CHANGELOG.md

Change Log

All notable changes to this project will be documented in this file. See standard-version for commit guidelines.

2.1.0 (2018-02-18)

Bug Fixes

• (#104) improve poison message handling (9d5ccc3)
• (#109) randomize initial connection index based on pid or hostname (2d6f372)
• (#74) correct errors in direct reply queue implementation (cccdf09)
• (#78) make it possible to handle messages from a unique queue using the alias/friendly name (50045a8)
• correct issue where publishes and requests before connection or configuration had completed would get lost (1f58c0b)
• remove remaining lodash references in core that got missed (9aa3b04)

Features

• (#87) add support for scatter-gather pattern via request (db23b48)
• (#93) add support for bulkPublish operation (d2df5ea)
• add support for purging queues (19a678d)

2.1.0 (2018-02-18)

Bug Fixes

• (#104) improve poison message handling (9d5ccc3)
• (#109) randomize initial connection index based on pid or hostname (2d6f372)
• (#74) correct errors in direct reply queue implementation (cccdf09)
• (#78) make it possible to handle messages from a unique queue using the alias/friendly name (50045a8)
• correct issue where publishes and requests before connection or configuration had completed would get lost (1f58c0b)

Features

• (#87) add support for scatter-gather pattern via request (db23b48)
• (#93) add support for bulkPublish operation (d2df5ea)
• add support for purging queues (19a678d)

rabbot

2.0.0

• Allow publisher confirms to be turned off per exchange
• Moving towards ES6 features
• Clean up cases where arguments were being re-written (perf issue)
• Reduce cases where properties are introduced to hashes after instantiation (perf issue)
• Rewrite tests into single modules with individual config to make contributing & maintenance easier
• Move build system to Travis
• Use coveralls to track code coverage
• Remove lodash dependency
• Remove when dependency
• Change logging to bole
• Change commit style to conventional commits
• Adopt semistandard as format style
• Improve README badges
• Break docs into multiple docs
• Add code of conduct
• Add contributor guidelines
• #80 - fixes issue where multiple messages were nack'd in nobatch mode
• #42 - fixes issue regarding broken build status
• Fixes for: #120, #107, #97, #42

1.0.x

1.1.0

• feature (#72) - support sending persistent messages to default exchanges (thanks @bmatson!)
• feature (#66) - support passing SSL certs as string (thanks @mrfelton!) making progress callback an optional argument
• #71 - leave error handler attached to connection object to try and prevent errors from being thrown from the AMQP library
• #69 - fixes an issue where the *.connection.configured event was not correctly being raised
• #64 - added mention of deadLetterRoutingKey option to queue declaration
• #57 - fixes (with breaking change) by removing dependency on deprecated when feature and

1.0.8

• #52 - serialize arrays as JSON by default
• Change connection timeout to default to 2 seconds
• #49 - bug fix for exchange recreation when reconnecting
• #33, #54 - bug fixes around graceful shutdown

1.0.7

• #50 - deleting queue after starting subscription causes channel error
• bug fix - ensure correct SNI is used with SSL connections when using multiple endpoints
• improvement - switch UUID lib to uuid since node-uuid is deprecated

1.0.6

• #38 - Correct race conditions in queueFsm and exchangeFsm causing errors during reconnection
• #37 - Add ability to capture Rabbit generated queue names
• #36 - Make options optional for addExchange and addQueue
• #19 - Add support for publishing arrays directly (as buffers)
• Added Drone build
• Update dependencies to latest
• Defect - queues and exchanges with the same name shared channels causing serious problems
• Improvement - add support for default exchange
• Improvement - changed from jshint to eslint

1.0.5

• Improvement - remove Vagrant in favor of just using Dockerfile, updated instructions

1.0.4

• Bug Fix - Publish no longer throws errors if made before calling addConnection or configure or if exchange does not exist

1.0.3

• Bug Fix - #26 - fixed a bug preventing bindQueue from working on unique queues
• Bug Fix - #22 - addConnection did not return a promise (thanks @mkozjak)
• Enhancement - publishing a number as a body gets converted to a string (thanks @brandonpsmith)
• Enhancement - routing keys can be changed on binding (thanks @droidenator)

1.0.2

• Bug fix - corrected bug causing connection.unreachable event not to be prefixed with connection name correctly (thanks @Cyri-L)

1.0.1

• Bug fix - uri parsing should not including leading slash between host specification and vhost

