python - appengine, urlfetch, and the content-length header -

I have the Google Appengine app to request pages from another server using urlib2 POSTs. I recently enabled gzip compression on other servers running the Apache 2, and Apenpage page requests began to fail on key-error, indicating that 'content-length' is not in the header.

I'm not explicitly telling gzip Apengine accepted requests in my requests, but it is possible that Appengine is adding that header Googling has not given any clear indication that Appengine's urlfetch has gzip A header has been added to accept the encoding.

Apache 2, if I miss the correct way, then leaves the header with length of content when the response is compressed, but non-compressed responses to the same server should not be affected.

Does anyone have any insights as to what is happening, why the content length header is being omitted? According to this thread :) On an Appenzine Java newsgroup, Google usually accepts urlfetch requests on 'Acceptance', and '' Adoption '' on the urlfetch requests.

So obviously, Appenpine adds an acceptable-encoding: On the incoming requests on the Internet lies the gossip header, and uncompresses the response, but no content-length in the header for the decoded does not include The size of the data so if the external server will provide the gypsy responses, the net result for the epiphany script (all the pre and post by the app engraved above After the processing behavior) there is loss of content-length headers.