94ade5d9dd
stable to default
2013-01-24 13:15:52 +00:00
4e730491c7
Update copyright 2013
2013-01-23 11:40:49 +00:00
18877f91c8
merge stable into default
2013-01-08 11:51:06 +00:00
23f036d082
correctly enable test
2013-01-08 11:50:37 +00:00
7fcb2ea1be
actually enable the test
2013-01-08 11:39:30 +00:00
ff017197ca
fail properly
2013-01-08 11:36:57 +00:00
e235a8d42a
fix bug in test
...
it's a throw, not an exit
2013-01-07 18:08:40 +00:00
eeaf125c38
Less severe failure when waiting on unselected channels
2013-01-07 17:03:35 +00:00
0095e042a3
Use less state
2012-12-20 16:16:32 +00:00
3d23af7af5
Prevent waiting for confirms unless selected
2012-12-20 15:00:24 +00:00
455e8c0f9a
It seems more obvious to treat the no password variant as the special case.
2012-09-20 17:00:55 +01:00
87994f521f
Update tests
2012-09-18 11:18:01 +01:00
33c01884cb
Merge bug25057
2012-08-15 13:52:24 +01:00
1c05823f35
add functions to unregister handlers
...
I have added tests for the return handler only since it is the easiest
to check. the confirm handler would be easy as well if it wasn't for
`setup_publish', which does not let you republish messages to the
exchange - I can add them if you think we need them. I would have to
set up some more code for the flow one as well.
2012-07-16 16:23:44 +01:00
7749390ecf
move internal definitions etc out of amqp_client.hrl
...
...in order to avoid application name space polution
2012-07-02 16:54:31 +01:00
fb57809497
And again, globally raise the timeouts.
2012-05-16 12:23:14 +01:00
69665f3ff6
Merged bug24624 into default
2012-02-06 16:12:43 +00:00
692884fc78
Merge bug24500
2012-02-03 15:36:20 +00:00
74ffe957cd
Merge bug24606
2012-02-03 13:05:25 +00:00
c8f222c60e
Merge bug24692.
2012-02-03 11:48:13 +00:00
924ccc903d
Move rabbit_misc:nodeparts/1 and rabbit_misc:makenode/1 into rabbit_nodes
2012-02-02 14:36:16 +00:00
44278dc2b8
Update copyright 2012
2012-02-02 13:04:28 +00:00
e0a21b2fa3
Unbreak the network client. Oops. Tests pass now :)
2012-01-25 16:00:19 +00:00
85e8ec4951
deconvolute logic
...
The first problem pointed out in the previous commit no longer applies.
The only change in semantics is that wait_for_confirms_or_die will wait for
all acks and nacks before exit(nacks_received). It shouldn't matter much.
2012-01-24 17:50:16 +00:00
976b651560
add timeout to wait_for_confirms[_or_die]
...
This also fixes two bugs:
1) Handling nacks never worked. It would do a gen_server:call inside a
gen_server:call and would just lockup. We circumvent the problem now, by
calling handle_close/4 instead of close/3;
2) In channels_manager, after the first erronous frame, the state would be
replaced by 'ok'. That's fixed now.
2012-01-24 17:03:05 +00:00
88d184be09
Relax default timeout
2012-01-20 13:12:20 +00:00
93c86f1958
make tests work on R15
2012-01-16 07:44:59 +00:00
2763345bd0
Make wait_for_confirms not hang under the influnce of multiple=true.
2011-10-18 17:48:03 +01:00
3e2419ab22
make tests a bit more comprehensive
2011-10-17 11:53:54 +01:00
9d7c6fbbd3
add more checks to the tests
...
There was also a bug: "amqp://host" would parse to vhost "", instead of vhost
"/".
