Jørgen Austvik [Sun, 11 Dec 2022 15:09:53 +0000 (16:09 +0100)]
[nanoleaf] More color for less network calls (#13893)
* [nanoleaf] More color for less network calls
This is a refactoring that moves the "get panel color" out of the
panel handler and into a separate class, with callbacks.
This makes us do only one REST call to get colors instead of one per
panel that is a thing. It also lets us retrieve colors for all panels -
also those that doesn't have a thing in OpenHAB,
While testing this out, I found a bug where solid colors set in the app
wasn't reflected in neither the controller nor panel channels, and that
should also be fixed now.
Cody Cutrer [Sun, 11 Dec 2022 14:54:10 +0000 (07:54 -0700)]
[homekit] delay starting until a particular StartLevel (#13877)
* [homekit] delay starting until a particular StartLevel
instead of a blind "delay by seconds", the new default won't even
attempt to start until items are loaded (both file-based and UI-based),
with an additional config to let the user delay it any further (in
case they have items coming from JSR223 scripts that can't run until
at least SL20 anyway).
Upgrades the included openhab-js version to 3.1.0, which uses the new
caches from core (introduced in
https://github.com/openhab/openhab-core/pull/2887) and provides many
doc improvements.
Removes the SharedCache from the addon because this functionality is
now provided by core (see
https://github.com/openhab/openhab-core/pull/2887).
Holger Friedrich [Sun, 11 Dec 2022 11:47:55 +0000 (12:47 +0100)]
[knx] Improve handling of serial gateways (#13897)
* [knx] Improve handling of serial gateways
Take over initialization logic from KNX IP gateway for serial gateway.
Properly re-initialize serial gateway after configuration changes done in UI.
Fixes #13892.
Signed-off-by: Holger Friedrich <mail@holger-friedrich.de>
Mark Herwege [Sat, 10 Dec 2022 22:32:25 +0000 (23:32 +0100)]
[systeminfo] QuantityTypes and state descriptions cleanup (#13804)
* State descriptions cleanup
* Converted channels to QuantityType, adjusted default translations
* Channel definitions and percent to QuanityType
* Changed default state descriptions from MB to MiB
Signed-off-by: Mark Herwege <mark.herwege@telenet.be>
Gerd Zanker [Sat, 10 Dec 2022 10:18:13 +0000 (11:18 +0100)]
[boschshc] Support smoke detector (#13760)
* Add smoke detector service and add it to twinguard handler
* Add handler for smoke detector
* Support for smoke detector
added smoke detector device with SmokeDetectorCheckService
added SmokeDetectorCheckService to TwinguardHandler
added tests for smoke detector code
updated smoke detector code to latest boschshc version after cherry-picks of initial code from Christian Oeing
* re-generate i18n file, refactoring of smoke detector code to use abstract base classes
* Fix typos
* Add unit test for PlayPauseType commands
* Add unit test for SmokeDetectorCheckState
* Re-add null annotation
* Fix warning
Signed-off-by: Christian Oeing <christian.oeing@slashgames.org> Signed-off-by: Gerd Zanker <gerd.zanker@web.de> Signed-off-by: David Pace <dev@davidpace.de>
The I/O Linc has a feature to match or invert the status of the contact with its control messages. The binding was written expecting the messages to be inverted. This is the opposite of how insteon recommends setting up the garage kit, and not described well by the original "Note" at the bottom of the section. Updated the instructions to specify the input is OFF when linking, and changed the note to better describe what happens if you don't. All based off of my own experience with the sensor: https://community.openhab.org/t/insteon-io-linc-garage-door-contact-slow-update/141469
Jørgen Austvik [Thu, 8 Dec 2022 20:00:11 +0000 (21:00 +0100)]
[Nanolaef] Visual State Bugfix (#13880)
* Nanoleaf Visual State fix
Fix the visual state channel name.
Also:
- Better name (from state to visual state) of the (new) channel
- Better logs which has helped us debug the problem
- Some more information on when it will work in the README
Cody Cutrer [Thu, 8 Dec 2022 12:52:30 +0000 (05:52 -0700)]
[homekit] persist all known accessories, to prevent loss of homekit information (#13484)
* [homekit] persist all known accessories, to prevent loss of homekit information
See the readme for more details, but basically this keeps track of every
accessory we've ever created, and if it no longer exists, presents a dummy
accessory instead. If the accessory comes back, nothing is lost in the Home
app; if you meant to prune it permanently, you have to run a console command.
there are also several fixes to prevent presenting the device with missing data -
such as when the bundle stops, _don't_ explicitly remove the accessories until
the server has stopped.
we also don't increment the configuration version unless the configuration
has _actually_ changed (so removing and re-adding the exact same thing
won't trigger the device to reconnect). this even works across restarts
of the bundle, because we're persisting all the accessory information
for dummy information anyway.
* [homekit] Address review comments for accessory persistence
* update hap-java to 2.0.4
* remove unused local variable
* [homekit] ensure accessories are replaced in a batch
so that HAP-Java can maintain subscriptions with the new objects
* [homekit] log individual dummy accessories up to 5
* [homekit] Tweak readme for dummy accessories slightly.
