This is the mail archive of the
java@gcc.gnu.org
mailing list for the Java project.
Re: Package-private access checking in GCJ 3.3
>>>*Except* for ACC_SUPER, everything else is there and has
>>>the same value as defined by GCJ - what does ACC_SUPER mean?
>>>Is it to specially flag references like "super.fooBar"?
>
>
> Eric> Also read JVMS chapter 4 to see the history of these
> Eric> values. ACC_SUPER is used for classes (it affects how super.m()
> Eric> is resolved by the virtual machine). Note that a class is never
> Eric> marked synchronized; ACC_SYNCHRONIZED is used for methods, which
> Eric> are never marked super.
>
> Incidentally, I don't think our interpreter properly handles this bit.
> Ranjit, could you put this in bugzilla? Or if someone has been
> waiting to fix a relatively straightforward interpreter bug, have at
> it.
I would file a bug if only I could reproduce the bug. :-)
I have compiled classes with JDK 1.4.1 and with JDK 1.0.2 (the
former sets ACC_SUPER, while the latter does not - I verified
it with jcf-dump) and gij 3.3 seems to handle them just fine.
I do not want to file a bug without a reproducable
testcase, so can you please tell me how I can concoct
a suitable one - I mean, what combo do you think would
show the bug?
Ranjit.
--
Ranjit Mathew Email: rmathew AT hotmail DOT com
Bangalore, INDIA. Web: http://ranjitmathew.tripod.com/