summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Bumps version to 1.4.5 and updates Changelogv1.4.5Ask Solem2014-04-153-3/+22
|
* Handle SSL closed socket bugCraig Jellick2014-04-151-6/+11
| | | | | | There is a bug in the python socket library wherein attempting ssl.read/write on a closed socket raises an AttributeError instead of IOError. We need to catch that case and raise an IOError.
* Updates changelogAsk Solem2014-04-151-0/+32
|
* Adds table deserialization of types s, b, U, u, i, L, l, f.Ask Solem2014-04-151-12/+49
|
* Fix for Transport.__del__ for late threads at shutdown. Closes #36Ask Solem2014-04-151-3/+7
|
* Adds .coverage to gitignorev1.4.4Ask Solem2014-03-031-1/+1
|
* Bumps version to 1.4.4Ask Solem2014-03-032-2/+2
|
* Adds Craig Jellick to AUTHORSAsk Solem2014-03-031-0/+1
|
* Don't disconnect transport on SSL read timeoutCraig Jellick2014-03-031-0/+11
| | | | | | | | | | | | | | SSL socket read timeouts are not raised as socket.timeout exceptions. So, to have SSL socket read timeouts behave the same way as TCP socket read timeouts, we need to find them based off of their text and reraise as socket.timeout exceptions. This is similar to the fix in read_timeout in connection.py at: https://github.com/celery/py-amqp/blob/1.4/amqp/connection.py#L336-L337 and in fact the exception raised here is eventually caught by read_timeout and replaced with a socket.timeout exception. The problem is that the transport is already marked as disconnected at that point. So, we need to catch the timeout here in transport.py
* Updates ChangelogAsk Solem2014-03-031-0/+10
|
* Bumps version to 1.4.3 and updates Changelogv1.4.3Ask Solem2014-02-093-2/+13
|
* Adds Ionel Cristian Mărieș to AUTHORSAsk Solem2014-02-091-0/+1
|
* Merge pull request #33 from celery/fix-celery-1847Ask Solem Hoel2014-02-091-3/+2
|\ | | | | Don't read from the socket more than requested. This prevents stalling i...
| * Don't read from the socket more than requested. This prevents stalling ↵fix-celery-1847Ionel Cristian Mărieș2014-02-081-3/+2
|/ | | | issues (fixes celery/celery#1847).
* Missing version changev1.4.2Ask Solem2014-01-231-1/+1
|
* Bumps version to 1.4.2 and updates ChangelogAsk Solem2014-01-233-2/+11
|
* Always ignore server heartbeat if client heartbeat is 0/None (change for ↵Ask Solem2014-01-231-0/+4
| | | | issue #30)
* Docs introduction version is out of dateAsk Solem2014-01-171-1/+1
|
* Bumps version to 1.4.1 and updates Changelogv1.4.1Ask Solem2014-01-143-8/+16
|
* Fixes heartbeat and NoneType error. Closes celery/celery#1790Ask Solem2014-01-141-4/+5
|
* Make sure exchange auto_delete deprecation warning is visible. Closes ↵Ask Solem2014-01-131-2/+10
| | | | celery/kombu#296
* Fixes header in Changelogv1.4.0Ask Solem2014-01-131-0/+1
|
* Bumps version to 1.4.0 and updates ChangelogAsk Solem2014-01-133-2/+32
|
* flakesAsk Solem2014-01-1315-111/+125
|
* Funtests passingAsk Solem2014-01-131-16/+21
|
* Remove additional Py2.5 compatibility checksAsk Solem2014-01-137-72/+9
|
* Fixes for #30Ask Solem2014-01-131-9/+24
| | | | | | | | | - Sent back the negotitated heartbeat interval in tune_ok - Made `server_heartbeat` and `client_heartbeat` attributes of the class so that they are not dynamically created. - Renamed `heartbeat_interval` -> `heartbeat` for backwards compatibility.
* Adds Dustin J. Mitchell to AUTHORSAsk Solem2014-01-131-0/+1
|
* Merge branch 'djmitche/heartbeat'Ask Solem2014-01-132-20/+89
|\
| * refactor heartbeats to follow rabbitmq-java-client's modelDustin J. Mitchell2014-01-101-28/+32
| |
| * use a monotonic clock insteadDustin J. Mitchell2014-01-092-4/+59
| | | | | | | | | | This transplants the monotonic backport from kombu into this package, and uses it for heartbeats.
| * Revamp heartbeat handlingDustin J. Mitchell2014-01-081-19/+29
| | | | | | | | | | | | Be conservative in heartbeat behavior: send at the minimum interval negotiated by the server and the client, and do not artificially abort a connection if we do not receive any heartbeats.
* | Merge branch 'aguinet/master'Ask Solem2014-01-131-1/+1
|\ \
| * | Closes #27Adrien Guinet2013-12-171-1/+1
| |/ | | | | | | SSL socket timeout can return errno.ENOENT
* | Adds Dominik Fässler to AUTHORSAsk Solem2014-01-131-0/+1
| |
* | Merge branch 'dfa/master'Ask Solem2014-01-133-48/+11
|\ \
| * | Allow 'None' type in tables and arraysDominik Fässler2013-12-173-6/+11
| | |
| * | Remove duplicated tests casesDominik Fässler2013-12-171-42/+0
| |/
* | Adds promise.callsAsk Solem2013-12-051-1/+4
|/
* Adds bitdeli1.3Ask Solem2013-11-151-0/+3
|
* Updates Changelog and bumps version to 1.3.3v1.3.3Ask Solem2013-11-113-2/+13
|
* Save data read in transport before raising exception.Tommie Gannert2013-11-111-23/+30
| | | | | | | | | If the exception raised is due to a timeout, we might read from the transport again. In the old code, the data was discarded, causing loss of synchronization. This ensures the data we see is preserved for next time. Conflicts: amqp/transport.py
* Bumps version to 1.3.2 and updates Changelogv1.3.2Ask Solem2013-10-293-2/+13
|
* Message.channel must be channel object, not id. Closes #24Ask Solem2013-10-292-3/+8
|
* Do not set disconnected on EGAIN/EINTRAsk Solem2013-10-252-8/+28
|
* Bumps version to 1.3.1 and updates Changelogv1.3.1Ask Solem2013-10-243-2/+12
|
* Implements Connection.connected. Closes #22Ask Solem2013-10-242-12/+35
|
* AMQPError __str__ cannot return NoneAsk Solem2013-10-141-1/+1
|
* Doc fixesv1.3.0Ask Solem2013-10-041-1/+1
|
* Bumps version to 1.3.0Ask Solem2013-10-043-5/+5
|
View Lorry Controller Status