class file extensions in JDK 7, for ASM 4.0

I was looking at JDK 7 new features, from http://openjdk.java.net/projects/jdk7/features/

It seems that 3 JSRs define extensions to the class file format:
- JSR 292 for invokedynamic
- JSR 294 for modular programing
- JSR 308 for annotations on Java types
or do I miss something?

all these JSRs are marked as inactive, and I did not found official documents describing the proposed changes to the class file format. This is strange because I read somewhere that JDK 7 is supposed to be released at the end of this year?

The most recent documents I found for JSR 294 and 308 are:
- http://openjdk.java.net/projects/jigsaw/doc/classfile.html
- http://types.cs.washington.edu/jsr308/specification/java-annotation-design.html#class-file

JSR 294 adds a new access flag (ACC_MODULE), as well as four new class attributes:

Module_attribute {
  u2 attribute_name_index;
  u4 attribute_length;
  u2 module_id_index;
}

ModuleRequires_attribute {
  u2 attribute_name_index;
  u4 attribute_length;
  u2 requires_length;
  { 
    u2 requires_index; 
    u1 flags; 
  } requires_table[depends_length];
}

ModulePermits_attribute {
  u2 attribute_name_index;
  u4 attribute_length;
  u2 permits_length;
  { u2 permits_index } permits_table[permits_length];
}

ModuleProvides_attribute {
  u2 attribute_name_index;
  u4 attribute_length;
  u2 provides_length;
  { u2 provides_index; } provides_table[provides_length];
}

providing support for this in ASM is not complicated (with four new methods in ClassVisitor
and a new constant in Opcodes)

JSR 308 adds two new attributes: RuntimeVisibleTypeAnnotations and RuntimeInvisibleTypeAnnotations
they contains extended_annotation values, which are like annotation, with these additional fields at the end:

    u1 target_type;    // the type of the targeted program element, 36 possible values!
    union {
        offset_target {
            u2 offset;
            // if within parameterized type or array type
            u2 location_length;
	    u1 location[location_length];
        };
        localvar_target {
            u2 table_length;
            {
                u2 start_pc;
                u2 length;
                u2 index;
            } table[table_length];
            // if within parameterized type or array type
            u2 location_length;
	    u1 location[location_length];
        };
        methodparam_target {
            u1 parameter_index;
            // if within parameterized type or array type
            u2 location_length;
	    u1 location[location_length];
        };
        field_target; // no definition ???
        typeparam_target {
            u1 param_index;
            // if within parameterized type or array type
            u2 location_length;
	    u1 location[location_length];
        };
        typeparam_bound_target {
            u1 param_index;
            u1 bound_index;
            // if within parameterized type or array type
            u2 location_length;
	    u1 location[location_length];
        };
        typearg_target {
            u2 offset;
            u1 type_index;
            // if within parameterized type or array type
            u2 location_length;
	    u1 location[location_length];
        };
        supertype_target {
            u2 type_index;
            // if within parameterized type or array type
            u2 location_length;
	    u1 location[location_length];
        };
        typeindex_target; // no definition ???
        wildcard_bound_target {
            u1 wildcard_location_type;
            target_info;
            // if within parameterized type or array type
            u2 location_length;
	    u1 location[location_length];
        }
        empty_target {
            // if within parameterized type or array type
            u2 location_length;
	    u1 location[location_length];
        };
    } target_info; 

quite complex! Especially to keep such a structure up to date, when doing bytecode modifications (there are many indices and code offsets to elements that could be modified by a bytecode transformation). Also handling these extensions in AnnotationVisitor will not be simple!

Anyway, before thinking about that, we need to know if these proposed changes are quite stable, if they can still be modified, if they will effectively be included in Java 7, etc. Any news about that?

Eric

-- 
You receive this message as a subscriber of the [email protected] mailing list.
To unsubscribe: mailto:[email protected]
For general help: mailto:[email protected]?subject=help
OW2 mailing lists service home page: http://www.ow2.org/wws