2011-10-17 11:40:23 +01:00
ee25d6a734
support mixed case schemes + more docs
2011-10-17 11:14:58 +01:00
909e86d358
support R12B-5 and better docs
2011-10-10 15:04:00 +01:00
baf6480c62
rename url -> uri; do params_network conversion in amqp_uri
2011-09-27 17:35:40 +01:00
cc1c548251
refactor + handle entities + url tests
2011-09-26 16:52:57 +01:00
1d78937835
use confirms to ensure messages ended up in queues...
...
so we don't rely on 'purge' not overaking 'publish'.
2011-08-18 18:27:43 +01:00
da87a41245
use 'purge' instead of passive 'declare' to determine msg count...
...
since as of bug 24194 a 'declare' may overtake a 'publish'
2011-08-18 18:13:10 +01:00
61ec3a9e53
Rework return from new_connection so that we detect sooner if there's been a problem creating the connection
2011-08-09 12:42:04 +01:00
1de9c86bb0
merge default into bug24018
2011-08-09 12:11:18 +01:00
fc7355471e
Merging bug23838 to default
2011-08-08 14:23:57 +01:00
58f30b55b7
ipv6 + ssl test
2011-07-29 18:26:52 +01:00
f4b5f3a2b6
add ipv6 test
2011-07-29 18:03:45 +01:00
26c55bb4c2
update copyright year
2011-07-28 14:30:09 +01:00
af5f1c85e4
direct tests run again
...
(and covertly strip trailing whitespace)
2011-07-28 14:01:36 +01:00
6ca8225d0d
merge from default
...
Direct tests still fail. There might also be a test or two missing.
2011-07-28 13:46:07 +01:00
680957725b
merge default into bug23117
2011-07-18 14:52:56 +01:00
e583b7978e
Merge bug24229 to default
2011-07-14 16:36:36 +01:00
23926b369a
add wait_for_confirms_or_die/1
2011-07-11 09:57:04 +01:00
19af12bb6a
remove useless check
...
I'm not sure we actually have a way to check that the two events,
returning from wait_for_confirms and the all published messages being
acknowledged, happen in the right order. It's especially problematic
since we don't have shared mutable variables like in Java.
2011-07-11 09:04:37 +01:00
b53d0a9eb1
add tests
2011-07-08 17:58:15 +01:00
d6fe655d23
don't inspect gs's From tuple; re-add channel:subscribe
...
I don't like the idea of having subscribe in the consumer modules.
Subscription is something all the consumers need to handle, so it
seems odd to have to do:
amqp_selective_consumer:subscribe(...
or even
(amqp_channel:get_consumer_module()):subscribe(...
It seems far more natural to do:
amqp_channel:subscribe(...
The only thing the old subscribe function could do that the current
call('basic.consume'{}... ) can't is subscribe make a different pid as
the subscriber, so I re-added
amqp_channel:subscribe(Channel, BasicConsume, Subscriber)
Now, this is exactly the same as running
amqp_channel:call('basic.consume'{}...) from the Subscriber process.
Instead of splitting the From tuple, we're now just adding the Sender
to all amqp_channel calls. The only time when this is currently used
is sending a basic.consume the the consumer module.
This also allows us to treat subscribes and consumes the same way, as
opposed to the old code that had a lot of special handling for
consumes in selective_subscriber.
2011-07-06 11:20:24 +01:00
62e52f44ea
all consumer module functions may returns errors
2011-07-05 17:18:56 +01:00
f2d58c1951
remove call to test that hasn't existed since 2.1.0
2011-07-04 12:28:22 +01:00
0b8d25b4e6
auto-select ports for normal and ssl connections
...
Also, unbreak SSL tests.
2011-07-04 12:19:04 +01:00
5752c1f3ff
fix a couple of warning and remove a dangling comment
2011-06-30 13:17:14 +01:00
1f19dcad5d
re-enable selective's default consumer and add a test for it
...
We're also doing something with the monitors now: selective keeps
track of which consumers are still alive; when it receives a message,
it either routes it to the appropriate consumer if it was registered
and is still alive, or to the default consumer if it exists, or drops
it otherwise. Monitors are used to remove dead consumers.