* Be consistent with proper usage of useDummyAccessories.
* Make the sentence more clear about the effects of having dummy accessories.
In recent PR #13824, I added closing the context when the engine is closed, but core seems to have problems with that.
It logs: "[ERROR] [ipt.internal.ScriptEngineManagerImpl] - Error removing ScriptEngine
java.lang.IllegalStateException: The Context is already closed."
Sami Salonen [Tue, 6 Dec 2022 12:02:45 +0000 (14:02 +0200)]
[mqtt] Avoid parallel streams with common thread pool to avoid deadlocks (#13621)
To mitigate issue https://github.com/openhab/openhab-core/issues/3125
(common thread pool exhaustion when combining parallel streams with
synchronization or locks)
Florian Hotze [Mon, 5 Dec 2022 17:40:15 +0000 (18:40 +0100)]
[jsscripting] Add an alternative MIME type/alias (#13851)
The alternative MIME type text/javascript is also used by Nashorn, therefore we add the same extension as for the main MIME type.
The alias is introduced to simplify the usage in SCRIPT transformation profiles.
* Fix syntax problems in the bndtools file
* Add null check to avoid potential NPE
In some cases during testing, a NPE occured, this should solve it.
Cody Cutrer [Mon, 5 Dec 2022 16:19:35 +0000 (09:19 -0700)]
[bondhome] Initial contribution (#13459)
* First commit on newly created branch, taking code from c8b8e210dfd23f98526763782eadbca49509baf9
* [bondhome] update snapshot version, and some typos
* [bondhome] Address (most) comments from prior review from #7260
* [bondhome] simplify channels
* lastUpdate is unnecessary; turn on persistence or add a rule on update if
you care to keep track of it
* use a single string command channel for all shoot-and-forget commands, instead
of multiple switch channels
* use a rollershutter channel for shades (accepting UP, DOWN, STOP, 0%, and 100%)
* on all dimmer channels, accept ON and OFF, as well as 0% to imply OFF, instead
of having to write rules to control ON/OFF state separately.
* if the dimmer channel exists, prune the corresponding power channel, since
the dimmer channel is now a pure superset of its functionality
* overload fan#speed to be ceiling fan or a fireplace's fan, depending on the
device type
* [bondhome] add bundle to the BOM pom
* [bondhome] clean up BondDeviceHandler a bit
* there's no need to delay initialization; ThingManager won't
even attempt to initialize a child thing until its bridge is online
* Remove some extra initialization checks that can never be false
* slightly refactor some methods to return early, rather than
nest a giant `else`
* remove some info logging that will get triggered in normal usage
* [bondhome] fix bridge discovery
* Bridge property and config serial number need to be the same name
* Don't arbitrarily delay the BPUPListener
* Automatically update the IP if the BPUPListener finds it
* Provide the new bridge with its discovered IP to avoid an additional
DNS query
* Don't get the bridge version after every keep-alive response
* [bondhome] trigger end-device discovery as soon as the bridge comes online
* [bondhome] remove internal binding version
* [bondhome] change addr property to string
Certain values seen in the wild when interpreted as a long are too big for that
storage. Also, the Bond API documentation describes the addr property on a
device to be a string.
OpenHAB already has infrastructure to have things update their
channel definitions when a binding is updated.
* [bondhome] ignore any device that starts with _
In v3 of their API, Bond added a new special entry of __. Because no valid
device id would start with an underscore, ignore everything that starts with an
underscore to fix v3 and maybe futureproof.
* address review comments
mostly adding i18n to error states, and cleaning up error handling of
HTTP requests.
* use builtin translation services
instead of plumbing our own provider through
* use System.nanoTime instead of currentTimeMillis
so that it will be a monotonic clock, not (as) susceptible to the clock
changing
* [bondhome] ignore BPUP messasges that aren't state
In recent firmware, bond is now sending action messages via BPUP as well as state.
This change ignores all messages that aren't state.
* [bondhome] Improve error handling, and remove dummy constants
Just use a single BondException class to communicate any sort of
error from within bond, and avoid throwing, catching, and re-throwing
the same (or slightly modified) exception.
Also remove dummy constants that might give the wrong impression
of the details of your Bond device. Then implement proper null checks,
especially setting a configuration error if key config properties
aren't set on the thing.
