[rfc][icedtea-web] bringing applets out of browser (part2)

[Jiri Vanek]

On 12/11/2014 09:38 AM, helpcrypto helpcrypto wrote:
> May I know what is all this about?
>
> Remeber we are using an Applet, which uses JSObject to send/receive info from WebPage
> We know Chrome will dump NPAPI next april/september, but -imho- no changes are needed for
> firefox/safari.
>

Yes, I know. Although I wont to implement some JSObject stubs, the *serious* applets will not 
work/or will be useless without browser. This initiative is for simple applets or java games, 
moreover legacy ones which are applets instead of javaws or just because it was cool to create applets.

THis is not intended to replace npapi or bring some extraordynary feature. Nor even affect existing 
ones. It is simplification of one usecase.


J.
>
> Regards.
>
>
>
> On Tue, Dec 9, 2014 at 3:04 PM, Jiri Vanek <jvanek at redhat.com <mailto:jvanek at redhat.com>> wrote:
>
>     This is first implementation of -html.
>
>     Well more easy then expected, but not working as expected... So this is moreover preview and
>     work is still in progess (any hint to current impl welcomed!)
>
>     Well it do what its description says. The param is html page, it finds an applet on it, and then
>     lunch it... So some shortcut to already implemented shortcut, but it was not intended to be.
>
>     TRight now it have many flaws:
>             1) generate jnlp instead of runn directly plugin bridge
>                 cannot (sometimes)load html resource, beause it is actually on /tmp
>             2) plugin.jar should be on javaws classpath too:(
>             3) too often needs -nosecurity
>
>     1) really should be fixed asap (by me, during this changeset) othewrwise it looses sense
>     2) those is dont know right now, many appelts needs JSObject, but having it, any runs, but more
>     of them fails in runtime later. So it is really question whether to burden javaws with
>     plugin.jar or work differently
>     3) this s "I dont understan now: because it is doing the same as generated jnlp shortct from
>     previous patch (yes this one is on to of it) . Jnlp shortcut works, but this /tmp one dont...
>     Anyway must be fixed in this changeset or later too.
>
>
>     Enjoy!
>        J.
>
>