"in" does not validate properly

[donaldducky]

I have a model with validations like

type:
  in: ['a','b']

When I save with

model.save({ type: 'a' })

The 'in' validator doesn't get passed valueToSet properly:

  "in" : function(whitelist, attributeName, model, valueToSet) {
    console.log(whitelist, attributeName, model, valueToSet);
    return _.include(whitelist, valueToSet) ? undefined : "in";
  },

Gives the log output:

["a", "b"]
"type" 
d
 undefined

valueToSet is undefined somewhere along the way.

My quick fix, for now:

  "in" : function(whitelist, attributeName, model, valueToSet) {
    valueToSet = model.get(attributeName);
    return _.include(whitelist, valueToSet) ? undefined : "in";
  },

[donaldducky]

Nevermind! I was using it wrong.

When I had it passing nothing into save:

model.save({ }, { ... });

I guess validations aren't run on what's already set, only what's passed in on save?

[nealstewart]

If you set it with {silent: true} the validations won't be run.

[nealstewart]

Can you post more code? I'm curious about what you're doing.

[donaldducky]

I had something setup like this:

var MyModel = Backbone.Model.extend({
  url: 'test',
  validate: {
    type: {
      in : [ 'a', 'b' ]
    }
  }
});

var b = new MyModel({ type: 'b' });
b.save({}, {
  success: function(model, response){
    console.log('success');
  },
  error: function(model, response){
    console.log('error', model, response);
  }
});

I was expecting it to pass validation since the type is "in" the list.

Setting it up like this triggers the save call to the server.

var MyModel = Backbone.Model.extend({
  url: 'test',
  validate: function(attributes) {
    if (_.include(['a','b'], attributes.type)) {
      return 'no a';
    }
  }
});

var a = new MyModel({ type: 'b' });
a.save({}, {
  success: function(model, response){
    console.log('success');
  },
  error: function(model, response){
    console.log('error', model, response);
  }
});

So, when using the validations plugin, it doesn't seem to trigger the same way as backbone's, in this case.

[donaldducky]

Let me know if you need more info!

[nealstewart]

Cool. That's bad, so I'm going to reopen this.

[nealstewart]

Closing this issue in favour of the exact problem at #12