This is the mail archive of the
java@gcc.gnu.org
mailing list for the Java project.
Re: question about -fno-optimize-static-class-initialization
- From: Alexandre Petit-Bianco <apbianco at cygnus dot com>
- To: java at gcc dot gnu dot org
- Date: Fri, 30 Nov 2001 13:14:00 -0800 (PST)
- Subject: Re: question about -fno-optimize-static-class-initialization
- References: <3C07F0B8.2040600@bothner.com>
- Reply-to: apbianco at cygnus dot com
Per Bothner writes:
> One big piece is that it appears that the code for
> -fno-optimize-static-class-initialization seems to be conncted with
> the definitive assignment code
Have you read this?
http://gcc.gnu.org/ml/java/2001-08/msg00014.html
Basically we attach local booleans to static class initialization and
we let the definite assignment code figure whether a class is
definitely initialized within a static method. Constant propagation
helps us getting rid if the unnecessary -- at the end of the method
body walk, we know which class(es) a method is known to definitely
initialize and we use the information when appropriate.
Does this help?
./A