IcedTea PPC patches: ok to commit?

[Christian Thalinger]

On Thu, 2007-11-08 at 13:21 +0000, Gary Benson wrote:
> Christian Thalinger wrote:
> > On Thu, 2007-11-08 at 12:47 +0000, Gary Benson wrote:
> > > diff -ru openjdk.orig/corba/make/common/Defs-linux.gmk
> > > openjdk/corba/make/common/Defs-linux.gmk
> > > --- openjdk.orig/corba/make/common/Defs-linux.gmk       2007-11-08
> > > 10:52:39.000000000 +0000
> > > +++ openjdk/corba/make/common/Defs-linux.gmk    2007-11-08
> > > 10:55:41.000000000 +0000
> > > @@ -196,7 +196,11 @@
> > >  
> > >  EXTRA_LIBS += -lc
> > >  
> > > -LDFLAGS_DEFS_OPTION  = -z defs
> > > +ifneq ($(findstring ppc,$(ARCH)),)
> > > +  LDFLAGS_DEFS_OPTION  = -Wl,-z -Wl,defs
> > > +else
> > > +  LDFLAGS_DEFS_OPTION  = -z defs
> > > +endif
> > 
> > As I already said on IRC I'd prefer to use -Xlinker instead of -Wl,
> > but I'm not an expert here.
> 
> I was actually going to change to -Xlinker, but when I went to edit
> the files I found that almost all of them were using -Wl, in the same
> command line.  I thought it looked neater to use the same option.

I see.  I don't know then...

> 
> > The thing I want to mention is, we need this on more architectures
> > than ppc.  So doing a findstring is not optimal.  Wouldn't it be
> > better to use -Xlinker/-Wl on all architectures?
> 
> I have no objection to this, but might it break things?
> I'm no expert here either...

It did not break anything for me on i386 and x86_64.

- twisti