Buglet in gcj

Alexandre Petit-Bianco apbianco@cygnus.com
Mon Oct 8 21:37:00 GMT 2001


Bryce McKinlay writes:

> Yeah, this is PR 1328. Its a nasty one since you can end up with garbage 
> on the stack when the code is actually expecting null.

The code it place (that I wrote) should be replaced by what's already
in check-init.c. The code should be adapted to check field
initialization across constructor calls, which the current code tries
to do.

./A



More information about the Java mailing list