896825 - "Fix" bug 896776 in a way that is safe for leo

Status: RESOLVED FIXED

(Firefox OS Graveyard :: Gaia::System, defect)

Description

[Kyle Huey (Exited; not receiving bugmail, old account, do not use)]

+++ This bug was initially created as a clone of Bug #896776 +++

What I really want to do in bug 896776 is probably too invasive for leo.  So let's do something simpler.

Comment 1

[Kyle Huey (Exited; not receiving bugmail, old account, do not use)]

Attachment: Pointer to Github pull request: https://github.com/mozilla-b2g/gaia/pull/11107

Pointer to Github pull-request

Comment 2

[Justin Lebar (not reading bugmail)]

Blocks a blocker.

Comment 3

[(inactive after 6/18) Alive Kuo [:alive]]

As I said I'm working on a new system to resolve this kind of problems.
I want to know:
* The problem is closeCallback isn't reset to null or isn't invoked?
* How if we have a static closeCallback for each frame and won't modify that callback after transition?

Comment 4

[Kyle Huey (Exited; not receiving bugmail, old account, do not use)]

(In reply to Alive Kuo [:alive][Paris work week 7/22-26] from comment #3)
> As I said I'm working on a new system to resolve this kind of problems.
> I want to know:
> * The problem is closeCallback isn't reset to null or isn't invoked?

It is not invoked.  closeCallback calls removeFrame, and if it is never invoked the iframe is never removed from the DOM.

> * How if we have a static closeCallback for each frame and won't modify that
> callback after transition?

We need to ensure that we invoke the callback.  That is the real problem here.

Comment 5

[(inactive after 6/18) Alive Kuo [:alive]]

Yeah I know this problem...sigh it's not totally resolved by some workaround. So I'm working a brand-new transition control to fix all this kind of problems. Thanks for investigation.

Comment 6

[Justin Lebar (not reading bugmail)]

> How about if we have a static closeCallback for each frame and won't modify that callback 
> after transition?

That could work, but you would have to be careful not to leak closeCallback, because closeCallback would probably keep the iframe alive.

Comment 7

[Justin Lebar (not reading bugmail)]

In bug 896776 Alive says he has a plan to change a lot of code so as to fix this on master.  That's great, but then I think we should land this on master as well as b2g18, so both trees are leak-free.

Comment 8

[(inactive after 6/18) Alive Kuo [:alive]]

(In reply to Justin Lebar [:jlebar] from comment #7)
> In bug 896776 Alive says he has a plan to change a lot of code so as to fix
> this on master.  That's great, but then I think we should land this on
> master as well as b2g18, so both trees are leak-free.

Sure!

Comment 9

[[:fabrice] Fabrice Desré]

Pushed to master: https://github.com/mozilla-b2g/gaia/commit/3d64d2aefaa80919d183c95c2fd95a94f2b7702a

Comment 10

[John Ford [:jhford] CET/CEST Berlin Time]

Uplifted 3d64d2aefaa80919d183c95c2fd95a94f2b7702a to:
v1-train: 25c7466682f1148da685abc1f2177a3221290d84

Comment 11

[Tim Guan-tin Chien [:timdream] (please needinfo)]

v1.1.0hd: 25c7466682f1148da685abc1f2177a3221290d84