Skip to content

Latest commit

 

History

History
257 lines (209 loc) · 12.3 KB

RELEASE_NOTES.md

File metadata and controls

257 lines (209 loc) · 12.3 KB
Raw

Version 2.0-rc1

  • RxTxComm has been successfully replaced with jSerialComm and tested with RTU
  • The whole codebase has been re-vamped to use modern constructs and practises, not a single class has remained untouched
  • Codebase is compatible with JDK 1.6
  • Javadoc has been extensively improved and fixed
  • A logging framework (log4j) has been included with a wrapper to improve performance and allow token passing
  • We have some tests!! Unit tests have been added for all automation friendly protocols (UDP and TCP)
  • Unit tests have been added that independently verify the j2mod TCP Slave functionality
  • Threading and concurrency has been hugely improved - removed synchronisation errors and corrected the handling of thread shutdown
  • Brought some sanity to the Facade package so that it is now usable in a production setting
  • Removed the serial BIN transport - it was proprietary and not taken up by any hardware vendors that I could find
  • Fixed the UDP transport for both Master and Slave so that it is some way approaching robust
  • Fixed the code layout - everything has headers, the layout is consistent, variable names and scope make sense etc.
  • Build system incorporates GPG signing and is compatible with one-click deployment to Maven Central through Sonatype

Version 2.0-rc2

  • Added ability to specify timeout for Facade classes ModbusTCPMaster, ModbusUDPMaster and ModbusSerialMaster
  • Fixed all the Unit ID checking when running as a Slave (myriad of problems - cannot see how it has ever worked)

Version 2.0-rc3

  • Fixed the timeouts for UDP/TCP and Serial listeners
  • Added more test cases

Version 2.0-rc4

  • Fixed an issue with the TCPTransaction class not initialising the Transport correctly
  • Removed all redundant casts
  • Made Transport available from Facade classes so that listeners can be added for Serial events

Version 2.0-rc5

  • Serial timeouts were being applied to a null transport

Version 2.0

  • Replaced log4j with slf4j

Version 2.1

  • Corrections to ASCII transport - thanks transparentink
  • Added fix for infinite listening loop - thanks martentamerius
  • Removed duplicate request handling code
  • Fix unit tests on linux #9 - thanks martentamerius
  • Improved logging of errors and warnings with stack trace #10 - thanks martentamerius
  • Facilitate multiple processimages in ModbusCoupler. #11 - thanks martentamerius
  • WriteMultipleRegisters (Function Code : 0x10) does not work. #14 - changed use of signed shorts to unsigned
  • Serial/RTU slower on 2.x #16 - worked around an issue with jSerialComm where setting timeouts on an open port injected a 200ms delay

Version 2.1.1

  • Corrected a problem with applying the timeouts when using the facade pattern

Version 2.1.2

  • Retrieve the localPort from opened socket if port is set to 0 #19
  • CR characters lost on ModbusSerialTransport #18
  • Corrected timeout setting issues on serial ports and harmonised setting timeout with IP methods

Version 2.1.3

  • Added support for Modbus RTU over TCP #21 - thanks axuan25
  • Code requires guards around a number of the functions #22

Version 2.2.0

  • Modbus Slave Multiple Port Support #23
  • Set ModbusCoupler to deprecated in favour of the new ModbusSlaveFactory and ModbusSlave sub-system
  • Fixed some javadoc errors - thanks jan
  • Timeout error on socket listener - thanks jan
  • Now allows for slaves to share same socket but different protocols

Version 2.2.1

  • Amend TCPMasterConnection socket timeout handling #27 - thanks David

Version 2.2.2

  • Fixed NPE in close method of SerialConnection if Comms port was never successfully opened

Version 2.3.0

  • Adjusted write timers with fiddle factors to allow for idiosyncratic hardware
  • jSerialComm library decoupled from implementation #28 - thanks Felipe

Version 2.3.1

  • Length of ReadWriteMultipleResponse indicates 1 byte too much? #31 - thanks elasticoder
  • Function code 0x17 first performs a read and then the write. #32 - thanks elasticoder
  • Double socket creation in ModbusTCP. #33 - thanks stoorm5

