, and
Titus Brown's SWAP from Paste ) and I am having difficulty mounting a
django-app under an existing website a particular locat
5 matches
Mail list logo