On Wed, 23 Jun 1999 23:16:50 -0400 (EDT) 
 Dan Winship <danw@MIT.EDU> wrote:

 > 1) bsd.own.mk sets NOPROFILE for the powerpc, but if you remove that
 >    and build something with profiling, you get results that look
 >    perfectly believable.

Cool.  The compiler used to abort().

Do we have the MCOUNT goo for PowerPC?

 > 2) src/{domestic/,}libexec/ftpd/Makefile turns off optimization on the
 >    powerpc "to prevent a core dump in cc1". If you remove that and
 >    compile with optimization, it builds fine, and basic testing
 >    doesn't turn up any problems with the resulting executable
 > 
 > Both of these hacks date to 1997. I'm guessing egcs fixed the relevant
 > problems. Does anyone know any reason I shouldn't revert both of them?
 > (Particularly Jason, since you're the one who put both of them in.)

I say undo the hacks :-)

        -- Jason R. Thorpe <thorpej@nas.nasa.gov>