Calling _super in a CP has the wrong context when retrieved from an action

[offirgolan]

In an Ember.Object, when you call this._super in a CP, it will retrieve the parent's value of that same property. What was found is that when you fetch a CP that has a _super call in an action handler, the _super context is that of the parent's action and not of the parent's CP.

Ember Twiddle - Tested on Ember 1.13.13 to Canary

When you click the save button, the following should happen:

1. Grab validations object which also grabs inherited validations if possible
2. the my-component action should be called which at the end calls _super (defined in CRUD mixin)
3. The CRUD save action should then be called

What is actually happening:

1. Grab validations object
2. _super in validations objects calls the CRUD mixin save action instead of inherited validations cp
3. Do stuff in my-component save action as intended and then call _super (defined in CRUD mixin)
4. The CRUD save action should then be called

[offirgolan]

cc @stefanpenner @rwjblue

[stefanpenner]

Yup, seems like a bug.

@offirgolan if you have cycles a failing test would be sweet, im going to skim the code and see what may be going sideways.

[stefanpenner]

I believe the issue is, the cp has no super so it doesn't bother super wrapping it. But if it does not super wrap it, it doesn't reset the actions _super when entering the CP.

https://github.com/emberjs/ember.js/blob/master/packages/ember-metal/lib/mixin.js#L105-L107.

The above code should insert a no-op super wrapper if no super descriptor is found, rather then circuiting.

I believe this should be a pretty quick fix, but it is unlikely I will be able to look at it more closely until Sunday. If no-one else does, and if i forgot please ping me :P

[offirgolan]

👍 Sweet will do. I dont have too much free time unfortunately, but if I find a spare moment, ill whip up a test. Thanks for taking this on!

[rwjblue]

Submitted #13242 with a couple tests and fix. Needs review.

[offirgolan]

👌 thanks Rob

On Sunday, April 3, 2016, Robert Jackson notifications@github.com wrote:

Submitted #13242 #13242 with a
couple tests and fix. Needs review.

—
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub
#13230 (comment)

Thanks,
Offir Golan

[rwjblue]

See #13242 (comment) for an update. I'm going to close this for now, and track the underlying issue in #13242 instead.

Thank you for reporting, hopefully we can get something in place to help prevent this sort of trollage in the near future.