* [bondhome] avoid setting device status when bridge just went offline
* address static analysis tool problems
Also-by: Sara Damiano <sdamiano@stroudcenter.org> Also-by: Keith T. Garner <kgarner@kgarner.com> Signed-off-by: Cody Cutrer <cody@cutrer.us>
openhab-bot [Mon, 5 Dec 2022 14:56:34 +0000 (15:56 +0100)]
New Crowdin updates (#13848)
* New translations miele.properties (Danish)
* New translations jruby.properties (Italian)
* New translations ecowatt.properties (Italian)
* New translations webexteams.properties (Italian)
M Valla [Mon, 5 Dec 2022 11:57:25 +0000 (12:57 +0100)]
[openwebnet] Add support for Alarm (WHO=5) (#13694)
* [openwebnet] first changes for Alarm support
* [openwebnet] added first version Alarm handler
* [openwebnet] added ownIdTest for Alarm
* [openwebnet] added things def for Alarm and discovery for alarm zone
* [openwebnet] refreshDevice
* [openwebnet] fix text formatting
* [openwebnet] fixes
* [openwebnet] added other alarm channels
* [openwebnet] Alarm: codestyle and null checks
* [openwebnet] updated Alarm examples in README and some cleanup
* [openwebnet] changed zone channel from armed to state, cleaned README
Markus Michels [Mon, 5 Dec 2022 07:51:31 +0000 (08:51 +0100)]
[shelly] Plus 1/1PM Add-On support; fix roller pos on open/close state (#13520)
* Added #13493 Plus 1/1PM addon support; Fixed #13515 pos for roller in
open/close state; various null warnings removed; README updated
Signed-off-by: Markus Michels <markus7017@gmail.com>
* digitalInput, analogInput and voltage added for Add-On; README updated
Signed-off-by: Markus Michels <markus7017@gmail.com>
* shelly2-roller re-added (wrong assumption that Shelly 2 doesn't support
roller mode, because it has no power meter)
Signed-off-by: markus7017 <markus7017@gmail.com> Signed-off-by: Markus Michels <markus7017@gmail.com>
* Initialization for Californium 2.7.3 fixed (no multicast packets where
received)
Signed-off-by: Markus Michels <markus7017@gmail.com>
* Analogous input support for Plus Addon changed; Allterco made a format
change in the beta firmware
Signed-off-by: Markus Michels <markus7017@gmail.com>
* changes applied
Signed-off-by: Markus Michels <markus7017@gmail.com>
* README updated; Shelly Plus addon handling adjusted to latest firmware
(before was beta)
Signed-off-by: Markus Michels <markus7017@gmail.com>
* Update bundles/org.openhab.binding.shelly/README.md
Signed-off-by: Fabian Wolter <github@fabian-wolter.de>
* fix JSON for Shelly1 - ext_switch_enable is a boolean, not an int
Signed-off-by: Markus Michels <markus7017@gmail.com> Signed-off-by: Markus Michels <markus7017@gmail.com> Signed-off-by: markus7017 <markus7017@gmail.com> Signed-off-by: Fabian Wolter <github@fabian-wolter.de> Co-authored-by: Fabian Wolter <github@fabian-wolter.de>
watchers had to be refactored similar to jsscripting.
it supports watching any directory referenced from RUBYLIB,
as well as the gem home. it properly excludes lib and gem home
(as well as other gem homes if you have multiple jruby versions
installed) from loading as regular scripts.
this is a breaking change if you don't have RUBYLIB explicitly
configured, and you are using the old default directory.
it's expected that the detection of what files and gems any
given script uses will be self-identified by the script, presumably
by the helper library.
JRubyScriptEngineConfiguration was largely refactored as part of this.
* CONFIGURATION_PARAMETERS was renamed, and is no longer static, since
it's modified every time the configuration is changed
* OptionalConfigurationElement was simplified since default values
are always provided now. this also simplified lots of other code
that accesses the current settings.
maniac103 [Sat, 3 Dec 2022 12:29:52 +0000 (13:29 +0100)]
[homematic] Fix min/max values for rollershutters (#13821)
* [homematic] Fix min/max values for rollershutters
For dimmers, the 1.0 max value sent by CCU was already converted to
percent values in the item state description. Do the same thing also for
roller shutters.
Signed-off-by: Danny Baumann <dannybaumann@web.de>
Florian Hotze [Fri, 2 Dec 2022 23:17:39 +0000 (00:17 +0100)]
[jsscripting] Fix memory leak that crashes openHAB (#13824)
* [jsscripting] Fix memory-leak caused by com.oracle.truffle.host.HostObject
Fixes this memory leak by making the HostAccess for the GraalJSScriptEngine available in a static final variable instead of creating it for each new engine.
Solution proposed in https://github.com/oracle/graaljs/issues/121#issuecomment-690179954.
Sharing a single engine across all Contexts (as proposed in https://github.com/oracle/graaljs/issues/121#issuecomment-880056648) is not possible, because core expects a ScriptEngine.
Daniel Demus [Thu, 1 Dec 2022 21:47:11 +0000 (22:47 +0100)]
[wundergroundupdatereceiver] Bugfixes: Regenerate trigger channel with proper type and more metadata. Normalize channel names. It might be easiest to delete and allow recreation of channels. (#13327)
* [wundergroundupdatereceiver] LAST_QUERY parameter should not be mapped automatically
* [wundergroundupdatereceiver] All channeltype props need to be applied
Especially the channel kind
* [wundergroundupdatereceiver] Remove illegal characters from channel name
Additionally expand the channel naming test to assert the generated channelUID and test that _ in names isn't inadvertently replaced
* [wundergroundupdatereceiver] Don't default AutoUpdatePolicy on creation
* [wundergroundupdatereceiver] Migrate changed channel to trigger type
Signed-off-by: Daniel Demus <daniel-github@demus.dk>
projects / openhab-addons.git / log