Opening this bug to address the CCC review suggested changes:
(1) It seems that JDWP commands ConstantPool and ClassFileVersion
are optional capabilities - this is *explicitly* contained in
javadoc for appropriate JDI's methods. However, JDWP spec
doesn't describe that behaviour.
Could you please add something like that:
"Requires canGetConstantPool capability - see CapabilitiesNew."
to the above commands (if they will use the same "canGetConstantPool"
capability).
It results from that the following error:
"NOT_IMPLEMENTED If the target virtual machine does not support
the retrieval class version information."
(or
"NOT_IMPLEMENTED If the target virtual machine does not support
the retrieval constant pool information.)
should be added to JDWP ClassFileVersion Command as well.
(2) Could you please remove from the ClassFileVersion Command spec
redundant statement (I believe that "Since JDWP version 1.6." is sufficient):
"...
Returns the class file major and minor version numbers, as defined in the class
file format of the Java Virtual Machine specification.
Only works with 1.6 and later versions of JDWP.
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Since JDWP version 1.6.
..."
(3) Could you please fix typos in:
- javadoc for JDI ReferenceType.constantPoolCount() (entries -> enteries):
"Returns:
total number of constant pool enteries of a class. "
^^^^^^^^
- javadoc for JDI ReferenceType.constantPool() and JDWP spec for ConstantPool Command:
(Return -> Returns):
"Return the raw bytes of ..."
^
- JDWP spec for ClassFileVersion Command (infromation -> information):
"The class file version infromation is absent for primitive and array."
^^^^^^^^^^^
(4) It would be better to change JDWP spec for ConstantPool Command
from:
"ABSENT_INFORMATION The class file version infromation is absent for primitive and array."
to:
"ABSENT_INFORMATION The constant pool information is absent for primitive and array."
What do you think about this?
###@###.### 2005-05-20 16:28:56 GMT