multiple {{}} interpolations inside <option> are causing value created by ngValue to be replaced by <option>'s text

[fxck]

See http://plnkr.co/edit/MXxLZ5wkVj4Cq2e5qQ4e?p=preview

works fine

    <select ng-model="baz.val">
      <option 
        ng-value="foo.val" 
        ng-repeat="foo in bar">
        {{ foo.name + ' ' + foo.second_name }}
      </option>
    </select> 

ng-value replaced in this case

    <select ng-model="baz.val">
      <option 
        ng-value="foo.val" 
        ng-repeat="foo in bar">
        {{ foo.name }} {{ foo.second_name }}
      </option>
    </select> 

also it's happening only if you use ng-value, it works fine when you use value="{{ val }}", see #9840

[Narretz]

Funny, just today I stumbled upon https://code.angularjs.org/snapshot/docs/api/ng/directive/ngBindTemplate which is made for that exact purpose. I wonder if we should add this info to select documentation. Might get lost though, because it's crowed there.

[caitp]

Yes, lets get a docs fixup for this. PRs welcome.

[caitp]

(note: we are still not fixing the original bug --- but it would be good to update the documentation to be easier on people, as mentioned)

[awerlang]

This behavior is explained by looking at https://github.com/angular/angular.js/blob/master/src/ng/rootScope.js#L445:

if (watchExpressions.length === 1) {
  // Special case size of one
  return this.$watch(watchExpressions[0], function watchGroupAction(value, oldValue, scope) {
      // omitted for brevity
  });
}

forEach(watchExpressions, function(expr, i) {
  var unwatchFn = self.$watch(expr, function watchGroupSubAction(value, oldValue) {
      // omitted for brevity
      self.$evalAsync(watchGroupAction);
  });
  deregisterFns.push(unwatchFn);
});

Whenever there's more that one expression to evaluate, because of $evalAsync() that'll be done some time later...later than assignment at option at least.

But...why does https://github.com/angular/angular.js/blob/master/src/ng/directive/select.js#L297 overwrites the value for option, when it already has one? I could trace this assignment back to 904b69c.

So, I believe a correction may be to verify if an option there's a value already.