This is the mail archive of the
binutils@sourceware.cygnus.com
mailing list for the binutils project.
Re: Another patch for visibility.
- To: Martin von Loewis <loewis at informatik dot hu-berlin dot de>
- Subject: Re: Another patch for visibility.
- From: "H . J . Lu" <hjl at lucon dot org>
- Date: Wed, 31 May 2000 08:10:11 -0700
- Cc: drepper at cygnus dot com, binutils at sourceware dot cygnus dot com
- References: <20000530123111.A3491@valinux.com> <200005311237.OAA19033@pandora>
On Wed, May 31, 2000 at 02:37:02PM +0200, Martin von Loewis wrote:
> > if (symvis && (hvis > symvis || hvis == 0))
> > - {
> > - visibility_changed = true;
> > - h->other = sym.st_other;
> > - }
> > + h->other = sym.st_other;
>
> I remember talk about a revised version of the gABI changing parts of
> the symbol visibility semantics; is that text available?
Not yet.
>
> Because under the current draft text I have, I believe that patch is
> incorrect:
I removed "visibility_changed", which was only used to turn a global
symbol into local. It has nothing to do with attribute propagation.
I don't see how it will change visibility attributes.
>
> # If different visibility attributes are specified for distinct
> # references to or definitions of a symbol, the most constraining
> # visibility attribute must be propagated to the resolving symbol in
> # the linked object.
>
> Is there any code left in binutils that propagates the most
> constraining visibility to the resolving symbol?
This part is not changed.
H.J.