[BZ 5651900] Binders not getting destroyed when using composite.execute and calling mp.destroyChildren

[goonieiam]

Use case:

Mojits being created on demand via composite.execute(), then destroying them via mp.destroyChildren()

The first execution does the right thing, but upon doing the full cycle again(composite.execute then mp.destroyChildren()), the newly created child doesn't get destroyed.

The problem seems to be in destroyMojitProxy (mojito-client.client.js):

On a second run, this._mojits contains duplicated ids and this condition never evaluates to true, therefore never destroying it.

One thing I also noted is that this line :

this._mojits[id].proxy._destroy(retainNode);

creates a recursion, not sure if it has anything to do with the issue, but thought I'd mention.

if (this._mojits[id]) {
// TODO: activate call to unbindNode below:
// unbindNode(this._mojits[id].proxy._binder,
// this._mojits[id].handles);
this._mojits[id].proxy._destroy(retainNode);
delete this._mojits[id];
// We don't manage binder children automatically, but any time a
// new child is added or removed, we should at least give the
// application code a chance to stay up to date if they want to.
// The only gap is when a mojit destroys itself.
// onChildDestroyed is called whenever a binder is destroyed so
// any parents can be notified.
parent = findParentProxy(this._mojits, id);
if (parent && parent._binder.onChildDestroyed &&
Y.Lang.isFunction(parent._binder.onChildDestroyed)) {
parent._binder.onChildDestroyed({ id: id });
}
}

[rwaldura]

goonieiam: is this something you can provide a patch for? Do you know exactly what needs to be fixed?

[goonieiam]

I'd have to spend some time digging around, I don't have a clear solution I can provide. If anyone is able to point me in the right direction, I might be able to help.

[rwaldura]

this._mojits contains duplicated ids

Can you view the IDs? Do they look like YUI_3_4_1_blahblah or YUI_3_5_1_blahblah?

[rwaldura]

this._mojits contains duplicated ids

Can you view the IDs? Do they look like YUI_3_4_1_blahblah or YUI_3_5_1_blahblah?

[rwaldura]

BZ 5651900

[isao]

Big thanks to Gonzalo Cordero for finding this memory leak, AND identifying the offending code. The leak affects dynamic mojits, and has existed for quite some time. Pull #225 was submitted for mojito's mainline, develop.

Note I made two branches:

• mojit-leak off HEAD of develop which has currently has a broken mojito test app and uses YUI 3.5.1
• mojit-leak-036 back port on top of the last mojito 0.3.26 release