On Wed, 2005-03-09 at 15:04 +0000, Andrew Haley wrote:
Ziga Mahkovec writes:
> On Wed, 2005-03-09 at 14:44 +0000, Andrew Haley wrote:
> > Ziga Mahkovec writes:
> > > On Wed, 2005-03-09 at 14:08 +0000, Andrew Haley wrote:
> > > > Ziga Mahkovec writes:
> > > > > Thanks Andrew, I put the steps and log files here:
> > > > >
http://www.bootchart.org/misc/ecj/ecj-native.html
> > > >
> > > > You didn't try
> > > > -Dgnu.gcj.precompiled.db.path=/usr/lib/eclipse/eclipse.db without
> > > > jdtcore.jar.so.
> > >
> > > Updated page. I guess this is the same as running interpreted, since
> > > most of the class lookups fail.
> >
> > It is: it seems that you don't have a precompiled version of
> > /usr/share/java/jdtcore.jar in your eclipse.db. Either that, or the
> > corresponding precompiled library isn't readable.
>
> The corresponding library
> is /usr/lib/eclipse/plugins/org.eclipse.jdt.core_3.1.0/jdtcore.jar.so
> right?
I would guess so, but I don't know.
> I do have it in eclipse.db, but for the last two tests I renamed it, in
> order to force tools.jar to not use the gcjlib:// class loader.
I see. Of course, I have no idea why tools.jar is attempting to use
the gcjlib:// class loader, or why it is so slow when it does. It
really should not explicitly use this class loader.
I doesn't any more. We were doing that to avoid having to always add
jdtcore.jar to the classpath. I also have no idea why it's so slow but
I did notice a significant performance hit when compared to interpreted
mode.
Ziga, you can get the sources for java-gcj-compat from
sources.redhat.com:
cvs -d :pserver:anoncvs@sources.redhat.com:/cvs/rhug co java-gcj-compat
I've removed gcjlib:// loading in favour of using the database, so when
everything's set up properly ecj should run natively using
jdtcore.jar.so.
Tom