Component controller

[drpicox]

See PR: #13732
See Issue: #13683

Now $componentController looks for the directive that represents a component which is the one that satisfies:

• it has at least restrict:'E' (other are valid if they contain 'E')
• it has controller
• it has controllerAs.

Tests are added to evaluate under multiple scenarios commented in the original thread.

PS: I created the new PR to solve conflicts with minimum "dirtiness".

[petebacondarwin]

I think we can just check === 'E' rather than indexOf('E')

[drpicox]

I tried to add minErr,

angular.mock.$ComponentControllerProvider = ['$compileProvider', function($compileProvider) {
  var minErr = angular.$$minErr('$componentController');
  return {
    $get: ['$controller','$injector', function($controller,$injector) {
      return function $componentController(componentName, locals, bindings, ident) {
        // get all directives associated to the component name
        var directives = $injector.get(componentName + 'Directive');
        // look for those directives that are components
        var candidateDirectives = directives.filter(function(directiveInfo) {
          // components have controller, controllerAs and restrict:'E'
          return directiveInfo.controller && directiveInfo.controllerAs && directiveInfo.restrict === 'E';
        });
        // check if valid directives found
        if (candidateDirectives.length === 0) {
          throw minErr('nocomp','No \'{0}\' component found', componentName);
        }
        if (candidateDirectives.length > 1) {
          throw minErr('multicomp','Found {0} directive candidates for component \'{1}\'', candidateDirectives.length, componentName);
        }
        // get the info of the component
        var directiveInfo = candidateDirectives[0];
        return $controller(directiveInfo.controller, locals, bindings, ident || directiveInfo.controllerAs);
      };
    }]
  };
}];

but I'm getting this error from grunt test:

error:   TypeError: Cannot read property 'nocomp' of undefined
    at /Users/drodenas/Projects/drpicox/angular.js/docs/config/processors/error-docs.js:43:76
    at Array.forEach (native)
    at Object.$process (/Users/drodenas/Projects/drpicox/angular.js/docs/config/processors/error-docs.js:18:12)
    at /Users/drodenas/Projects/drpicox/angular.js/node_modules/dgeni/lib/Dgeni.js:202:28
    at _fulfilled (/Users/drodenas/Projects/drpicox/angular.js/node_modules/dgeni/node_modules/q/q.js:798:54)
    at self.promiseDispatch.done (/Users/drodenas/Projects/drpicox/angular.js/node_modules/dgeni/node_modules/q/q.js:827:30)
    at Promise.promise.promiseDispatch (/Users/drodenas/Projects/drpicox/angular.js/node_modules/dgeni/node_modules/q/q.js:760:13)
    at /Users/drodenas/Projects/drpicox/angular.js/node_modules/dgeni/node_modules/q/q.js:821:14
    at flush (/Users/drodenas/Projects/drpicox/angular.js/node_modules/dgeni/node_modules/q/q.js:108:17)
    at doNTCallback0 (node.js:419:9)

I have added the directory $componentController with files multicomp.ngdoc and nocomp.ngdoc to docs/content/error.

I ran out of time. ¿Any clue how to solve it?

Thanks!

PS: I did not pushed back the changes because are breaking check passes, but I can push them if you ask for it.

[gkalpak]

Not sure what might be causing the issue. IF you push the changes (with minErr), even if broken, it will be easier to identify the problem (and the solution 😃).

[drpicox]

Thanks!
Done, breaking commit away.

[petebacondarwin]

I'll take a look now

[petebacondarwin]

The problem is that we were not minifying the angular-mocks.js file (unsurprisingly) and so the errors.json file did not contain the minErr error information, which is needed.

So looking at this again, I realise now that we don't use minErr in the rest of ngMock and so I was wrong to suggest that we use it here. Sorry.

[petebacondarwin]

I will review and merge the rest of the PR without this last commit

[drpicox]

No problem at all, I used throw new Error in the first approach because is what I found inside the angular-mock.js code.
Anyway it was nice to explore how minErr is working.

Thanks!