[Vortex] weird behavior or misunderstanding of beep/vortex?

[Francis Brosnan Blazquez]

El vie, 09-06-2006 a las 16:19 -0300, Gustavo Sverzut Barbieri escribió:
Hi Gustavo,

Attached to the bugzilla [1] is found a patch that solves the problem
reported and also solves the segfault found at the vortex sequencer
loop. 

The first problem was cause by a checking at the vortex channel module
which wasn't replying with a SEQ frame, extending the window size, and
allowing to receive more data because the channel was flagged as "being
closed". The check has been updated to be more accurate, and only to
only perform the check for the channel 0. 

The problem was triggered by the fact that the client was requiring to
close the channel (last vortex_connection_close instruction inside
vortex-simple-client.c), but the server still had frames to deliver.
Because the client wasn't producing SEQ frames, making the server to not
send more frames, the client was stalled, waiting for the message to be
received, until the timeout is reached.

For the segfault found at the vortex sequencer module. It was caused
because the vortex sequencer was placing twice the same reference into
the "pending messages" queue for the channel. Once the vortex sequencer
do a check to avoid queueing messages that are already queue and being
resequenced, everything works fine.

Let us to know if this solves your problem,

Cheers!

[1] http://dolphin.aspl.es/cgi-bin/bugzilla/show_bug.cgi?id=321

-- 
Francis Brosnan Blazquez <francis at aspl.es>
Advanced Software Production Line, S.L.