There's a race between the consumer dying and messages for it: it's
possible for the consumer to die, and a few messages to get lost
before selective receives the DOWN and starts re-routing messages to
the default consumer.
2011-06-30 11:45:47 +01:00
28eb4c4190
tests pass
2011-06-29 16:32:32 +01:00
4e56303921
remove selective_consumer:cancel/2 and comments mentioning it or subscribe/2
2011-06-29 13:28:27 +01:00
af065e3315
merging in from default
2011-06-03 17:01:38 -05:00
5650f0fed0
Separate out different types for network and direct connections, don't check the password for direct ones.
2011-05-03 15:35:25 +01:00
caedd9f7b9
merging in from default
2011-04-07 10:28:12 -05:00
8639d5c88c
fixing tests in source tarball
2011-04-02 09:02:43 +01:00
f9af40d3d1
abstracting out unit tests
2011-04-01 22:14:31 +01:00
a6f8ce8295
merging in from bug23117
2011-04-01 19:20:00 +01:00
b98bc37dad
adding documentation for exit reasons in amqp_channel and amqp_connection
2011-04-01 18:45:39 +01:00
ea9ef4e131
fixing discrepancies between exit reason in the two types of connections
2011-04-01 17:56:39 +01:00
c740cf6057
fixing simultaneous channel close bug; adding test for it
2011-04-01 16:57:35 +01:00
964543d122
updating documentation for the consumer mechanism
2011-03-30 21:23:13 +01:00
077275abf5
adding amqp_direct_consumer and amqp_selective_consumer implementations; fixing tests
2011-03-30 18:55:19 +01:00
1f15d7edd1
changing the more usual exit reasons to {shutdown, Reason} for connections and channels
2011-03-29 16:02:00 +01:00
084239a7bf
merging in from default
2011-03-17 19:25:11 +00:00
59396dbef5
Added tests for consumer death notification
2011-03-04 17:11:41 +00:00
508f2d523b
merging in from default
2011-01-28 18:02:19 +00:00
8a0f9ef895
fixing connection_closing exit reason in channels
2011-01-28 17:37:44 +00:00
86fa202656
Merging bug23696 to default
2011-01-28 17:08:18 +00:00
0925cac294
Merging bug23410 to default
2011-01-28 13:16:29 +00:00
f65fcc0b7d
fixing subscribe nowait bug; adding test for it
2011-01-27 20:34:00 +00:00
1b053392c9
minor
2011-01-27 19:42:27 +00:00
91554a71e8
fixing amqqueue call bug; swapping test_direct and test_remote_direct in make test_suites and make all_tests; adding direct tests to common package test; updating rabbit_common.app
2011-01-27 17:19:16 +00:00
2359429a5a
merging in from default
2011-01-27 13:48:49 +00:00
47dbeb0e6d
rename ack_handler to confirm_handler and handle basic.nacks
2011-01-26 13:02:41 +00:00
e34348ff43
minor
2011-01-20 19:06:39 +00:00
79cca459ac
adding test suite for direct connection to different rabbit node; general test makefile cleanup
2011-01-20 18:51:19 +00:00
41f73f2a80
updating license year
2011-01-20 15:40:16 +00:00
03106aff67
fixing some merge conflicts; other cosmetic
2011-01-19 21:42:24 +00:00
28a197e8a2
fixing failing tests
2011-01-18 17:33:11 +00:00
3d1683e85e
merging in from default
2011-01-17 17:08:54 +00:00
6b711d0a3c
merging bug23562 into default
2011-01-17 15:49:04 +00:00
1185a327c1
Simplified nack tests in erlang client
2011-01-14 11:52:22 +00:00
852c066d25
Merge with default
2011-01-12 15:27:15 +00:00
02794d22dc
tests for basic.nack
2011-01-12 15:27:07 +00:00
604178e3ea
Beginnings of a test for basic.nack
2011-01-12 10:16:15 +00:00
e8002b9be0
Merging default into bug23055
2011-01-11 13:03:24 +00:00
80e6abcd41
Merge with default
