18838 - You can't access Nav4 Plugin methods using LiveConnect and JavaScript

Status: RESOLVED WONTFIX

(Core Graveyard :: Java: Live Connect, enhancement, P3)

Description

[Neil Marshall]

Currently you can't access any methods which a Nav4 plugin may have which
effectivly breaks a lot of pages.

The ones that come to mind would be ones that use the Flash fs command.
And the ones we are currently developing at work (To be released in a couple
weeks).

Comment 1

[Scott Furman]

We've been over this issue many times, so I'm virtually certain this is a
WONTFIX because of the enormous difficulty of the task, but it's not my decision
to make.  Due to the design of the old NN system for plugin methods, which
required the active cooperation of the JVM, any attempt at implementing this
functionality could only be partially effective.  If anyone is interested in
taking up this task, however, I can outline the broad steps necessary to write
the adapter code.

Anyway, passing this bug on to the owner of plugins.

Comment 2

[av (gone)]

Changing to REMIND

Comment 3

[Blake Ross]

reopening and marking future

Comment 4

[Blake Ross]

reopening and marking future...

Comment 5

[ekrock's old account (dead)]

Let's stop kidding ourselves. This is infeasible for technical reasons due to
the dependency on the now-defunct Nav4 JVM and the fact that we'd need to get
third-party JVMs hacked in baroque ways to try to fix this. The way this problem
will actually be solved is for plug-in vendors to upgrade their plug-ins to
support the Mozilla Plug-in API, not through hacking both Gecko and multiple
third-party JVMs to try to emulate the old calls. (Major plug-in vendors will
have finished their upgrades long before any such hackery would be completed,
and as fur notes it would likely be only partially successful at best anyway.)

Marking WONTFIX. Netscape definitely won't be supplying engineering resources to
attempt to fix this. If anyone chooses to reopen, they should reassign both eng
and qa contacts to [e-mail] (or themselves) and mark helpwanted. But
again, I think that would be people kidding themselves, and I think that the
nglayout module owners would be reluctant to take a half-working fix into the
codebase.