incubator-callback-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andrew Grieve <agri...@chromium.org>
Subject Re: [Android] Can we get rid of the Callback Server/Hanging GET now?
Date Thu, 13 Sep 2012 00:25:38 GMT
ONLINE_EVENTS and JS_OBJECT are the fastest and have no bugs that I've
found. As soon as 2.1 ships, let's make the switch. I don't think devs
should need to know about the bridge modes unless there becomes a reason to
expose this to them.

With several other options other than callback server, I think we should
get rid of it since it's a fair amount of code and complexity.


On Wed, Sep 12, 2012 at 3:41 PM, Filip Maj <fil@adobe.com> wrote:

> I would be in favor of dropping a deprecation-like notice and educating
> users about the differences.
>
> I would change the default bridge mode to the events one, say in 2.2 or
> 2.3. Then like 2.5 remove the callback server if we've gone through a
> couple release with no issues with the new bridge mode.
>
> My $0.02.
>
>
> On 9/12/12 12:38 PM, "Joe Bowser" <bowserj@gmail.com> wrote:
>
> >Hey
> >
> >In 2.1.0, we currently have the ability to use multiple bridges thanks
> >to Andrew's work.  However, we currently still have a series of issues
> >related to the fact that on Android 4.x, the routing tables decided to
> >take a vacation and never come back when there's no Internet
> >connection.  This means that the bridge freezes up and never comes
> >back.  This wouldn't be an issue if this wasn't our default bridge
> >method.  In addition to this, a large amount of memory usage on
> >Android is also taken up with this callback server.  So, I think we
> >should take this thing out behind the shed and put it out of its
> >misery.
> >
> >As far as what should replace it, I'm for the overriding of the online
> >event for replacing it, since it performs faster than the others, and
> >actually works across all the versions of Android based on what I've
> >tested so far (2.2.2 to 4.1.1).
> >
> >Any thoughts or reasons why this method should survive?
> >
> >Joe
>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message