2011-01-10 12:28:40 +00:00
a7fe9b6c72
merging in from default
2010-12-16 18:34:34 +00:00
308d3a5555
fixing ignored tests
2010-12-16 18:17:56 +00:00
88a8d6851a
merging in from default
2010-12-16 17:42:13 +00:00
b85eba361b
merging in from default
2010-12-16 15:41:06 +00:00
c1e4ac0a56
fixing some merge conflicts
2010-12-01 18:21:19 +00:00
e30ade2aaa
Some cosmetic fixes and a test for the pub ack functionality
2010-11-30 16:43:40 +00:00
554fa21b4b
Removed tests for requeue=false since we don't support this now
2010-11-30 15:36:31 +00:00
63de61d6c2
general refactoring in amqp_channel; checking if tag is already in use, on subscribe
2010-11-23 18:56:49 +00:00
ff4e91379b
extending timeout for method serialization tests
2010-11-23 15:59:29 +00:00
c23b52ac51
merging in from default
2010-11-16 15:11:03 +00:00
de890420dc
merging in from default
2010-11-16 13:32:19 +00:00
4b6831ab72
more consistent error reporting on failed handshake; heartbeating is now started between connection.tune and .tune_ok; the network connection follows the closing protocol if it fails to receive connection.open_ok
2010-11-04 22:18:37 +00:00
9bdcfde17f
updating license boilerplate
2010-11-04 17:35:36 +00:00
e99e48c851
fixing failing qos test
2010-11-04 17:17:50 +00:00
85abc07663
adding more tests for ensuring proper method ordering
2010-11-04 16:59:27 +00:00
9ddc8873d9
fixing methods overtacking each other
2010-10-28 18:35:58 +01:00
eed0fd2ab2
updating license boilerplate in all files
2010-10-28 16:37:10 +01:00
bb0d0a2b2e
changing connection_closing to a cast; updating amqp_dbg
2010-10-13 16:33:25 +01:00
0b4a9fce18
fixing merge conflicts
2010-10-12 19:57:17 +01:00
ec814c68b3
merging in from default
2010-10-12 18:39:16 +01:00
46a6500ee8
better error reporting on connection startup; fixing compile warnings
2010-10-11 20:45:19 +01:00
c35cd644e5
adding amqp_gen_connection module; all duplication between connection types has now been removed
2010-10-11 18:20:58 +01:00
be33721ad1
merging in from default
2010-10-05 12:23:07 +01:00
64b37f7c50
merging in from default
2010-10-04 12:51:05 +01:00
1540e75b28
Merging bug 23024 onto default. Drive-by correction to amqp_client.app.in
2010-10-01 16:50:53 +01:00
5df54a19a2
merging in from bug23024
2010-09-24 16:55:31 +01:00
13db1190fb
removing use of timer:sleep in tests; cosmetics
2010-09-24 12:33:59 +01:00
c8f83afbc2
Endless cosmetics
2010-09-20 15:44:12 +01:00
a1ee995808
connection now sends connection.close if it receives a channel method on channel0 or a connection method on a channel; better support for sending the server errors, using #amqp_error and rabbit_binary_generator:map_exception/3; tests for command_invalid errors due to server misbehaviour; other small fixes
2010-09-17 12:28:24 +01:00
d2676f736f
removing channel flow test
2010-09-15 17:30:55 +01:00
611ae43f21
fixing merge conflicts
2010-09-14 13:12:06 +01:00
78c6bc3e90
merging in from default
2010-09-14 12:57:32 +01:00
e1677c480d
Merge default into bug 23234 (now broker default works with erlang client on 23234 with new API)
2010-09-10 15:10:37 +01:00
cfadb215a2
Merging bug 22993 into default
2010-09-10 14:46:52 +01:00
92f9e7f233
Revert connection API to throw on error, as it does on default
2010-09-09 14:31:25 +01:00
50091a9c95
removing connection:start_link; making open_channel return {ok, Pid} | {error, Error}
2010-09-08 14:41:27 +01:00
