Changes how object mutation works behind the scenes in order to keep objects in sync automatically when their model changes.
* The way that objects are mutated and observed has not changed, openmct.objects.mutate and openmct.objects.observe should still be used in the same way that they were before.
* Behind the scenes, domain objects that are mutable are wrapped in a new MutableDomainObject that exposes mutator and observer functions that allow objects to be mutated in such a way that all instances can be kept in sync.
* It is now possible to retrieve MutableDomainObjects from the API, instead of regular domain objects. These are automatically updated when mutation occurs on any instance of the object, replacing the need for "*" listeners. Note that the view API now provides objects in this form by default. Therefore, you do not need to do anything differently in views, the domain objects will just magically keep themselves up to date.
* If for some reason you need to retrieve an object manually via openmct.objects.get (you should ask why you need to do this) and you want it to magically keep itself in sync, there is a new API function named openmct.objects.getMutable(identifier). Note that if you do this you will be responsible for the object's lifecycle. It relies on listeners which must be destroyed when the object is no longer needed, otherwise memory leaks will occur. You can destroy a MutableDomainObject and its (internal) listeners by calling openmct.objects.destroyMutable(mutableDomainObject). Any listeners created by calls to openmct.objects.observe need to be cleaned up separately.
* If the composition of a MutableDomainObject is retrieved using the Composition API, all children will be returned as MutableDomainObjects automatically. Their lifecycle will be managed automatically, and is tied to the lifecycle of the parent.
Any MutableDomainObject provided by the Open MCT framework itself (eg. provided to view providers by the View API, or from the composition API) will have its lifecycle managed by Open MCT, you don't need to worry destroying it.
* Implement 'save' method in Object API
* Refactor legacy persistence code to work with new save object API
* Added 'isPersistable' check to object API
* Fixed incompatibility between object API changes and composition policies
* Make save method private
Co-authored-by: Deep Tailor <deep.j.tailor@nasa.gov>
* Move to webpack build
* Use webpack for building openmct. Move SCSS to one folder and load
all core css up front. Remove bower, begin removing gulp in favor
of npm run.
* Uses eslint instead of jshint and jscs. Merge style checking rules
into .eshintrc.js, carrying over core parts of crockford style and
our adaptations. Current code base fails to pass the linter, want
to separate linter changes from fixes to linting rules.
* Support for Vue SFC with example
* Remove outdated examples
* Use HTML loader for html (supports relative imports of resources e.g.
images) and raw-loader for when javascript must be loaded as text.
* Updated Karma and Jasmine versions
* Added DOMObserver class. Supports promise-based testing of DOM changes
Update asynchronous test specs to use promises or done() instead of waitsFor/runs
* Modified ActionCapability to duplicate context object properties as own properties for better object equality comparisons
* Global find + replace to fix syntax issues
* Fixed various issues caused by non-deterministic runtime order of tests in Jasmine 3. Fixed issues caused by changes to determination of object equality
* Addressed review comments
* Resolved merge conflicts with master
* Fixed style errors
* Use spy.calls.count() instead of manually tracking
Stub in specs for policies added for Edit mode which control
the availability of the Edit and/or Edit Properties actions
in the view-control area, WTD-1062.
