Subject:
reflection and illegal package access
Date:
Wed, 04 Aug 1999 21:56:20 -0400
From:
Andrew Wason <aw@softcom.com>
To:
norris@netscape.com (Norris Boyd)
CC:
Howard Lin <howard@softcom.com>
If you run Rhino under JDK1.2 with a security manager:
java -Djava.security.manager=java.lang.SecurityManager
org.mozilla.javascript.tools.shell.Main
Then reflection fails for objects that are in a restricted access package
(e.g. sun.*). Rhino is reflecting based on the dynamic type of the object
instead of the declared static return type.
In this example, createImage is declared to return java.awt.Image, but it
returns sun.awt.image.OffScreenImage. Attempting to reflect this class
results in a java.security.AccessControlException for
java.lang.RuntimePermission accessClassInPackage.sun.awt.image.
Here is the script. You will need to type it in because you won't be able
to load it from a file due to the security manager.
var f = new java.awt.Frame();
f.setVisible(true);
var i = f.createImage(10,10);
Subject:
null arguments
Date:
Wed, 04 Aug 1999 13:22:35 -0400
From:
Andrew Wason <aw@softcom.com>
To:
norris@netscape.com
CC:
Howard Lin <howard@softcom.com>
When I try to pass a null argument to an interface implemented in JS, I get:
js: Cannot convert null to an object.
js: uncaught JavaScript exception:
org.mozilla.javascript.EvaluatorException: Cannot convert null to an object.
var b = new Packages.javax.swing.border.Border() {
getBorderInsets : function(c) {
return new Insets(0,0,0,0);
}
};
b.getBorderInsets(null);
Here is the stack trace where the exception is happening:
java.lang.reflect.InvocationTargetException:
org.mozilla.javascript.EvaluatorException: Cannot convert null to an object.
at
org.mozilla.javascript.tools.ToolErrorReporter.runtimeError(ToolErrorReporte
r.java:106)
at org.mozilla.javascript.Context.reportRuntimeError(Context.java:484)
at org.mozilla.javascript.Context.reportRuntimeError(Context.java:500)
at
org.mozilla.javascript.ScriptRuntime.toObject(ScriptRuntime.java:529)
at org.mozilla.javascript.Context.toObject(Context.java:1107)
at adapter0.getBorderInsets(<adapter>)
at java.lang.reflect.Method.invoke(Native Method)
at
org.mozilla.javascript.NativeJavaMethod.call(NativeJavaMethod.java,
Compiled Code)
at org.mozilla.javascript.ScriptRuntime.call(ScriptRuntime.java:1256)
at org.mozilla.javascript.Interpreter.interpret(Interpreter.java,
Compiled Code)
at
org.mozilla.javascript.InterpretedScript.call(InterpretedScript.java:49)
at
org.mozilla.javascript.InterpretedScript.exec(InterpretedScript.java:37)
at org.mozilla.javascript.Context.evaluateReader(Context.java:691)
at
org.mozilla.javascript.tools.shell.Main.processSource(Main.java, Compiled Code)
at org.mozilla.javascript.tools.shell.Main.main(Main.java:146)
Context.toObject does not allow wrapping nulls.
JavaAdapter.generateOverride should generate bytecode to check if an
argument is null and if it is not call Context.toObject.
I'll take a look at fixing this after the other JavaAdapter patches get
checked in so we don't get out of sync.
Andrew
--
Andrew Wason
SoftCom, Inc.
aw@softcom.com
* Accept patch from Andrew Wason <aw@softcom.com>:
Subject:
Re: partial interface problem
Date:
Wed, 04 Aug 1999 13:04:37 -0400
From:
Andrew Wason <aw@softcom.com>
To:
norris@netscape.com
CC:
Howard Lin <howard@softcom.com>
>I'm having a problem implementing a Java interface in JS where I don't
>implement all the methods, and one of the methods I don't define returns
>non-void.
I have a patch for this. I generate bytecode in
JavaAdapter.generateReturnResult to check the return type on the stack from
JavaAdapter.callMethod. If it is Undefined, return null.
I'm not positive this is the right way to fix this - maybe it should be
fixed closer to the source (e.g. prevent callMethod from returning
Undefined to begin with)
Andrew
--
Andrew Wason
SoftCom, Inc.
aw@softcom.com
Subject:
default JavaAdapter patch
Date:
Tue, 20 Jul 1999 15:35:01 -0400
From:
Andrew Wason <aw@softcom.com>
To:
norris@netscape.com
CC:
mccabe@netscape.com, rogerl@netscape.com
Attached is a patch to the patch I sent a while ago for the JavaAdapter stuff.
If a SecurityManager is installed, attempting to access the
"org.mozilla.javascript.JavaAdapter" system property can throw a
SecurityException. This should not prevent the default JavaAdapter
implementation from being used.