60dd766b0e
Be more robust in waiting for process termination
2010-09-07 16:56:44 +01:00
19c5eae78c
Support linked and unlinked versions in the supervisor branch too. Unfortunately, there's no supervisor:start, only supervisor:start_child. Luckily, all the children are dynamically started, thus we can unlink it after it's created and before it has any children added. However, I still can't get the negative_test_util:channel_writer_death_test test to pass
2010-09-06 17:35:25 +01:00
955520a7ec
Turns out to be very important to support both start_link and start variants
2010-09-06 17:17:44 +01:00
1a3eafcf08
Support older API still on default
2010-09-06 12:09:00 +01:00
ef2faba6b1
Make amqp_connection API more OTPish and better abstracted (changes lifted from bug 22993)
2010-09-06 11:47:01 +01:00
481bba47fd
forgot to update amqp_dbg
2010-09-03 13:27:26 +01:00
cac9609396
fixing merge conflicts; more consistent children types in supervision hierarchy; removing amqp_channel_util module entirely; better error handling code; other smaller fixes
2010-09-03 13:20:19 +01:00
e9b1b31bb4
merging in from bug22993
2010-09-02 12:44:48 +01:00
ab2980beee
removing supervisor info item in connections; updating amqp_dbg
2010-09-01 18:27:43 +01:00
e97750e98b
debitroting; moving all children start code in the supervisors; other small fixes
2010-09-01 16:30:42 +01:00
2b9a3b8d33
fixing failing tests
2010-08-05 16:33:15 +01:00
6d1ae7a0d2
minor amqp_dbg fix
2010-08-05 13:00:35 +01:00
8b8056bb92
merging in from bug22993
2010-08-04 15:09:21 +01:00
fae295c7e9
fixing leftovers; need to find a way to get connection exit reason if it fails to start and return it as {error, Reason} from amqp_connection_sup:start_link/2
2010-08-02 18:15:59 +01:00
caedde16fc
supervision children are now properly started in supervisor's init; adding amqp_connection_specific_sup
2010-08-02 15:24:23 +01:00
c0f0decf2e
adding amqp_channels_manager; adding two tests for opening and closing channels repeatedly and in paralel; removing some duplication between network and direct connections; other refactorings; other small fixes
2010-07-30 15:24:17 +01:00
eca87dda95
cosmetic
2010-07-26 20:05:32 +01:00
ed20971d86
fixing some merge conflicts; moving heartbeat from main reader to the network connection module - the two heartbeating processes are added to amqp_connection_sup; other small fixes
2010-07-26 18:23:39 +01:00
e51fb6db69
major refactoring to supervise all processes; amqp_connection_{network,direct} processes now lives under amqp_connection_sup which uses amqp_infra_sup; amqp_channel now lives under amqp_channel_sup which uses amqp_infra_sup; amqp_channel_sup lives under amqp_channel_sup_sup which lives under amqp_connection_sup; adapting all processes to work with the supervision model; changing the way to start connections in the API; other smaller fixes
2010-07-26 17:31:21 +01:00
5d367771e3
amqp_main_reader is now a gen_server
2010-07-21 17:26:47 +01:00
5c33dc5218
cleaning up all makefiles in the client; make -j works; client does not do unnecessary builds anymore; source tarball makefile does not contain unnecessary make targets (some of which would not work); other smaller makefile fixes
2010-07-19 13:07:33 +01:00
Simon MacMullen
Emile Joubert
Matthias Radestock
Francesco Mazzoli
Alexandru Scvortov
Matthew Sackman
Vlad Alexandru Ionescu
Rob Harrop