Version 2.3.2

  • Increased maximum queue length for incoming connections. #38 - thanks martentamerius
  • Incorrect javadoc for ModbusSlaveFactory #35 - thanks bertrik
  • Adds setRetries and setCheckingValidity services into ModbusTCPMaster #30 - thanks ericauguie
  • Added RTU over TCP #43 - thanks eli-mcgowan
  • Fixed setting timeout of TCPMaster does not get applied to transactions #41
  • Reduce an NPE - Serial port connection #40
  • Fixed issue with serial write bytes delay if data bits not specified

Version 2.3.3

  • Specify IP address in ModbusSlaveFactory #46 - thanks MindVark
  • Added getConnection() to AbstractSerialConnection - thanks nnadeau
  • Reading full stream instead of parts - thanks liebehentze

Version 2.3.4

  • Fixed NullpointException at handleRequest in AbstractModbusListener #47
  • Add checks to see if serial port is available during connect #45

Version 2.3.5

  • Fixed a typo in the failure debug message
  • Fixed the omission of the bit count when reading a DiscreteInput request

Version 2.3.6

  • Debug logger messages causing garbage collections even when DEBUG mode is not enabled #53
  • RS485 echo in RTU mode #51 - thanks david

Version 2.3.7

  • Fixed a format error in the thrown connection retry catch block
  • Propagate ModbusTCPListener timeout to the TCPSlaveConnection #59 - thanks javi
  • Not possible to override handleRequest in ModbusSlave #57 - thanks javi
  • Extending ModbusSlave exceptions list

Version 2.3.8

  • Added a sleep timer to the TCP/UDP/serial transaction retry loops
  • Added the timeout value to the serial write
  • Responded to vulnerabilities highlighted by FindBugs

Version 2.3.9

  • Critical regression introduced in previous release whereby the sharing of stream buffers was being trampled over
  • Serial CRC was not correctly checking both bytes

Version 2.4.0

  • The transport RTU over TCP was not being set correctly after a comms failure
  • The Slave interface was not always correctly assigning the headers (Transaction ID, Function Code. Unit ID) for all types of requests
  • Fully tested the RTU over TCP transport with some real hardware

Version 2.4.1

  • Fixed some error messages
  • Added more logging and made the TCP transaction a little more robust

Version 2.4.2

  • Fixed a critical issue with RTU over TCP when the connection is set to re-connecting. Initial connection correctly uses ModbusRTUTCPTransport but subsequent connections use ModbusTCPTransport - RTU Over TCP Resend Bug #71

Version 2.5.0

  • Upgraded to the latest version of JSerialComm - this fixes issues with closing Windows COM ports and many other problems It also allows us to manage the built-in latency timeouts within JSerialComm
  • Add unit tests for the serial implementation of RTU and ASCII (Windows only)
  • Made the closing and shutdown of a Serial Listener more robust
  • Fixed an issue whereby if a Serial Slave was created using the ModbusSlaveFactory is closed independently, and the then the slave is reused with different serial parameters, the old parameters would have been used instead
  • Removed all JavaDoc compilation warnings

Version 2.5.1

Note:- v2.5.1 is not backwards compatible with v2.5.0

Some changes may be required for applications using this version. The changes bellow detail where a non-compatible change has been made.

  • Removed the deprecated ModbusCoupler implementation - users should switch their code to use SlaveFactory and apply their ProcessImage instances to the slave directly The new implementation means that ProcessImage is not coupled to a slave and can be shared across slaves and unit IDs (NOT BACKWARDS COMPATIBLE)
  • Added a method getError to ModbusSlave to enable users to get any startup errors after a slave is opened #72
  • Fixed the issue whereby serial slaves will respond with error messages for requests that are targeted for them #62
  • Updated the documentation and byline
  • Moved the useRtuOverTcp flag to constructor #75 (NOT BACKWARDS COMPATIBLE)
  • Fixed getConnection.isOpen() still true after serial-port removal #74
  • Implemented inter-frame delays for serial transactions to match the Modbus specification #62
  • Fixed - Why doesn't DigitalOut extend DigitalIn #86
  • Made library an OSGi bundle #67 thanks amitjoy
  • Removed all final qualifiers where it prevents extensibility and doesn't add any value #85
  • Added default thread names to listeners and methods to get/set it on the slave #80
  • Fixed Error in ReadInputDiscretesRequest class SetBitCount() method. Can't access the last address. #87

