I was going to say that you wrote this, but then I remembered that Git
does that thing with files you move around where it looks like you own
the code. Can we go back to when this was added and ask whoever wrote
this what they were trying to do? I'm just guessing at this point.
On Mon, Feb 24, 2014 at 5:32 PM, Andrew Grieve <agri...@chromium.org> wrote:
> If that's the intention of the code, then I think we could probably change
> it since...
>
> When you navigate via HTML (e.g. click a link or set window.location), we
> don't reset the plugin manager, instead, this triggers the onReset() method
> on all plugins.
>
> pluginManager.init() is called *only* when you navigate via Java.
>
>
> On Mon, Feb 24, 2014 at 7:19 PM, Joe Bowser <bows...@gmail.com> wrote:
>
>> Right, this calls pause and destroy on all the existing plugins, and
>> clears the objects until they're called again.
>>
>> So, for example. Let's say you had the Camera plugin, which is
>> probably our nastiest plugin w.r.t memory usage, and you took a
>> picture. Then you wish to go back to what is basically a HTML app
>> that doesn't have any need for any plugins. It doesn't make sense for
>> that Camera plugin to keep holding onto the memory if you're
>> navigating to another URI where that plugin isn't needed. So, we
>> destroy the plugins that were instantiated and wait for that document
>> to make a call to another plugin.
>>
>> I think this behaviour does make sense, although I'm worried that we
>> may be missing a delete/destroy somewhere.
>>
>> On Mon, Feb 24, 2014 at 4:04 PM, Naik, Archana <na...@lab126.com> wrote:
>> > Oh I see.
>> >
>> > Here, by recreating I meant calling pluginManager->init() method.
>> > loadUrl(string url) in CordovaWebView calls loadUrlIntoView(final String
>> > url, boolean recreatePlugins)
>> > With second argument as true. Which does this
>> >
>> > if (recreatePlugins) {
>> > this.url = url;
>> > this.pluginManager.init();
>> > }
>> >
>> >
>> > I am wondering why calling init() here every time we load a new url.
>> >
>> > Thanks
>> > Archana
>> >
>> >
>> > On 2/24/14 3:57 PM, "Joe Bowser" <bows...@gmail.com> wrote:
>> >
>> >>On Mon, Feb 24, 2014 at 3:45 PM, Naik, Archana <na...@lab126.com> wrote:
>> >>> History stack reset? I thought loading url will add to the navigation
>> >>> history.
>> >>>
>> >>
>> >>Which history are we referring to? We have some old legacy methods
>> >>from the bad old days when we maintained our own history, because we
>> >>thought the browser history was broken (Android 3.x, 4.0.x).
>> >>
>> >>> Yes, overload helps to by pass this recreation but default call is with
>> >>> this flag set to true so internally when you use loadUrl() it will
>> >>> recreate plugins.
>> >>
>> >>We should only create plugins when we invoke them unless we're setting
>> >>plugins to be instantiated onload. There also could be some security
>> >>reasons to destroy and recreate the plugins, although none are coming
>> >>to mind now.
>> >>
>> >>
>> >>
>> >>>
>> >>> Archana
>> >>>
>> >>> On 2/24/14 11:13 AM, "Andrew Grieve" <agri...@chromium.org> wrote:
>> >>>
>> >>>>I think the history stack is reset when you use that API, so it
>> somewhat
>> >>>>does make sense to recreate the plugins. Not sure if there is a better
>> >>>>answer than that...
>> >>>>
>> >>>>I added the overload to allow not resetting the plugins, because I
>> think
>> >>>>that is a useful thing to want to do as well.
>> >>>>
>> >>>>
>> >>>>On Fri, Feb 21, 2014 at 5:24 PM, Naik, Archana <na...@lab126.com>
>> wrote:
>> >>>>
>> >>>>> Hi, Devs,
>> >>>>>
>> >>>>> Why do we recreate plugin every time an url is loaded? I am referring
>> >>>>>to
>> >>>>> loadUrlIntoView(url,bool) method.
>> >>>>> Other override which take only string(url), has this recreatePlugins
>> >>>>> boolean as true.
>> >>>>>
>> >>>>> Thanks
>> >>>>> Archana
>> >>>>>
>> >>>
>> >
>>
