On Sat, 2002-05-11 at 14:52, Jakub Jelinek wrote:

> The following patch kills a bunch of dynamic PLT relocations, mainly
> in sunrpc.

Thanks, I've applied the patch.  Although I see 26 plts go...

> (it is of course a JDK bug too, as it should handle unknown
> fds properly, but anyway).

I've told somebody at Sun about this and hopefully they stop writing
such stupid code in future.


> IMHO we should cook up a list of symbols libraries should be able to
> interpose (definitely all malloc functions, etc.) and kill the rest.

That's the intention.  So far I only have a few symbols which need to be
kept.  The malloc functions are the only official ones.  There are a few
more internal symbols.

I'll try to create a list of those which need to go.  Hopefully this
will encourage people to help finishing the work.  This is a very
time-consuming work which is why I haven't done all that already.

-- 
---------------.                          ,-.   1325 Chesapeake Terrace
Ulrich Drepper  \    ,-------------------'   \  Sunnyvale, CA 94089 USA
Red Hat          `--' drepper at redhat.com   `------------------------