Changes between Version 1 and Version 2 of Ticket #17187, comment 6


Ignore:
Timestamp:
Nov 11, 2011, 4:46:27 PM (13 years ago)
Author:
Leo Shklovskii

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #17187, comment 6

    v1 v2  
    1 A similar issue happens when running against uWSGI. Django attempts to do a read() even if the content-length is explicitly set to zero. The result is that the read() call blocks.
     1A similar issue happens when running against uWSGI. Django attempts to do a read() even if the content-length is explicitly set to zero. The result is that the read() call blocks. The repro in my case is really simple. Get the latest uWSGI, spin up simple django app, send a POST with nothing in it.
    22
    3 Djangop should properly distinguish between unknown content-length (when trying to read might make sense) and one explicitly set to zero.
     3There's some good explanation of this issue in Graham Dumpleton's blog: http://blog.dscpl.com.au/2009/10/details-on-wsgi-10-amendmentsclarificat.html
    44
    5 The repro in my case is really simple. Get the latest uWSGI, spin up simple django app, send a POST with nothing in it.
     5Django should properly distinguish between unknown content-length (when trying to read might make sense) and one explicitly set to zero, when trying to read is a mistake. As per the WSGI spec, it would be wrong for uWSGI to return anything in response to a read() call, even an EOF when it set the content-length to 0.
    66
     7
     8
Back to Top