Long strings store their length in 2 bytes at the front of the string, with the
bytes being a little-endian representation of the string length.

But the endpoint config output was spitting out the length as big-endian, not
little-endian.  This caused things to be pretty confused and led to buffer
overruns trying to read data that was not there.

0. Add basic helper categorization.
1. Externalize handlebars instance.
2. Desingletonize the handlebars instance.
3. All handlebars instances are now taken from single location.

1) Remove struct_is_fabric_scoped and event_is_fabric_scoped now that we have
   isFabricScoped/isFabricSensitive on those objects.  Add the relevant XML
   attributes instead of trying to guess based on fields.
2) Remove event_fabric_idx_field and struct_fabric_idx_field, since we are no
   longer guessing based on fields.
3) Synthesize the relevant struct item or event field with the right name, type,
   field id, per spec, if the struct is fabric-scoped or the event is
   fabric-sensitive.

This is a hard requirement for implementing Leave events properly for Matter:
those have a fabric_idx field but are _not_ fabric-scoped.

Minor workflow change.

* Fix the importer to properly assign packages. Earlier, readFile() methods didn't even init packages.
* Separate the logic to initialize session key values from package defaults.
* Fix conversion process which suffered from double package init.
* Clean up unit tests to eliminate double session initialization.

This commit adds the Cypress tests using Matter test data as well. Before this, we were only running tests against ZigbeePro data, so occasionally UI got broken for Matter and it was not caught, because it was specific to differences in Matter and ZIgbeePro.

* update tests and add matter data
* update check_summary tests

* Add fabric-scoped flags to schema.
* Add isFabricScoped to command loading
* Add fabric scope to command queries.
* Add fabric sensitivity to event queries.
* Add isFabricScoped to structs in XML, and load them correctly.
* Update queries with the fabric-scoped field for structs.
* Update schema diagram.
* Set events to being fabric-sensitive.

* download from SiliconLabs fork by default
* enable Github token to be passed in via --githubToken

0. Add the workflow logic.
1. Remove some old muddle around project.env.DEV

github: ZAP#309

github: ZAP#309

Adding string length validation check to the zap UI to warn the user when the values exceed the allowed length of the string
Github: ZAP#379

* Share the ipc socket same way as single-instance shares sockets.

* Some code cleanup.

* Increase feature level to trigger IDE tests.

* Make server load both zigbee and matter data.

* Fix the hang of the generation in the client/server case.

* Remove hardcoded background.

* Export events into the zap file.

* Do not include empty array keys in the zap file.

* Create skeleton function for event importing.

* Fix unit tests.

* Cleanup the attribute query.

* Import event data from zap file into the tables.

* Add unit test that validates that events are loaded.

* Return the event state in the session back to the UI as part of the state creation.

* Work the way up the UI to show event data.

* Properly display toggles based on what is loaded in the database.

* Fix the package-lock.json that was broken in an earlier faulty merge.

Add the end-of-run cleanup action to set package.json back to 0.0.0

https://github.com/project-chip/zap/issues/582

0. Add sparse enum to the unit test data.
1. Add helpers to generate the spare enum.
2. Add unit test for the "first_unused_enum_value" both modes.

create conversion result file during Studio upgrade process for reporting results

BUG: MCUDT-28898

* Automatically stamp versions at npm install time, to be ready for concurrency.

* Peg package-lock.json to the fake version as well.

* Cleanup the naming of axios global accessor functions.
* Add license checker to workflow.

* Add unit test for event field type checking.

Returning from insertDataType just like the other insert functions
Adding cluster id since it is being used by our code.
Removing dbcache usage since it seems to slow down the generation based on the following statistics
Upon generating z3Light app with dbcache: Average time taken across 5 runs: 24.8812
Upon generating z3Light app without dbcache: Average time taken across 5 runs: 19.272
These changes are saving around 5s locally
Adding the command argument count to the command queries to avoid additional sql calls for finding out the count within a block helper to reduce the complexity of the template
Bumping the feature level for zap
JIRA: EMZIGBEE-9385

* Update versioning scheme to on-the-fly inject a correct version for build.

* Add npm targets for version fake/real stamping.

* Update version-stamp with creating a real version in package.json

* Move a package.json logic to a function in script-util.

* Update printouts.

* When building, use a timestamp of a last git commit, NOT the current time.

* Export category and enable zapmatter target property.
* Add description to the zap file, so users can be presented with a nice dialog.

* Clean up some package queries and encapsulate db.

* Fix the loading of templates and zcl packages to include category and description.

* Increase feature level.

* Update generation templates handling with the new columns.

* Update preview and add the template name to the db.

* Update expected electron version to what we actually have.

JSON.stringify does not handle TypeError object

* https://github.com/project-chip/zap/issues/503

* https://github.com/project-chip/zap/issues/505

* https://github.com/project-chip/zap/issues/507

* https://github.com/project-chip/zap/issues/519

* adding ubuntu latest to enviornments of github action jobs

* Minor cleanup.

* Clean up some session creation in tests.

* Configure more lenient timeouts for github actions.

For strings especially, consumers really do want to know what size to
expect.

Adding helpers for zcl cli commands such that the right content is generated for Zigbee and other cleanup.