ractivejs / ractive-adaptors-backbone Goto Github PK

When having in a template like the following example using unescaped html triple mustache, it is not possible to get the the model updates after a user has change the input value.

<input type="text" value="{{{model.name}}}">

ractive.get('model.name')

It used to be that the event.context of a Ractive proxy event contained the backbone model. Now it just contains a plain object with the model properties. However, if I call this.get(event.keypath) then I do get a backbone model.

It seems to me that event.context should contain the backbone model and not just the properties.

It seems like there's only one assertion in the test suite:

https://github.com/ractivejs/ractive-adaptors-backbone/blob/master/test/tests.js

It'd be difficult to just merge PRs right now unless more tests are added.

Hi. This adaptor is pretty great and very helpful for our projects.

I am running into an issue where calling a .sort() on a collection does not prompt Ractive to re-render the HTML. The collection is indeed resorted, but this is not reflected in the DOM. Even with a manual ractive.update() call, no changes.

I will work on making a JSFiddle to make this reproducable, but i just wanted to start the thread to see if there was something obvious I was missing.

Ideally, I should be able to put require('ractive-backbone') in my app, but with a main property or an index.js, I need to instead use `require('ractive-backbone/Racktive-Backbone').

refs ractivejs/ractive#1392

Would be great to do after 0.2.0 once we've merged outstanding PRs.

Hi,

is there a reason to propagate a model change only for immediate child property ?

https://github.com/RactiveJS/Ractive-adaptors-Backbone/blob/master/src/Ractive-Backbone.js#L116

Thanks

Ractive has a ractive.merge() operation which minimises the amount of DOM manipulation involved in updating list sections. It should be possible to use the same logic with Backbone collections.

I'm seeing a blank page on: http://examples.ractivejs.org/backbone

Is it just me or is the demo broken?

jsfiddle example: http://jsfiddle.net/zw4Pc/2/

in the example i bind to each model's .on('change') as i add them but they never get called when editing in the inputs. my naive understanding thinks we should bind to ractive .set and .update in the backbone adaptor. what do you think?

as a side-note, how does one do two-way binding in a list? in the example if you make multiple items they all have the same values.

cheers!

ractive and backbone need to be under peerDependencies.

It might not work in browserify otherwise. x instanceof Backbone.Model may check x against a new instance of Backbone from the one the user's probably already using.

Hello, I have found a bug that seems to only surface in very rare instances. Under the reset method for models, this.value.reset( object ); is used. Since this.value is a Backbone model, calling reset() will most certainly fail. I fixed this by replacing the whole function with return false;. This might not be desirable, so instead perhaps this.value.clear().set(object); would be better?

it work in 0.4.0: http://jsbin.com/tovik/5/edit?html,js,console,output

but fail in 0.6.0: http://jsbin.com/tovik/4/edit?html,js,console,output

there is a way to make it work in this version: http://jsbin.com/tovik/6/edit?html,js,console,output

Grunt for a 180-line library may be overkill. Since moving from qunit to mocha, the only functions grunt fulfills today is (1) jshint, and (2) uglifyjs compression.

My suggestions:

• include jshint testing as part of the test suite. I think this is great because any open PRs will be tested against the test suite via Travis, hence checking PR's against jshint as well. This can be done via scripts: { test: "jshint ractive-adaptors-backbone.js && mocha" } or mocha-jshint.
• use npm run prepublish to minify the js file. The prepublish hook is the npm standard for invoking build steps. This is done by adding { scripts: { prepublish: "uglifyjs -m < index.js > index.out.js" } } to package.json.
• finally, remove dependency on grunt.

When a wrapped backbone collection contains a second backbone collection, the reset call of the ractive-adaptor blows the call stack :(

http://jsfiddle.net/va6jU/46/

Looks like the collection is wrapped twice. One time through misses the this.setting check and gets .reset. Which seems like it would be fine, except the this.value.reset( models ) triggers a 'reset' event, which triggers the collection wrapper to reset again (and again, and again).

Adding {silent: true} seems to fix this. That might be covering a symptom though.

Consider:

ractive.set( 'x', new Backbone.Model() );
ractive.set( 'x', new Backbone.Collection() );

This will treat the collection as a POJO and attempt to extend the model with it. That is:

// this doesn't make sense
model.set({ length: 0, models: [], _byId: {} });

There are some pros and cons to mergeing as opposed to seting as is currently done here. A con would be performance is better with a set, and a pro would be that elements would transition properly with merge.

Prior to 0.8, it was impossible to merge an array with itself, so it may not be wise to do so until 0.7 and earlier are mostly out of use. See ractivejs/ractive#2203