Closed GoogleCodeExporter closed 9 years ago
Operation 12 is not known to me.. what did you do exactly to make this appear?
Original comment by evert...@gmail.com
on 28 Apr 2008 at 7:04
Just checked with the PyAMF guys.. its new to them too..
These were the operations times known to me (from documentation)
http://code.google.com/p/sabreamf/source/browse/trunk/SabreAMF/AMF3/CommandMessa
ge.php
SabreAMF supports the PING and LOGIN command.. So i'm pretty curious what this
is from.
Original comment by evert...@gmail.com
on 28 Apr 2008 at 7:18
I don't know if that charles dump is the right thing, but what i did was a
consequent
request of a remote method, without waiting for the result and that's the error
i've got.
I will try to repeat it asap and give you feedback.
Original comment by stoica.i...@gmail.com
on 29 Apr 2008 at 3:45
I've attached all the charles dumps needed to hopefully understand what is the
issue,
they appeared in Charles in the order of the name of the files.
dump-1.chls
dump-2.chls
I hope they will help you.
Original comment by stoica.i...@gmail.com
on 1 May 2008 at 1:26
Attachments:
looks like Flex3 adds two new CommandMessage operations:
DISCONNECT_OPERATION : This operation is used to indicate that a channel has
disconnected.
MULTI_SUBSCRIBE_OPERATION : Used by the MultiTopicConsumer to
subscribe/unsubscribe
for more than one topic in the same message.
(see
http://livedocs.adobe.com/flex/3/langref/mx/messaging/messages/CommandMessage.ht
ml)
According to the ASDoc for DISCONNECT_OPERATION:
"Disconnects from the remote destination. Because this channel uses a
stateless HTTP
connection, it sends a fire-and-forget message to the server as it disconnects
to
allow the server to shut down any session or other resources that it may be
managing
on behalf of this channel."
I guess then a response is not required. Right now it's responding with a
fault, so
that can be avoided (and probably respond with nothing). The PyAMF guys are
responding with what looks like a null result (http://pyamf.org/ticket/325).
The following patch takes a similar approach to the PyAMF implementation. I
wondering if it would be more efficient to drop the connection and respond with
nothing.
Original comment by michael....@henryschein.com
on 8 Jul 2008 at 1:55
Attachments:
Thanks a lot, this fix has been committed to SVN. A release will follow within
a week..
Original comment by evert...@gmail.com
on 16 Sep 2008 at 10:51
Original comment by evert...@gmail.com
on 16 Sep 2008 at 10:52
Original issue reported on code.google.com by
stoica.i...@gmail.com
on 28 Apr 2008 at 6:48Attachments: