This is the mail archive of the
java@gcc.gnu.org
mailing list for the Java project.
Re: Completed Plan for --gc-sections on gcj [IGNORE my previous posts]
- From: Adam Megacz <gcj at lists dot megacz dot com>
- To: bryce at waitaki dot otago dot ac dot nz, java at gcc dot gnu dot org
- Date: 06 Nov 2002 14:02:57 -0800
- Subject: Re: Completed Plan for --gc-sections on gcj [IGNORE my previous posts]
- Organization: Myself
- References: <868z06qs9l.fsf@nowhere.com>
Adam Megacz <gcj@lists.megacz.com> writes:
> If you invoke a given method, every method that overrides it is
> included in the link, as well as everything that those methods refer
> to.
Actually, I realized that this isn't as bad as I thought. The linker
already knows not to even think about including a .o if no symbols in
it are referenced (this happens even without --gc-sections), so you'd
only get this extraneous inclusion if you were using static
fields/methods of the class, but never planned on instantiating the
class (and hence didn't need any of its virtual-dispatch methods).
I think I can live with this; such cases are not that common.
- a
--
Microsoft follows standards the same way fish follow migrating caribou.