Re: A patch for visibility.

[Martin von Loewis <loewis at informatik dot hu-berlin dot de>]

> > I don't know what these are good for, and they eventually have to go,
> > but I don't think we should break the tools on that target if there is
> > no reason. Your patch would do that.
> 
> This only means on V850 you will never be able to comply with the
> current specs and never will have visibility.

It means that and more: it means that binutils break the current usage
of st_other on v850. I agree that it does not comply with the future
revision of the gABI that is currently being drafted, it certainly
complies with the *current* spec.

I saw no reason to break these targets when I originally wrote the
visibility code; I've also asked Cygnus for a position whether it
would be acceptable to break them. If your position is an official
Cygnus one (we don't care about breaking it and will find other
solution), then that's fine with me. I just wanted to refute the claim
that HJ's recent patch does not break anything - it does.

Regards,
Martin