Andrew
--
Andrew Wason
SoftCom, Inc.
aw@softcom.com
Subject:
Rhino reflection patch
Date:
Wed, 28 Jul 1999 18:14:52 -0400
From:
Andrew Wason <aw@softcom.com>
To:
norris@netscape.com
CC:
mccabe@netscape.com, rogerl@netscape.com, Howard Lin <howard@softcom.com>
When JavaAdapter generates an adapter class, it does not take into account
the types of method parameters when wrapping the generated methods arguments.
This means that if a non-public class implements a public interface the
non-public class type will be wrapped instead of the declared public
interface - and methods cannot be invoked via the wrapper.
I have attached sample code (reflect-demo.zip) which shows this. The
JavaScript caller.js generates an adapter implementing the CallerInterface
interface. CallerInterface has a method (doSomething) which takes an
argument of type pkg.Interface. pkg.Target is a non-public class that
implements pkg.Interface. If an instance of pkg.Target is passed to the
CallerInterface adapter doSomething method, an Error is thrown because
pkg.Target.doSomething is called (instead of pkg.Interface.doSomething) and
pkg.Target is not public.
I have attached a patch to Context.java, ScriptRuntime.java and
JavaAdapter.java. I overloaded toObject in Context and ScriptRuntime to
take a 3rd argument which is the declared type of the object being
wrapped. This is passed to NativeJavaObject.wrap so that it generates the
correct wrapper. I changed JavaAdapter.generateOverride to generate
bytecode calling Context.toObject passing the declared Class type of the
argument.
Context.java also includes my previously submitted patch for dealing with
SecurityExceptions and the JavaAdapter property (because this patch has not
been checked into CVS yet).
Andrew
--
Andrew Wason
SoftCom, Inc.
aw@softcom.com
reflect-patch.txt
Name:
reflect-patch.txt
Type:
Plain Text (text/plain)
reflect-demo.zip
Name:
reflect-demo.zip
Type:
Zip Compressed Data (application/x-zip-compressed)
Encoding:
base64
2) add a test that acts as a sample for doing oberloaded methods on xpconnect wraped natives.
3) add a NOT_IMPLEMENTED method to wrapped native for getting the prototype JS object.
4) Set the global object of the JSContext as the wrapped native JSObject's parent when creating this JSObject. This makes JS code compiled against the wrapper actuall work!
5) fix the refcounting on factories in the tests/components module
Thanks go to Andrew Wason <aw@softcom.com> for finding this problem and providing the patch:
If we implement a Java interface in JavaScript, and a method in that interface returns a Java object, we get a java.lang.IncompatibleClassChangeError: org/mozilla/javascript/Wrapper exception.
We have attached a sample JavaScript file which duplicates the error.
The problem is JavaAdapter is generating an INVOKEVIRTUAL bytecode to call Wrapper.unwrap, but Wrapper is an interface and so INVOKEINTERFACE should be used instead. As a result of this change, the IFEQ bytecode generated needs to jump more bytes. We have attached a patch that fixes the problem.
Andrew
--
Andrew Wason
SoftCom, Inc.
aw@softcom.com
- normalized initial MPL comment to match the format of others in the tree, including an initial -*- Mode line.
- removed RCS $log$, etc. comments. We use CVS, and they just make spurious changes...
- js_DecompileValueGenerator had rusted due to bytecode/source-note changes,
or maybe parts of it never worked right. Anyway, it now does not induce a
crashing underflow in the decompiler. As part of this fix, it now takes a
checkStack flag telling whether to look for the jsval v argument on the JS
stack. The calls from ImportProperty, js_SetProperty, and js_DeleteProperty
pass in v a jsval for the property id, which should not be sought after on
the stack (it might happen to be there due to o['p'] = 2, but we want to
decompile o["p"], not "p").
- js_DecompileValueGenerator would load a generating pc even if the value v
did not match the pc's corresponding stack item! Oops. This lead to less
than idea diagnostics.
- js_DecompileValueGenerator was also not mapping JSOP_TRAP to the real op at
a sufficiently early and univeral point in its control flow.
- Fix PopOff to assert and check for stack underflow in the decompiler, and
beef up PushOff too (it asserted, but did not check).
- js_ReportIsNotFunction now avoids JS_InternString by indexing directly into
cx->runtime->atomState.typeAtoms with the result of JS_TypeOfValue.
- Removed unnecessary local GC root reserved by non-zero trailing member of
obj_eval's JSFunctionSpec initializer.
Add an #if defined(__arm) check to tell jsdtoa.c about the (somewhat bizzare) endian-ness of the ARM architecture.
Thanks go to Stefan Hanske <sh990154@mail.uni-greifswald.de> for this fix.
