summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* populate slave's msg_id_ack with sync'ed messages pending ackbug25394Matthias Radestock2013-01-192-24/+30
|
* eager sync of messages pending ackMatthias Radestock2013-01-195-24/+23
|
* merge bug25409 into bug25394Matthias Radestock2013-01-199-85/+131
|\
| * merge bug25395 into bug25394Matthias Radestock2013-01-190-0/+0
| |\
| * \ merge bug25397 into bug25395bug25395Matthias Radestock2013-01-190-0/+0
| |\ \
| | * \ merge default into bug25397bug25397Matthias Radestock2013-01-191-64/+57
| | |\ \ | | | |/ | | |/|
| | | * merge default into bug25397Matthias Radestock2013-01-171-64/+57
| | | |\
| * | | \ merge default into bug25395Matthias Radestock2013-01-194-77/+103
| |\ \ \ \ | | |/ / / | |/| | |
| | * | | pass 'unacked' flag to BQ:fold funMatthias Radestock2013-01-154-18/+21
| | | | | | | | | | | | | | | | | | | | so it can distinguish between 'ready' messages and those pending ack
| | * | | merge bug25397 into bug25395Matthias Radestock2013-01-152-15/+45
| | |\ \ \ | | | |_|/ | | |/| | | | | | | | | | | | which involved tweaking the variable_queue_with_holes result slightly, and ripping out a whole bunch of now-redundant test code.
| | | * | merge default into bug25395Matthias Radestock2013-01-151-23/+68
| | | |\ \
| | | * \ \ merge default into bug25395Matthias Radestock2013-01-141-56/+80
| | | |\ \ \
| | | | * | | refactorMatthias Radestock2013-01-131-19/+14
| | | | | | |
| | | | * | | include pending_acks in 'fold'Matthias Radestock2013-01-131-13/+46
| | | | | | | | | | | | | | | | | | | | | | | | | | | | we implement this as a zipper over three iterators
| | | | * | | merge bug25397 into bug25395Matthias Radestock2013-01-131-2/+2
| | | | |\ \ \
| | | | * \ \ \ merge bug25397 into bug25395Matthias Radestock2013-01-121-30/+30
| | | | |\ \ \ \
| | | | * \ \ \ \ merge default into bug25395Matthias Radestock2013-01-121-21/+24
| | | | |\ \ \ \ \
| | | | * \ \ \ \ \ merge bug25397 into bug25395Matthias Radestock2013-01-121-56/+52
| | | | |\ \ \ \ \ \
| | * | | \ \ \ \ \ \ merge bug25400 into bug25397Matthias Radestock2013-01-151-64/+57
| | |\ \ \ \ \ \ \ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | which is rather conflict-ridden; essentially requiring the rewriting of the crux of the change in bug25400 in the context of new iterator.
| | | * \ \ \ \ \ \ \ \ merge degault into bug25397Matthias Radestock2013-01-151-23/+68
| | | |\ \ \ \ \ \ \ \ \ | | | | | |_|_|_|_|_|/ / | | | | |/| | | | | | |
| | | * | | | | | | | | merge default into bug25397Matthias Radestock2013-01-141-56/+52
| | | |\ \ \ \ \ \ \ \ \ | | | | |_|/ / / / / / / | | | |/| | | | | | / / | | | | | |_|_|_|_|/ / | | | | |/| | | | | |
| | | | * | | | | | | cosmeticMatthias Radestock2013-01-131-2/+2
| | | | | |_|_|_|/ / | | | | |/| | | | |
| | | | * | | | | | only pass the minimum necessary state to 'next'Matthias Radestock2013-01-121-23/+26
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | i.e. the IndexState. The remainder of the State is encapsulated inside the iterator state. Technically we only need q{1-4} and delta, but it's simpler and more obvious to just pass a read-only State around.
| | | | * | | | | | return MsgStatus only from iteratorMatthias Radestock2013-01-121-10/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | leaving the message reading to the fold
| | | | * | | | | | merge default into bug25397Matthias Radestock2013-01-121-56/+52
| | | | |\ \ \ \ \ \ | | | | | | |_|/ / / | | | | | |/| | / / | | | | | |_|_|/ / | | | | |/| | | |
| | | | | * | | | pass State to iteratorMatthias Radestock2013-01-121-51/+43
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | We want to be able to zip this iterator with other iterators that also manipulate the vqstate. Hence we must pass the State explicitly rather than keeping it opaque inside the iterator state. Also, some refactoring on read_msg.
