Re: Bug#44460: Fails to comply with the proposed Java Policy
On Monday 6 September 1999, at 16 h 7, the keyboard of cpbs@debian.org wrote:
> As far as I can see, the java-common package is entirely pointless as
> regards java bytecode compilers and JVMs depending on it. It seems to
> contain absolutely nothing they could need.
java-common contains the Java Policy and will contain some utilities, which
are useful for some people (a program which creates a CLASSPATH string from a
list of directories containing jars, for instance).
> /usr/bin/ and /usr/sbin/ are not
> needed;
A bug in java-common. Feel free to submit a bug, I respect bug submitters, I
will work on it.
> There is no reason to force all end users of JVMs and bytecode
> compilers to have the policy document installed. It is intended for
> developers, not end users.
No, the Policy is intended for everybody. It can help users understand how
their packages work (I plan to add a section in the Policy which will not be
normative but targeted to the system administrators). The 'users should not
know what we do' is not my way of thinking. Emacs does the same, for instance,
you get the Emacs policy even if you don't *need* it.
> Thus, I see no reason that any JVM or bytecode compiler need depend on
> java-common. If there is a genuine reason, please document it in your
> policy document.
See above. (It is possible that the Policy lack a 'Rationale' section but,
anyway, I will not mix it with the normative part of the Policy.)
> In addition, you don't seem to be aware that the Java programming
> language and the ".class" bytecode format are two entirely different
> things,
Thanks. Before you, nobody on debian-java was aware of the difference.
> package, "bock", has nothing to do with the bytecode format at all; it
> compiles Java source code to C.
I regard it as a Java compiler, then. Of a "special" kind.
> I think your proposed policy is intended to govern JVMs which interpret
> bytecoded class files, and compilers which compile against and/or
> generate bytecoded class files.
No. It applies to gcj as well, which compiles Java sources to native code.
> I'm closing this bug report, due to its bogosity.
Well, closing the bug without even a prior discussion is something we should do more often: the release manager would be delighted to see the number of critical bugs decrease.
Reply to: