Re: A patch for visibility.

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

> Martin, you can stop arguing.  Visibility on undefined references make
> no sense.  Both HJ and I were there when this extension was designed
> and decided.  And just to be sure we asked today for clarification.
> Any future comment is just noise.  The patch seems to do the right
> thing and the intent is to not have any of the bits set for undefined
> references.

My point is: If there are undefined symbols with visibility, it is an
error. That error must be detected by the linker. The patch should be
dead code; if it isn't, there is an error elsewhere.

I'll stop arguing when you convince me, based on facts (and quotes
from normative texts would count as fact here...), rather than by
proof-by-authority ("I was there, therefore I know how it is meant").

In the end, it is the maintainer's decision to accept or reject a
patch, so you can ignore me if you think I'm wrong - it's just that I
think you are :-)

Regards,
Martin