Patch: PR java/5848

Alexandre Petit-Bianco apbianco@cygnus.com
Wed Mar 6 23:15:00 GMT 2002


Tom Tromey writes:

> Frankly this patch feels like a hack.  It seems to be in the spirit
> of this function, in that it handles another case by unwrapping
> extraneous layers of trees.  However, I'm concerned because there
> might very well be other cases that ought to be handled here.

This is the kind of thing the needs to be re-written, because it's the
implementation is sub optimal and not in the spirit of what is done in
the rest of the compiler (we ought to use COMPONENT_REF.)

> I'm going to rebuild libgcj with this shortly.  Meanwhile, Alex,
> what do you think?

It looks OK. For this kind of cases, I'd like to run it through my set
of tests before we check it in, especially since I think we should
check that into the 3.1 branch.

./A



More information about the Java-patches mailing list