421855 - JavaScript-to-Java LiveConnect not working on Firefox 3 Mac betas. Still working on Windows.

Status: VERIFIED FIXED

(Plugins Graveyard :: Java (Java Embedding Plugin), defect, P2)

Description

[bill]

User-Agent:       Mozilla/5.0 (Macintosh; U; Intel Mac OS X; en-us) AppleWebKit/523.15.1 (KHTML, like Gecko) Version/3.0.4 Safari/523.15
Build Identifier: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9b3) Gecko/2008020511 Firefox/3.0b3

Calls to java from chrome code have stopped working in the Firefox 3 betas on the Mac. They still work in Windows. Any call to the java object from chrome JS (e.g. java.System.out.prinln("hello")) produces a cascade of LiveConnect errors in the System console window:

Mar  9 11:32:31 billorcuttsmacbookpro [0x0-0xa10a1].org.mozilla.firefox[1283]: ### MRJPlugin:  getPluginBundle() here. ###
Mar  9 11:32:31 billorcuttsmacbookpro [0x0-0xa10a1].org.mozilla.firefox[1283]: ### MRJPlugin:  CFBundleGetBundleWithIdentifier() succeeded. ###
Mar  9 11:32:31 billorcuttsmacbookpro [0x0-0xa10a1].org.mozilla.firefox[1283]: ### MRJPlugin:  CFURLGetFSRef() succeeded. ###
Mar  9 11:32:33 billorcuttsmacbookpro [0x0-0xa10a1].org.mozilla.firefox[1283]: MRJ Plugin JEP: JavaScript-to-Java LiveConnect failed -- no security manager!
Mar  9 11:32:35: --- last message repeated 6 times ---
Mar  9 11:32:33 billorcuttsmacbookpro [0x0-0xa10a1].org.mozilla.firefox[1283]: initialization error: Can't load class netscape/javascript/JSObject
Mar  9 11:32:33 billorcuttsmacbookpro [0x0-0xa10a1].org.mozilla.firefox[1283]: MRJ Plugin JEP: JavaScript-to-Java LiveConnect failed -- no security manager!
Mar  9 11:32:35: --- last message repeated 6 times ---
Mar  9 11:32:33 billorcuttsmacbookpro [0x0-0xa10a1].org.mozilla.firefox[1283]: initialization error: Can't load class java/lang/Object
Mar  9 11:32:33 billorcuttsmacbookpro [0x0-0xa10a1].org.mozilla.firefox[1283]: LiveConnect was unable to reflect one or more components of the Java runtime.
Mar  9 11:32:33 billorcuttsmacbookpro [0x0-0xa10a1].org.mozilla.firefox[1283]: Go to http://bugzilla.mozilla.org/show_bug.cgi?id=5369 for details.

This works in Firefox 2 and continues to work on FF3 betas on Windows.

This extension demonstrates the bug- 
http://simile.mit.edu/repository/java-firefox-extension/java-firefox-extension-2.0.0.xpi




Reproducible: Always

Steps to Reproduce:
1. Attempt to use LiveConnect with a FF3 beta on a mac from chrome xul.
2. Observe the errors in System console log.
3.
Actual Results:  
LiveConnect is noop

Expected Results:  
LiveConnect works

Techniques for using Java from extension code are detailed here: http://developer.mozilla.org/en/docs/Java_in_Firefox_Extensions

Comment 1

[Steven Michaud [:smichaud] (Retired)]

Yup.  This was broken (in the Java Embedding Plugin) by a change in
trunk-based Mozilla.org browsers' nsIScriptSecurityManager interface.
(The change happened on 2008-01-04.)

It will get fixed in the next release of the Java Embedding Plugin.
This should be soon ... but since it's waiting on another change to a
different Mozilla.org interface, I'm not sure exactly when.

Comment 2

[Steven Michaud [:smichaud] (Retired)]

I should qualify what I said:

It's JavaScript-to-Java LiveConnect that's broken in Mozilla.org trunk
builds (as of 2008-01-04).

Java-to-JavaScript LiveConnect still works fine.

Comment 6

[Damon Sicore (:damons)]

Steven, when are you going to release a new version of JEP?  Is this going to make it into the RC?  Need to know ASAP.

Comment 7

[Steven Michaud [:smichaud] (Retired)]

I thought it was going to be soon ... but I have a problem, which has to do with bug 408329.  Basically, to make it feasible for me to do a quick release that also fixes bug 408329 on the 1.8 branch, I need some changes on the 1.8 branch.

I'll post a more detailed note at that bug.

Comment 8

[Mike Schroepfer]

(In reply to comment #7)
> I thought it was going to be soon ... but I have a problem, which has to do
> with bug 408329.  Basically, to make it feasible for me to do a quick release
> that also fixes bug 408329 on the 1.8 branch, I need some changes on the 1.8
> branch.
> 
> I'll post a more detailed note at that bug.
> 

Given that the next branch build and trunk build are offset a bit can we get this fixed on trunk first?

Comment 9

[Steven Michaud [:smichaud] (Retired)]

> Given that the next branch build and trunk build are offset a bit
> can we get this fixed on trunk first?

I'd like to avoid releasing one new version now (for the trunk), and
then another one right afterwards (for both the trunk and branch).

But (at bug 408329) it looks like the branch will be changed to match
the trunk.  If that happens I won't need to release two new versions
of the JEP.

(I also assume that no further changes to nsIPrincpal will happen on
the trunk before the Firefox 3 release.  That would also require an
additional new JEP.)

So ... standing on this (possibly rickety) ladder of assumptions, I
plan to take Thursday and Friday off from my Mozilla.org work and
release a new JEP later this week.

Comment 10

[Steven Michaud [:smichaud] (Retired)]

I've just released JEP 0.9.6.4.  See bug 428606.

Comment 11

[Steven Michaud [:smichaud] (Retired)]

(Following up comment #9)

> (I also assume that no further changes to nsIPrincpal will happen on
> the trunk before the Firefox 3 release.  That would also require an
> additional new JEP.)

Oops -- it's changes to nsIScriptSecurityManager (and to its UUID)
that I'm mainly worried about.

Comment 13

[Steven Michaud [:smichaud] (Retired)]

Fixed by landing of JEP 0.9.6.4 on trunk (see bug 428606).

Comment 15

[Marcia Knous [:marcia]]

https://litmus.mozilla.org/show_test.cgi?id=5283 added to Litmus.

Comment 17

[Carsten Book [:Tomcat]]

verified fixed using Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9pre) Gecko/2008050621 Firefox/3.0pre and also the Litmus Test

--> Verified fixed