Package-private access checking in GCJ 3.3

[Tom Tromey]

>>>>> "Ranjit" == Ranjit Mathew <rmathew@hotmail.com> writes:

Ranjit> This change however exposes another problem - we do not use
Ranjit> any bit to indicate package-private access and so 
Ranjit> java_accstring_lookup( ) returns the closest available set
Ranjit> flag, if any, and "" otherwise. This makes for some rather
Ranjit> confusing error messages. :-)

Do you see a nice way to solve this?  Could you post a sample program
and message so we know the details?  Maybe we could follow what some
other compiler (jacks, JDK) does.

Ranjit> (BTW, we use the same bit for both ACC_SUPER and 
Ranjit> ACC_SYNCHRONIZED and these are the only two that share a
Ranjit> bit - is this by design or is this an oversight?)

Design.  See java.lang.reflect.Modifier -- these values are specified
by Java, not us.

Ranjit> I have only "bubblestrapped" with this change and tested 
Ranjit> against very simple testcases.

Make sure you rebuild libgcj from scratch.  We may have library bugs
that we don't know of due to this compiler bug.

Tom