weak for Darwin

[Benjamin Kosnik]

>There is, however, a complication.  The Darwin loader doesn't really
>handle weak symbols, or symbol overriding of any kind, unless you use
>flat namespaces.  After the afore-mentioned discussion, we concluded
>that the best thing would be if the Darwin loader did handle this, and
>it's planned to be implemented (you'll note that the magic words
>"implemented now" did not appear in that sentence).

Why hack up the FSF sources for this? Couldn't you just keep this on
your vendor branch?

-benjamin