| | | | | * | | | merge default into bug25397Matthias Radestock2013-01-121-48/+52
| | | | | |\ \ \ \ | | | | | | |_|/ / | | | | | |/| | |
| | | | | | * | | implement vq:fold in terms of an iteratorMatthias Radestock2013-01-111-48/+52
| | | | | | | | |
| * | | | | | | | add xmerl to plt so we get fewer 'Unknown functions' in dialyzerMatthias Radestock2013-01-191-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | mochijson2 depends on it
| * | | | | | | | eliminate "Function X has no local return" dialyzer errorsMatthias Radestock2013-01-194-1/+23
| | | | | | | | |
| * | | | | | | | cosmeticMatthias Radestock2013-01-191-6/+4
| | | | | | | | |
* | | | | | | | | add a testbug25409Matthias Radestock2013-01-191-0/+16
| | | | | | | | |
* | | | | | | | | add BQ:purge_acks/1Matthias Radestock2013-01-193-3/+11
|/ / / / / / / /
* | | | | | | | tiny refactor on variable_queue_with_holesMatthias Radestock2013-01-191-3/+4
| | | | | | | |
* | | | | | | | cosmeticMatthias Radestock2013-01-191-2/+2
| | | | | | | |
* | | | | | | | merge bug25383 into defaultMatthias Radestock2013-01-180-0/+0
|\ \ \ \ \ \ \ \ | | |_|_|_|_|_|/ | |/| | | | | |
* | | | | | | | fix testMatthias Radestock2013-01-181-1/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | connections only show up in 'list_connections' after the protocol header has been sent
* | | | | | | | various reader related changes for AMQP 1.0Matthias Radestock2013-01-183-31/+74
|/ / / / / / / | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | - mechanism for the reader to 'become' a different reader. - become the 1.0 reader if an AMQP 1.0 header is presented by a client and the reader is present. That way we can support 1.0 on the same port as 0-{8,9,9-1}. - defer starting of the channel_sup_sup and do that in the reader. This allows the AMQP 1.0 reader to start its own versio of the sup. It also makes aborted connections less costly. - track connections in an ets table rather than implicitly via the supervisor. That way AMQP 1.0 connections can exclude themselves, since they are already tracked via their direct connections.
* | | | | | | simplifying refactor on rabbit_mnesia:discover_clusterMatthias Radestock2013-01-171-9/+6
| | | | | | |
* | | | | | | merge bug25400 into defaultTim Watson2013-01-172-29/+24
|\ \ \ \ \ \ \ | |/ / / / / /
| * | | | | | filter out pending acks when folding over deltabug25400Matthias Radestock2013-01-151-6/+14
| | | | | | |
| * | | | | | testMatthias Radestock2013-01-151-23/+10
| | |_|_|_|/ | |/| | | |
* | | | | | stable to defaultSimon MacMullen2013-01-161-4/+9
|\ \ \ \ \ \
| * \ \ \ \ \ Merge bug25401Simon MacMullen2013-01-160-0/+0
| |\ \ \ \ \ \
| * | | | | | | Prevent explosion if someone passes a list.bug25401Simon MacMullen2013-01-161-4/+9
| |/ / / / / /
* | | | | | | cosmeticMatthias Radestock2013-01-161-4/+2
| | | | | | |
* | | | | | | remove superfluous conditionMatthias Radestock2013-01-161-2/+2
| | | | | | |
* | | | | | | Merge bug25393Simon MacMullen2013-01-1612-141/+283
|\ \ \ \ \ \ \
| * | | | | | | some more reader connection state abstractionMatthias Radestock2013-01-151-14/+12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | and a slightly more logical (and efficient) handle_frame clause order
| * | | | | | | optimising refactor of check_xrefMatthias Radestock2013-01-151-1/+2
| | | | | | | |
View Lorry Controller Status