2) rooted out all instances of JS_ReportErros and replaced with JS exception throwing.
3) added more exception message types.
4) Added service (also reflected into JS) for getting JS and native stack objects to be used in exceptions.
5) Fixed broken test.
6) updated TODO list.
to define an unbound global variable used to enumerate an empty object, when
the standard says no variable should be defined.
- Clean up minor nits, make comment style match prevailing, tighten up for..in
element-indexing special case comments.
- Fix js_Invoke to take CONSTRUCT and INTERNAL flags arg instead of a construct
boolean arg and the JSStackFrame internalCall mis-located boolean non-arg.
Use these flags correctly in all calls (previous version failed to set the
bogus frame.internalCall around js_Invoke calls from jsfun.c and jsstr.c).
2) make the wrappedJS QI check for a match with chained wrappers before trying QI on the wrapped JS object.
3) add support for brendan's new JSArgumentFormatter callback to do auto js arg conversions - with test.
- Fix js_GetAtom fallibility by returning &dummy on assert-botch "can't happen"
index out of range case.
- js_InitAtomMap needn't bother nulling ale->next with tmp save
- js_InitAtomState explicit tail fusion for FROB via goto bad, and early memset
(I know, JSRuntime is cleared already and it contains the atom state ... but
jsatom.c doesn't know that).
- Clear all ATOM_ flags save ATOM_PINNED when creating a new atom.
- Cleanup xtra, ALIGNNUM, etc. useless variables, use JSVAL_ALIGN and JS_MAX.
Despite my stupidity here, I must say that over-includers elsewhere of both
jspubtd.h and jsapi.h, and even jscntxt.h, with system headers intermixed, suck
a lot too.
"external" calls compiled from scripts, which have stack budget depth slots
under the return value slot for the call expression's generating pc.
- New internalCall packed boolean flag in JSStackFrame to distinguish callers
of js_InternalCall so the generating pc isn't mistakenly stored depth slots
below stack space allocated on-the-fly for internal call args/rval.
- Fixed bugzilla bug 8042, finally by storing the call expression's generating
pc -- but only for external calls.
- Whitespace policing, gratuitous parenthesization reduction, etc.
2) breakout js ops into their own file so that the flattening project is less complicated
3) add a noisy object to the test components - this object prints addrefs, release, QIs, ctor, and dtor. This is useful for refcounting tests.
4) cleanup some component registration stuff.
Support for jsGet_ and jsSet_ prefixes to methods for explicit getter
and setter definition.
Addition of "importClass" and "importPackage" top-level functions.
The beginnings of a history object accessible from the shell.
2) added where I was not checking for JSVAL_NULL from brendan.
3) got rid of 'xpcom32' references.
4) added nsIEcho::ReturnInterface as a loopback for testing the above.
5) removed some redundant retval setting code.
- check static members of instances in JavaMembers.put
- do not unwrap Wrappers before calling NativeJavaMethod.findFunction
or NativeJavaObject.coerceType; both methods may need extra information
provided by the wrapper.
- separate Java signatures for resolving overloaded methods and script
signatures for error messages, so we can distinguish primitive types
from classes.
- separate Java signatures for resolving overloaded methods and script
signatures for error messages, so we can distinguish primitive types
from classes.
- prevent a NativeJavaClass from being treated as a wrapped instance of
java.lang.Class
- correct bug which preferred the *less* specific of two classes in
NativeJavaMethod.preferSignature
- add new LC3 conversion rules to NativeJavaObject.coerceTypes.
- coerce JS numbers to Java numbers or chars only if the JS number is in
range.
- JS_PushArguments{,VA} and JS_PopArguments to convert a list of C/C++ values
passed as actual arguments into an 'jsval *argv' on the JS stack.
- Clean up and robustify ConvertArgs test command in the js shell.
To wit:
Attached are a couple of small fixes for building in the js/ subdir.
These are for FreeBSD-CURRENT, and for building in a seperate OBJDIR.
The FreeBSD one is a quick and dirty. The real fix is pobably just to
use OS_LDFLAGS as defined by configure...
Re: netscape.javascript.JSObject ?
Date:
Thu, 03 Jun 1999 17:52:42 -0700
From:
Frank Mitchell <frankm@eng.Sun.COM>
Organization:
Java Products Engineering
To:
Norris Boyd <norris@netscape.com>
References:
1 , 2 , 3 , 4 , 5 , 6 , 7 , 8
Norris Boyd wrote:
>
> Sorry--missed the checkin of a new file. It's there now.
>
> I'd also added a small change for the "inheritance" of JavaScript array methods.
Actually, I've already done that (and for String as well). It still
fails some LC3 regression tests, though.
I'm including a tarfile that includes the previous changes and the new
ones.
Frank
