Author xdegaye
Recipients Pierrick.Koch, andy_js, giampaolo.rodola, xdegaye
Date 2013-06-13.10:39:36
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <1371119977.52.0.416536457021.issue17925@psf.upfronthosting.co.za>
In-reply-to
Content
Attached are two test cases for this patch.

test_simple_producer still fails with the new patch because it should be:

  self.producer_fifo.extendleft([first, data])

instead of:

  self.producer_fifo.appendleft([data, first])

The order of the items in the list is reversed, as documented in
deque.extendleft. So the attachment also includes the corrected patch
with this single change.

I still think  that if num_sent == 0, then 'first' should be put back
in the queue. This means that initiate_send should not attempt anymore
to send an empty string, which is confusing anyway, and therefore at
the beginning of initiate_send, when 'if not first', then we should
return in all cases and not only when 'first' is None.
History
Date User Action Args
2013-06-13 10:39:37xdegayesetrecipients: + xdegaye, giampaolo.rodola, Pierrick.Koch, andy_js
2013-06-13 10:39:37xdegayesetmessageid: <1371119977.52.0.416536457021.issue17925@psf.upfronthosting.co.za>
2013-06-13 10:39:37xdegayelinkissue17925 messages
2013-06-13 10:39:37xdegayecreate