summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* cosmeticMatthias Radestock2010-03-031-12/+9
|
* Apparently, consistency is a good thingMatthew Sackman2010-03-032-3/+3
|
* Made the very generic invoke call into a slightly more specific flush callMatthew Sackman2010-03-033-14/+18
|
* Deal with the possibility of there being no queues to go throughMatthew Sackman2010-03-031-2/+5
|
* Implemented the restMatthew Sackman2010-03-033-10/+47
|
* Fixes to limiter - logic mistake in maybe_notify and some refactoring and ↵Matthew Sackman2010-03-032-19/+23
| | | | cosmetics
* Incomplete implementationMatthew Sackman2010-03-022-18/+65
|
* Problem: if the limiter is blocked due to a client flow, and the client then ↵Matthew Sackman2010-03-022-24/+29
| | | | issues a qos = 0, we can't have the limiter exiting. Thus limit needs to become a call, plus associated changes. Note that we *have* to know whether or not the limiter is about to stop - if we just wait for the exit signal then there's a race with a new qos setting being sent to an expiring limiter and the limiter thus getting lost. Also, the channel *must* be responsible for telling the queues to forget the limiter, otherwise again there's a race between a new limiter starting up and telling the queues about it, and the old limiter dying and telling the queues they're unlimited. This is why the limit_queues function cannot be moved into the limiter - the channel must drive this.
* lower priority of conserve_memory below that of 'info'Matthias Radestock2010-02-281-1/+1
| | | | | | | This really is more of a principled change than of practical significance since we are unlikely to end up with channels having mailboxes full of conserve_memory messages. The change brings 'default' closer to the 'bug21673' branch
* change priorities in amqqueueMatthias Radestock2010-02-281-3/+3
| | | | | We change all '8' to '7' - this has no effect but brings 'default' closer to 'bug21673'
* Merging bug 22188 into defaultMatthew Sackman2010-02-191-4/+4
|\
| * Use a suffix other than .rpm for file lists, to avoid confusion with rpms.David Wragg2010-02-191-4/+4
|/
* Added tag rabbitmq_v1_7_2 for changeset bacb333d7645Matthew Sackman2010-02-160-0/+0
|
* release v1.7.2rabbitmq_v1_7_2Matthew Sackman2010-02-162-0/+9
|
* Merge bug22363 into defaultDavid Wragg2010-02-151-2/+7
|\
| * Well without this, I can break the test frequently (> 50% of the time, with ↵Matthew Sackman2010-02-151-2/+7
| | | | | | | | [rabbit_tests:test_pg_local()|| _ <- lists:seq(1,10000)]. With this, it's still not brilliant - but then it actually can't be fixed properly, without extending pg_local. This fix is "good enough" to stop my machine barfing on it, though it is still racy - "we" could process our monitors and go onto the final pg_local test before pg_local processes the monitors in its mailbox (well, actually before the monitors *arrive* in its mailbox).
* | Merging bug 22361 ⇨ defaultMatthew Sackman2010-02-151-3/+2
|\ \ | |/ |/|
| * Avoid the use of spawn_executable on non-Windows platformsDavid Wragg2010-02-151-3/+2
|/ | | | | | | The Windows-only issue addressed in bug22282 requires the use of spawn_executable, which was introduced in OTP R13. But for other platforms we'd like to continue to support R12, and so we should stick to plain spawn.
* merge headsMatthias Radestock2010-02-124-9/+5
|\
| * Merging bug22262 ⇨ defaultMatthew Sackman2010-02-111-4/+0
| |\
| | * Both of the macports makefile targets are deploy-time, so merge themDavid Wragg2010-02-111-4/+0
| | | | | | | | | | | | | | | | | | | | | We need to know the REAL_WEB_URL in order to generate the protsfile and index the resulting macports tree, and all this has to happen at deploy time. So there is no point in distinguishing the indexing step.
| * | unix2dos has gone away on some systems, and was always an alias for todos ↵Matthew Sackman2010-02-111-1/+1
| | | | | | | | | | | | anyway (well actually fromdos) so todos should work on all systems currently
| * | Merge bug22336 into defaultDavid Wragg2010-02-112-4/+4
| |\ \ | | |/ | |/|
| | * Don't set env vars to the empty stringMatthew Sackman2010-02-092-4/+4
| | |
* | | cosmeticMatthias Radestock2010-02-111-2/+3
| | |
* | | minor optimisationMatthias Radestock2010-02-111-21/+20
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | when handling channel closure in the queue process, don't do any unnecessary work when we end up auto-deleting the queue. Specifically we don't rollback any transaction or attempt to deliver any messages. The formmer happens in the termination code anyway. The latter is a change in behaviour, but one that is permitted by the spec, and arguably deleting the queue as quickly as possible. The bug21673 branch already has changes along similar lines.
