- Important rpc fix w/ [email protected], fixes rpc replyTo queue leak
- Fixed some wrong
sinon.assert.calledOnce
calls asserting args -> sinon.assert.calledWith
- Updated all out-of-date dependencies
- Important rpc fix w/ [email protected], fixes rpc request sendOpts
Application:
- Added
app.prefetch(...)
to support prefetch options forconsumerChannel
Readme: - Added RPC section and links to amqplib-rpc
Application:
- Fix
app.close
not closing thepublisherChannel
Context:
- Added
context.checkQueue()
- Added
context.checkReplyQueue()
- Use
delegates
for message accessors/getters oncontext
- Added
context
delegates formessage
fields
getterproperties
gettercontent
accessormessageAcked
accessor
- Added
context
delegates formessage.properties
headers
getter
- Added
context
delegates formessage.fields
exchange
accessorroutingKey
accessordeliveryTag
getterconsumerTag
getterredelivered
getter
Package:
- Remove unused
ee-first
- Update standard to 6.0.5
Context:
- Added
context.toJSON()
Context:
- Moved
request
logic out toamqplib-rpc
- Moved
reply
logic out toamqplib-rpc
Package: - Removed
uuid
Application:
- Added optional
Context
arg tomessageHandler
for easier testing NoAckErr: - Improved stack Debug:
- Fixed issues w/ filepaths logged CastToBuffer:
- Moved to external module
cast-buffer
Readme: - Added link to coworkers-test Package:
- Removed callsite dependency
Application:
- Fix broken
new Application(schema)
constructor now acceptsschema
oroptions
Application:
- Remove close/exit handlers before closing rabbit connection and channels in
close
ClusterManager:
- Propagate process.env to workers
Application:
- Sends process messages for connect and close states (useful for process management)
- 'coworkers:connect', 'coworkers:connect:error', 'coworkers:close', 'coworkers:close:error'
- SIGINT handler now throws uncaught exception if graceful shutdown fails
- When using clustering master process
connect
will wait for all workers to connect to rabbitmq- if any fail, it will yield an error and shutdown all other workers
- Unexpected
connection
'close' events now throws an error to kill the process - Unexpected
consumerChannel
'close' events now throws an error to kill the process - Unexpected
producerChannel
'close' events now throws an error to kill the process - Unexpected
connection
'error' events now throws the error to kill the process - Unexpected
consumerChannel
'error' events now throws the error to kill the process - Unexpected
producerChannel
'error' events now throws the error to kill the process
ClusterManager:
start
waits for all worker forks to finish, if any errors it will throw a special error
Added retries w/ exponential backoff to killWorker
and spawnWorker
If spawning a worker fails repeatedly it will throw an uncaught exception to crash the master
Fix require in index.js Readme updates
Initial release Application:
- use, queue, connect, close, error handler Context:
- connection, consumerChannel, publisherChannel,
- queueName, message, deliveryTag, messageAcked, state
- queueOpts, messageOpts,
- ack, nack, ackAll, nackAll, rejectAll,
- publish, sendToQueue, request, reply