Age | Commit message (Collapse) | Author | |
---|---|---|---|
2015-02-17 | Changed the way that we process a single job. | Steve Kemp | |
We now work on the assumption that a single "job", pulled from the queue, might contain multiple "test" objects. These are the instances of the protocol-testers which we actually execute. This is part of #9558. | |||
2015-02-05 | More simplification - don't pass state around. | Steve Kemp | |
Rather than passing our settings-object around, as well as specific settings that are read from it, just pass the object. The worker can read the settings directly if/when it needs to. | |||
2015-01-21 | Read the queue-type from the settings-object. | Steve Kemp | |
This replaces redis as a hard-wired default. | |||
2015-01-19 | Updated to use our queue-abstraction. | Steve Kemp | |
This means we can queue/dequeue to either Redis or Beanstalkd. | |||
2013-08-01 | Updated the cool-off period to work after every failed test, bar the | Patrick J Cherry | |
last one. | |||
2013-05-15 | Use log_message() not puts(). | Steve Kemp | |
2013-05-14 | We only sleep between tests if we saw a failure. | Steve Kemp | |
2013-05-14 | Ensure repeat_count and sleep values are integers. | Steve Kemp | |
2013-05-14 | Allow retry-behaviour to be tweaked. | Steve Kemp | |
We can now sleep between retesting, and we configure the number of repeats via the configuration file. | |||
2013-05-14 | Comment update. | Steve Kemp | |
2013-04-23 | The name of the beanstalkd tube is passed into the constructor now. | Steve Kemp | |
2013-01-24 | Updated to log duration of test-runs with our alerters, if they | Steve Kemp | |
implement a "duration" method. | |||
2012-12-12 | The dequeuing process passes the settings to the wroker. | Steve Kemp | |
2012-11-30 | Hard-wire repeat for the moment. It'll come from the configuration | Steve Kemp | |
file shortly. | |||
2012-11-26 | Updated so that each alert-type has its dedicated alert-target. | Steve Kemp | |
2012-11-26 | Allow multiple alerters. | Steve Kemp | |
2012-11-26 | Explicitly watch a "tube" - a named queue - to avoid stomping on | Steve Kemp | |
anything else running under beanstalkd. | |||
2012-11-24 | Renamed 'custodian/protocoltest.rb' -> 'custodian/protocoltests.rb' to ↵ | Steve Kemp | |
match what we did with the alerts. --HG-- rename : lib/custodian/protocoltest.rb => lib/custodian/protocoltests.rb | |||
2012-11-24 | Set the alerter-target, if we can find one. | Steve Kemp | |
2012-11-24 | Use the factory-method to instantiate an alerting method. | Steve Kemp | |
2012-11-24 | Move the protocol-test includes into a single file to avoid duplication. | Steve Kemp | |
2012-11-24 | Simplify test-execution logic, and count the test-runs from 1 rather than ↵ | Steve Kemp | |
zero. | |||
2012-11-24 | Invert the success/failure of tests if appropriate. | Steve Kemp | |
2012-11-23 | Updated the alerter to be in the Custodian name-space, and the worker | Steve Kemp | |
to invoke it from that. Currently alerts are disabled. TODO: Fix this. | |||
2012-11-22 | Removed json library - we don't need/use it. | Steve Kemp | |
2012-11-22 | Updated to use the new string-body from jobs, and the new factory to ↵ | Steve Kemp | |
generate tests. | |||
2012-11-22 | Moved the worker into the custodian namespace. | Steve Kemp | |
2012-11-22 | Renamed class Custodian -> Custodian/Worker | Steve Kemp | |
--HG-- rename : lib/custodian.rb => lib/custodian/worker.rb |