When InterfaceAdapter is used, the wrong thisObj is used

[rPraml]

I had an issue when I pass a JS object to Java class and an InterfaceAdapter is constructed:

Java interface

public interface AdapterInterface {
        int m1(int i);
        int m2();
}

Corresponding impl. in javascript:

function Obj() { this.myObj = {one: 1} }
Obj.prototype.m1 = function(i) { return i + this.myObj.one } // this.myObj is undefined
Obj.prototype.m2 = function() { return 7 }

When I pass this object in my AdapterClass, I get the error: this.myObj is undefined
This PR provides a test and a possible fix

[rPraml]

thisObject is an instance of JavaAdapterInvokeTest.AdapterClass and we want to call a method on target

[gbrail]

This seems logical to me but it's a pretty fundamental thing to the Java embedding support and I'd like to see someone who works with this stuff regularly take a look before merging it.

[p-bakker]

Not a regular user of code that uses code constructs like the testcases, calling a java method passing as argument a JavaScript objects that implements a java interface, but the change looks ok to me

[gbrail]

I'd love to see a few of the questions answered on the PR above and then see if it makes sense to merge this.

[rPraml]

This test shows the problem. this.myObj and this.seven were undefined, when the js object is passed in an InterfaceAdapter
The other tests are all passing without the fix and are for completeness, not to break anything else

[rPraml]

@gbrail I resolved the merge conflicts and updated the test with some documentation

[rPraml]

Hmm. Seems that there might be a flaky test

ozillaSuiteTest > [3846, js=testsrc/tests/js1_5/Regress/regress-211590.js, opt=0] > runMozillaTest[3846, js=testsrc/tests/js1_5/Regress/regress-211590.js, opt=0] STANDARD_OUT
    In "testsrc/tests/js1_5/Regress/regress-211590.js":
     FAILED! Math.random should be random : Expected value 'between 48% and 52%', Actual value ' is 47.990' 
     FAILED! Math.random should be random : Expected value 'between 48% and 52%', Actual value ' is 47.990'

Will investigate later

[p-bakker]

Converted this to a draft as it currently requires more work as the pipeline fails

Please remove the draft status once the pipeline passed and the PR can be considered for merging,.

This way it allows the committers to focus on ready-to-merge PRs

[rPraml]

Not a regular user of code that uses code constructs like the testcases, calling a java method passing as argument a JavaScript objects that implements a java interface, but the change looks ok to me

I've added a more realistic testcase where i use var runnable = new MyRunnable() and pass this object as constuctor argument for java.lang.Thread:
https://github.com/mozilla/rhino/pull/1453/files#diff-05cf7412910fd42af1934af105265073d54b9fd36c8d433d53b3e51a4fae8650R106
(without the fix, the test fails)