In article <199410050532.WAA21453@netcom6.netcom.com>,
J.T. Conklin <jconklin@netcom.com> wrote:
: Cc: jtc@cygnus.com, tech-userlevel@NetBSD.ORG
: In-Reply-To: <m0qsKzi-000JILC@twwells.com> from "T. William Wells" at Oct 4, 94 09:18:09 pm
: > Weak references, of the kind I'm familiar with, aren't a solution
: > either. How would yours solve the problem?
: I was under the impression that using weak references is one of the most
: common ways of making a C library ANSI compliant.

Back when I paid attention to these things, "weak references"
were all or nothing; if the user defined the weak reference, the
library got the user's definition. Of course, that type of weak
reference wouldn't help.

If I'm understanding you correctly, a weak reference in the
library can only be satisfied by a definition in the library.
(Or a definition encountered later by the linker?) And an
ordinary reference will not be satisfied by a definition included
for a weak reference unless there isn't already a definition.  If
so, that'll solve the problem.

Be that as it may, until *something* is done to alleviate
namespace pollution, we shouldn't be making it worse!