Re: A patch for visibility.

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

> This does not make any sense.  The visibility is a property of a
> defined symbol.

The ABI says

# Second, if any reference to or definition of a name is a symbol with
# a non-default visibility attribute, the visibility attribute must be
# propagated to the resolving symbol in the linked object. 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.

So the intent is clearly that visibility on references must also be
supported.

> > Besides: there are still some systems that use st_other in an other
> > meaning. It would probably be good if we treat those gracefully - I
> > tried the best I could.
> 
> They are treated correctly.

Not in HJ's patch, AFAICT: It would clear some bits in st_other, no
matter whether they indicated visibility or not.

Regards,
Martin