* | | cosmetic: moving things aroundMatthias Radestock2010-02-111-20/+20
| | | | | | | | | | | | to shrink the diff with bug21673
* | | minor refactor: s/Delivered/IsDeliveredMatthias Radestock2010-02-111-12/+12
|/ / | | | | | | | | | | | | ...in all cased where it refers to the state of a message rather than an event that has just occured A similar rename was done in the bug21673 branch, so this change brings default closer to that
* | Merging bug22332 onto defaultMatthew Sackman2010-02-1064-199/+199
|\ \
| * | update copyright notices to say 2010 where appropriateDavid R. MacIver2010-02-0964-199/+199
| | |
* | | minor cosmetic changes and refactoring in amqqueue_processMatthias Radestock2010-02-101-27/+25
| |/ |/| | | | | these bring it more in line with the bug21673 branch
* | Merging bug 22289 into defaultMatthew Sackman2010-02-097-179/+229
|\ \
| * | Fix rabbitmq-multi issue with parentheses in pathDavid Wragg2010-02-092-12/+18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | rabbitmq-multi would fail to start rabbitmq-server.bat if the path had parentheses in the first few characters of a path component (but not after the few few characters, due to short names being involved; so cases such as 'Program Files (x86)' do not trigger the issue). This fix makes rabbitmq-multi run rabbitmq-server.bat indirectly via cmd.exe, working around the erlang issue from bug20526. So we no longer need to involve short names, and by sprinkling double quotes in the right places we can handle parentheses too. To test this properly you need a path component with both spaces and parentheses in the first few characters, e.g. "foo ()".
| * | Merge default into bug22289David Wragg2010-02-0917-111/+798
| |\ \
| * | | Convert all batch files to use deferred expansionDavid Wragg2010-02-086-118/+154
| | | |
| * | | Remove exclamation mark from an error message.David Wragg2010-02-081-1/+1
| | | | | | | | | | | | | | | | | | | | It will be more trouble than it is worth to preserve this with deferred expansion.
| * | | Protect %*David Wragg2010-01-311-1/+2
| | | |
| * | | Convert rabbitmq-server.bat to use delayed expansionDavid Wragg2010-01-311-51/+58
| | | |
* | | | Merge bug22295 into defaultPaul Jones2010-02-093-3/+3
|\ \ \ \ | |_|/ / |/| | |
| * | | Stripped out sndbuf and recbufMatthew Sackman2010-02-053-3/+3
| | | |
* | | | Merge bug22303 into masterDavid Wragg2010-02-093-13/+75
|\ \ \ \
| * | | | Don't hardcode erts path, instead, searchMatthew Sackman2010-02-081-1/+3
| | | | |
| * | | | Some cosmetics, but also make service_manager_path be based off erlang_home ↵Matthew Sackman2010-02-081-7/+18
| | | | | | | | | | | | | | | | | | | | if it's not set
| * | | | Detect version of os_mon, and if we have version > 2.2.1 then we trust its ↵Matthew Sackman2010-02-042-6/+55
| | |_|/ | |/| | | | | | | | | | output
* | | | Merge bug22259 into defaultDavid Wragg2010-02-081-1/+1
|\ \ \ \
| * | | | Everything uses su except for macports which uses sudo. On OS X, man sudo ↵Matthew Sackman2010-02-041-1/+1
| |/ / / | | | | | | | | | | | | suggests that -E is needed to preserve environment. Manually checked and indeed it is. Hence added -E to the one call of sudo in entire codebase
* | | | merge bug22310 into defaultMatthias Radestock2010-02-052-8/+10
|\ \ \ \ | |_|/ / |/| | |
| * | | merge default into bug22310Matthias Radestock2010-02-0513-58/+665
| |\ \ \ | |/ / / |/| | | | | | | in order to resolve merge conflict going the other way
* | | | Merging bug 22309 onto defaultMatthew Sackman2010-02-052-2/+11
|\ \ \ \
| * | | | handle attempt to close connection with a non-connection pidMatthias Radestock2010-02-052-2/+11
| | | | | | | | | | | | | | | | | | | | | | | | | which was previously hanging at best and causing things to explode at worst
View Lorry Controller Status