MSR with reverse proxy

[toggledbits]

Build 21168

• 0000220 Using default values on Telegram notification action causes UI to (incorrectly) flag error in field.
• 0000222 A broken reference in the "Run" action may cause an exception.

An experimental (still) exporter/feeder for InfluxDB is also available for testing.

POTENTIAL BREAKING CHANGE FOR HOMEASSISTANT

Because "state" values on HomeAssistant can be any data type, the data type for the value_sensor capability attribute value was destabilized; it's supposed to be numeric. But some Hass entities were assigning string values, because they were generally being identified as a sensor class device, but had non-numeric state values. The float-iness of value_sensor.value is now enforced (and once again consistent across all controllers). But this enforcement means any non-convertible string (string that doesn't contain a number) will result in null in value_sensor.value. To make sure that the original string is still available, a new string_sensor capability has been added, and Hass sensor class devices where the type cannot be explicitly determined from the device class will have both the value_sensor (with a float, if possible) and string_sensor (with the string) capabilities/values. What this change may break is any condition in triggers or constraints that use string tests against the value_sensor.value attribute (e.g. testing if value_sensor.value is equal to the word "idle"). You will need to switch those conditions to use string_sensor.value. Numeric tests are not affected by this change and so do not need to be modified. You will only run into trouble if you have been testing a value_sensor against a string. A check has been added with a warning notification to simplify finding these cases in your rules if you have any.

[toggledbits]

Build 21189

• 0000224: Unable to save rule set or reaction is only change is to switch a notification profile (doesn't enable save button)
• 0000223: Support for newer-style Hass color mode attributes (RGB[W] devices)

Fix an issue with the /variable/set HTTP API throwing an error (not new, but newly discovered and no PR open; attn @cw-kid ). Support for Hass up to 2021.7.1. Updated to new lexpjs with numeric coalesce operator. Internal improvements in the loading of extensions (notifiers, controllers, etc.). Update icon library to latest version. Now supports InfluxDB natively.

[toggledbits]

Build 21196

This build contains the eZLO interface (development level). There are no other significant user-facing changes, although many changes have been made in preparation for the upcoming 1.0 release (this is release candidate 2).

Users of the archive package (i.e. not docker) please npm update --no-save in your reactor directory after unpacking the archive.

Please refer to the documentation for configuration instructions.

I expect lots of issues in every range of possibility. I don't have many devices on my eZLO Plus to test with, so we're going to cover a lot of new ground, mostly by long distance, as I mentioned before. Please open PRs in Mantis for everything you find. Please post/upload all of the files prefixed ezlo_ from the logs directory with every report. If you are reporting an issue with a specific device, please be as specific as possible about which device it is and how it's not working. If you are including a log snip, please remember to include plenty of context (extra lines) both before and after the part you think is relevant (more is better than less).

In addition to the required configuration above, please add the following to the end of your logging.yaml file:

  EzloController:    # this line indent TWO spaces
    level: 7         # this line indent FOUR spaces

[toggledbits]

Build 21197 (1.0 release candidate 3)

• Additional capabilties for eZLO devices
• PR 0000229: Fixed an issue with attempt NaN write when certain sun conditions cannot be calculated in some "extreme" latitudes (e.g. sunset in Tromso NO in mid-summer: sun doesn't set; no astro dusk/dawn in mid-summer northern UK, etc.).
• PR 0000228: Fixed an issue with sequence (..after..) condition option not reselecting predecessor condition on subsequent rule edit.

I don't regard this as a vital update; this is mostly just informational. But if you're affected by either of the above PRs, you may want to make the upgrade.

On track for 1.0 release in 7/19. I've already changed the structure of the download directory.