• Add section to README about logging

• Feature - add default and custom strategies for returned mandatory messages

1.0.0 - issues addressed from wascally

• 103 - gulp test no longer throws expect.js errors
• 111 - support exclusive subscriptions when calling startSubscription
• 112 - add shutdown method to allow node to exit
• 116 - support multiple serializers
• 119, 115, 107 - resubscribe to all previously subscribed queues after reconnect
• 121 - rejectUnhandled should now correctly reject unhandled messages
• 122 - guarantee support for type and routing keys set to ''
• Feature - unique property on queue creation allows for "single use" exclusive queues with TTL
  • "hash" adds a numeric value
  • "id" uses client id for clearer ownership
• Feature - add machine and process info to connection properties for easier identification in management console
• Feature - consumer tags are now based on "client id"
• Feature - handlers can now be scoped/limited by queue name
• Feature - allow custom limits for deferred published messages
• Feature - add support for AMQP URIs when defining connections
• Feature - add ability to stop a running subscription to the queue
• Improvement - use "client id" to create response queue names
• Improvement - more warning level logging around channel and connection disruption
• Improvement - ensure UTC timestamps are put on published messages
• Improvemet - reject reply promise with error if no replyTo address is provided
• Improvement - set appId to "client id" on publish
• Improvement - introduce limit to the number of messages an exchange will cache waiting for a connection
• Improvement - handle blocked and unblocked broker events internally
• Improvement - limit the number of messages stored per exchange while waiting on a connection
• Bug fix - do not resolve close on connection until all publishes have been confirmed
• Bug fix - use reject instead of nack when queues are in noBatch
• Bug fix - noop nacks and rejections when consuming in no-ack mode
• Bug fix - channels closed by the broker should still be re-acquired with redefined primitives
• Bug fix - check for existence of headers before checking for direct-reply-to header. Thanks Matt Young (@mashu-daishi)
• Breaking - no longer recover automatically from a user terminated connection
• Breaking - limit number/duration of connection retries
• Breaking - messages are not auto-re-published when a connection or channel fails

Significant or Breaking changes

• reply signature has changed to support control over serialziation
• response queues are named very differently, shouldn't break code, but worth noting
• wascally.iomonad logging namespace changed to wascally.io
• timeouts added to requests

wascally - preserved for history

0.2.*

0.2.10

• 107 - Correct improper connection timeout configuration and update README
• Update dependencies to latest versions
• Use latest whistlepunk log api

0.2.9

• 91 - Provide publish timeout
• 98 - No-op duplicate exchange, queue and binding declarations
• 96 - Fall back to routing key if no type is provided for routing
• 94 - Update amqplib dependnecy to 0.4.0 to support Node 4
• 77 - Add explanation of connection events
• 95 - Add test coverage for wildcards in handle
• Change log namespaces to use '.' delimiter instead of ':'
• Add example for VirtualBox in Vagrantfile.sample - thanks, @josephfrazier
• Correct configuration and README regarding ports, credentials, and clustering - thanks again, @josephfrazier

0.2.8

• Bug fix - Resolve issue #64 working with node-config no longer throws exceptions by avoiding get call
• Minor improvement to README to better explain publish promises

0.2.7

• Bug fix - correct memory leak in req/res by leveraging new postal feature

0.2.6

• Bug fix - calling destroy on a queue or exchange should defer until they are in ready state.
• Update machine and monologue versions
• Remove when.promise and event handles in favor of deferred promises and transitions

0.2.5

• 65 - Bug Fix: setting replyQueue to false caused publish to fail silently.
• 63 - Add uri property to connection object emitted for 'connected' event.
• 61 - Bug Fix: correct SSL URIs - thanks, @longplay
• Improvements to connection clean up (specifically around resolving outstanding messages on queues)
• ConnectionFSM - Only emit 'connected' when establishing a new connection, use 'already-connected' otherwise.

0.2.4

Thanks to @dvideby0 and @neverfox for identifying and providing code to help reproduce bugs 39 and 57.

• 39, 57 - Bug fix: acking responses did not resolve them causing them to pile up in a response queue
• Bug fix: setting replyQueue to false throws exceptions
• Bug fix: replying from a service with replyQueue set to false fails
• Spec update - request spec in integration tests failed to ack messages causing closeAll to hang

0.2.3

• 47 - Added support for noBatch to queues, thanks @derickbailey!
• 45 - Bug fix: port option was being ignored, thanks @esatterwhite

0.2.2

Bug fix - bindExchanges flat out broken. Thanks to @leobispo for the catch & fix.

0.2.1

Special thanks to @neverfox for finding and reportig 38 & 39 - both serious problems and very difficult to find/reproduce in any kind of automated test.

• 38, 39 - I/O getting blocked when publishing at high frequencies (think for/while loops).
  • Removed one-time failed event handler from publish call
  • Cache reject callbacks from publish
  • On publish confirmation, remove reject from deferred array
  • On exchange connection failure, invoke all rejects in deferred array
• 37 - document use of close and closeAll calls
• Correct improper use of .then( null, ... ) which was creating additional promises.
• Update whistlepunk version
• Include biggulp to simplify the gulpfile (yay?)

0.2.0

• Add logging support via whistlepunk
• Add logging statements to assist with troubleshooting/debugging
• 24 - Connection should not close until after all queues have completed batch processing (only applies to user initiated connection shutdown)
• 30 - Escape passwords to be connection URI safe
• 17, 19 - Unhandled messages
• Nack unhandled messages by default
• Provide configurable strategies for handling unhandled messages
• 26 - Support custom reply queue definitions per connection
• Add behavioral specs to improve coverage and testing story
• Fix bug in reject batching implementation
• Refactor of exchange and queue implementation into channel behavior and FSM
• Reject exchange and queue creation promises on failure
• Reject publish and subscribe calls on failed exchanges and queues
• Bug fix - closing a connection didn't reliably clean up channels, exchanges and queues
• Bug fix - a failed connection that had been closed would continue to attempt reconnecting in the background
• Bug fix - configure doesn't reject the promise if a connection cannot be established

prerelease 8

• Add connection timeout
• Add @derickbailey to contributor list

prerelease 7

• Add demos and documentation to better explain handlers
• Allow replies to provide a replyType without specifying more parameter
• Add support for per-message expiration
• Add support for reject (nack without re-queue)
• Code clean-up / addressing linting errors
• Fix README issues
• Correct typo in spec
• Code clean-up / addressing linting errors