Gordon Sim wrote:
(*) bug in c++ broker is captured in JIRA: https://issues.apache.org/jira/browse/QPID-548
I've checked in a fix for that issue and the corresponding ruby test now passes.
However there is then a failure on a later test (see trace below), caused I think by the broker closing the socket without waiting for a connection.close-ok. Thats not a critical one and I don't myself plan to fix it for M2.
test_channel_close(Channel): CONNECTION CLOSED: 504, Unknown channel: 1, 20, 40
writer Qpid::Closed .worker Qpid::Closed Codec::EOF ./qpid/codec.rb:239:in `read' ./qpid/codec.rb:247:in `unpack' ./qpid/codec.rb:186:in `octet' ./qpid/connection.rb:59:in `read' ./qpid/peer.rb:90:in `reader' ./qpid/peer.rb:74:in `spawn' ./qpid/peer.rb:72:in `spawn' ./qpid/peer.rb:55:in `start' ./qpid/client.rb:83:in `start' ./qpid/test.rb:32:in `connect' ./tests/channel.rb:38:in `test_channel_close'