Version 2.5.2

  • Removed a lot of synchronized decorated methods to put the onus on the caller and speed up processing. #88

Version 2.5.3

  • Added check of serial response to make sure it is from the same unit and function code as the request
  • Return the new value of the register after writing to it (#91)
  • Fix data length in MaskWriteRegisterResponse. Add facade methods for Mask Write Response function. Add unit tests. (#92)
  • Upgraded jserialcomm to v2.3.0

Version 2.5.4

  • Serial slave not reading response from other slave #76
  • Upgraded jserialcomm to v2.4.1

Version 2.5.5

  • Upgraded jserialcomm to v2.5.1

Version 2.5.6

  • Added a finally block to ensure that afterMessageWrite event handler is invoked, even if the message fails to write

Version 2.5.7

  • ModbusTCPTransaction Error message is not shown in a specific case #100

Version 2.5.8

  • Upgraded jserialcomm to v2.5.3
  • ClassCastException when timeout?? #96 - made all Master facade methods thread safe
  • Code cleanup to lower Sonar score

Version 2.5.9

  • Desensitised the serial writes to allow more time to send

Version 2.6.0

  • Sanitised the setBaudRate method to make it part of the serial parameters only
  • Removed all the redundant test data for the old command test
  • Made the default SerialConnection more resilient to the connection not yet being open
  • Added retries to the serial port connection with a retry delay
  • Fixed a serious synchronisation problem with the ope/close methods in the use of SerialPort
  • Added a modpoll style command line interface to test serial connections

Version 2.6.1

  • Upgraded to jSerialComm 2.6.0

Version 2.6.2

  • Downgraded jSerialComm until jSerialComm #277 is resolved
  • Upgraded log4j to prevent security issue
  • added ModPoll class to mirror ModPoll exe capabilities
  • Created fat executable jar for testing ModPoll features
  • SHOULD NOT BE USED

Version 2.6.3

  • Fix missing logging properties in assembly
  • Fixed the release so that it isn't the fat jar that is released (2.6.2 error)
  • SHOULD NOT BE USED

Version 2.6.4

  • Fixed the release so that it isn't the fat jar that is released (2.6.3 error)
  • SHOULD NOT BE USED

Version 2.6.5

  • Upgraded to jSeralcomm 2.6.2 that contains fix for jSerialComm #277
  • Added some code cleanups as suggested by Sonar
  • Fixed issue where slave listeners (TCP/UDP/Serial) fail silently if they cannot bind to a port
  • Bumped log4j version for testing
  • Added isConnected() method for al master facade classes #103

Version 2.7.0

  • TCP Connection idle timeout #108 thankyou akochubey2004
  • Changed tests to use 127.0.0.1 rather than localhost

Version 3.0.0

  • Fixed some stray documentation
  • Add support for explicitly configuring RS-485 mode #117 thanks sirhcel
  • Removed checks on readInputDiscretes #114 thanks kazuyatada
  • Switched to scheduled executor service instead of timer #115 thanks mkurt

Version 3.1.0

  • Updated to latest version of jserialcomm and fixed log4j version

Version 3.1.1

  • Add remaining RS-485 configuration parameters for Linux #118

Version 3.2.0

  • Move parts of DEBUG logging to TRACE #126
  • Bump jSerialComm for fixed POSIX signal handling (#138)
  • ModbusTCPMaster: Fix incorrect reconnect parameter documentation (#137)
  • Added equals & hashcode method to ModbusTCPMaster & TCPMasterConnection (#127)
  • modbus slave factory getSlave fix
  • Bumped the version

Version 3.2.1

  • Add capability to bind the Socket to specific local interface. (#135)
  • Serial port is left hanging in OS if it gets physically disconnected (#144)
  • Modbus Serial new event to serial disconnected (#148)
  • Updated jserialcomm to 2.10.4

Version 3.2.2

  • Add support for providing an optional java.net proxy for. (#151)
  • Added support for testing serial comms on Mac/Linux