Re: RFC: reflection access checks

[Dalibor Topic]

On Sun, Jan 15, 2006 at 12:35:53AM +0100, Christian Thalinger wrote:
> Hi!
> 
> Today Michael told me about a CACAO bug filed to debian
> (http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=348048).  While
> checking what's wrong, i also checked sun's jvm and this is the
> resulting stacktrace:
> 
> Exception in thread "main" java.lang.IllegalAccessException: Class 
> TestReflect can not access a member of class priv.TestClass with modifiers 
> "public"
>         at sun.reflect.Reflection.ensureMemberAccess(Reflection.java:57)
>         at java.lang.reflect.Method.invoke(Method.java:317)
>         at TestReflect.main(TestReflect.java:23)
> 
> Seen this, i think we also should implement something like sun does in
> ensureMemberAccess, in order that not all vm implementors have to write
> the (more or less) same code.
> 
> Comments?

I believe the reflection code would profit from getting a nice
VMInterfaceLift for the next release :)

I'd like to see it, since it's a security relevant part of the class
libraries, and currently everyone is implementing that in their own
fashion, I guess (and it's probably faster that way for some VMs). 

It'd best go in some reflection specific VMInterface, so that VMs that 
want or have to do it in native code can do it there.

cheers,
dalibor topic

> 
> TWISTI
> 
> 
> _______________________________________________
> Classpath mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/classpath