Closed ssalonen closed 3 years ago
Note to reviewers. Same comment I had in PR #11 applies here:
Methods now raise EOFException. Should be fine as the modbus transport in openhab-core already handles IOException which is parent of EOFException.
'Integration' tests against temporary modbus tcp server in openhab-core pass with the modified jamod:
[INFO] Results: [INFO] [INFO] Tests run: 1074, Failures: 0, Errors: 0, Skipped: 0 [INFO]
This confirms that the basic functionality should not be broken with this fix.
Positive test result from the user https://community.openhab.org/t/modbus-binding-with-sma-inverter-missing-lower-byte/119467/35?u=ssalonen
This resolves long-standing issue with SMA inverters, see https://github.com/openhab/openhab-addons/issues/5347
@kaikreuzer could you please review? Perhaps we can get this to the next oh release as well.
@ssalonen Shall I do a new release or do you expect further changes in the next days/weeks that we could cumulate?
@kaikreuzer thanks. I'm not expecting any further changes in the short term... We have had exceptionally many issues identified now recently :)
@ssalonen Version 1.3.3.OH has been deployed to our Artifactory.
Additional correction to PR #11. PR #11 Only fixed Modbus TCP server implementation (readRequest). This is not used in openHAB binding at all.
Pre-compiled modbus transport that uses this version of jamod: org.openhab.core.io.transport.modbus-3.1.0-2021-04-20-tcp-transport-fix-SNAPSHOT.zip
Intention to fix SMA inverter behaviour which seems to split the modbus response in multiple tcp packets. Apparently this seems to be rare otherwise. See https://community.openhab.org/t/modbus-binding-with-sma-inverter-missing-lower-byte/119467/29 and https://github.com/openhab/openhab-addons/issues/5347 .
TODO
Signed-off-by: Sami Salonen ssalonen@gmail.com