* Backport #2443

* OnOffServer adaptions for OO/2.3 tests (#2335)

* Ensure state.offWaitTime is updated periodically

Test Harness is validating this property

* state.offWaitTime should not be reset in 'off'

Test Harness does not like it 😎

* Enable OO/2.3 tests

* OO/2.3 step 15 fix

(cherry picked from commit 484abe6)

* Backport #2335

* add basic CI tests for 0.15.x branch

---------

Co-authored-by: Richard Sulzenbacher <rsulzenbacher@users.noreply.github.com>

* backport #2322

(cherry picked from commit b5a8540)

* Add safeguard for invalid Dns Answers (#2454)

While we have no clue why this can happen at all it seem to happen that an answer is somehow collected that has an undefined recordtype and also no Byte data which leads to issues on encodeRecord.
See ioBroker/ioBroker.matter#539

(cherry picked from commit 2a51092)

* Fix guards in CommissioningController.resetStorage() (#2455)

fixes #2451
... and remove some redundant checks

(cherry picked from commit fb122c2)

* Enhance Pase session check (#2468)

... and make sure it does not considers pase sessions that are in closing.

(cherry picked from commit 82346dd)

* Fix ep change detection (#2486)

* Fix Endpoint change detection and logic in Legacy Controller

* Enhance storage to store only fabric scoped data and version only when wildcard, but automatically
* Add mechanism to cleanup stored data
* Fix endpoint change detection to also update, or remove endpoints correctly
* introduce new events for endpoint changes
* Allow to process reads with attribute change detection
* Do not send a Read when we update structure because we know we have updated values

* Formatting

(cherry picked from commit 08c9b11)

* changelog and log removal

* Process and store Read/Subscribe result chunked (#2488)

* Process and store Read/Subscribe result chunked

This is another optimization of the "legacy controller" to already bring future convenience to it. This change already process changes to data including storage to happen while data is received already.
This can allow reconnection to devices with bad connection because data might fill up over time. I was able to reconnect to a device that was else never finishing.

The new API will do that automatically, so lets use that already here.

The code is not too nice and optimized because potential backport candidate (depending on the conflicts)

* [execute-controllertests] review feedback

(cherry picked from commit cdcc097)

* [execute-controllertests] changelog and tests

* [execute-controllertests] tests

* [execute-controllertests] tests

* Also accept Non-Authoritative Mdns Responses (#2502)

* Also accept Non-Authoritative Mdns Responses

... including testing of it

* review comments

(cherry picked from commit bae8894)

# Conflicts:
#	packages/protocol/src/mdns/MdnsServer.ts
#	packages/protocol/test/mdns/MdnsTest.ts

* Fix the update of child endpoints in PairedNode (#2503)

Child were not properly updated when they were already existing

(cherry picked from commit 093eb9f)

* changelog

* Fix Assert dependencies check (#2509)

(cherry picked from commit 837cabf)

* Correctly store and use FabricIndex in Resumption records (#2521)

* Correctly store and use FabricIndex in Resumption records

The absence of the fabricIndex lead to wrong mappings when nodeIds were colliding for a commissioned device.
The new code is backward compatible and represent the former logic without fabricIndex, but work correctly with fabricIndex. Also logging is adjusted

* Update packages/protocol/src/session/SessionManager.ts

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
(cherry picked from commit 9e01b8b)

* Generate a random FabricId for new controllers (#2522)

To prevent Fabric Id collisions we now generate random fabric IDs when controllers are initialized the first time and no defined adminFabricId is defined.

(cherry picked from commit 3fe1e7c)

* changelog

* Fix for Logger returning on first higher log level (#2528)

(cherry picked from commit 809669d)

* changelog and chip test adjustment

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: David Masshardt <david@masshardt.ch>

* manually port #2544

* Remove exchange earlier

... no need to wait until we closed the session

(cherry picked from commit 11c5fd9)

* Add Exchange direction indicator

(cherry picked from commit f9bf5c0)

* Fix fabric scoped event read without payload (#2565)

When the Read was fabric scoped and the event had no payload (which e.g. happens on the shutDown event) then the event was skipped.
Fix for Backporting

(cherry picked from commit 65b7975)

* Fix fabric scoped event reads (#2572)

We were not correctly checking for the field and fabric sensitive usage. This is fixed now including comprehensive testing

(cherry picked from commit 2a00c54)

* Close OccurrenceManager on node shutdown (#2566)

* Close OccurrenceManager on node shutdown

To also ensure non-volatile stores gets closed properly.
The PR also fixes that the promise of close is returned correctly.

* Add comment for ServerNode and OccurrenceManager

Added comment regarding ServerNode usage and OccurrenceManager closure.

---------

Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>

(cherry picked from commit 1a719aa)

* Manual merge of Initialize controller into an own environment (#2567)

* Close OccurrenceManager on node shutdown

To also ensure non-volatile stores gets closed properly.
The PR also fixes that the promise of close is returned correctly.

* Add comment for ServerNode and OccurrenceManager

Added comment regarding ServerNode usage and OccurrenceManager closure.

---------

Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>

(cherry picked from commit 1a719aa)

* Ignore unexpected responses on invoke (#2568)

The Matter SDK seems to have a bug that also sends invoke responses when the response should be suppressed (not happening for group responses). So we add logic to ignore such unexpected responses, but still react on statusresponses with an error.

Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>

(cherry picked from commit cf23abc)

* Ensure that the PairedNode meta data are updated in storage (#2569)

We need to update the data in storage until we move away from PairedNode so that the details are available and not outdated.

Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
(cherry picked from commit 6ef4457)

* finalize and changelog

---------

Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>