On Tuesday, July 17, 2001, at 02:10 PM, Ian Holsman wrote:
Chuck Murcko wrote:
I'll have a look at the dechunk stuff, and see if a more generic fix can be done.
On Tuesday, July 17, 2001, at 12:54 PM, Chuck Murcko wrote:
Well, if the dechunk filter is not removing intermediate eos inserted by the http_in or core_in filters, as it should, then this allows the whole set of chunks through, and we should close the connection when finished, and we make the extra call into the dechunk filter to reset its state machine.
For sure we have to make one last ap_get_brigade() call to reset the state on the ap_dechunk_filter(). Then it will read and return the entity trailers and append the (proper) eos bucket, and reset its state to awaiting headers, instead of awaiting trailers.
The real fix might lie in losing the awaiting trailers state, and just do all that right away when the last chunk comes into the dechunk filter, and return to awaiting headers state directly. Is there some good reason not to do this?
Chuck Murcko Topsail Group http://www.topsail.org/
