Re: egcs-1.1 release issues

[David Edelsohn <dje at watson dot ibm dot com>]

>>>>> Jeffrey A Law writes:
Jeff> Found 'em.  They aren't really regressions.

Jeff> gcc-2.8.1 and egcs-1.0.3a do not issue a warning when you use -fPIC
Jeff> with ABI_AIX, but egcs-1.1 does.  This causes the testsuite to think
Jeff> those tests failed their test for excess errors.

	Oh, that?  Really?  That was an unforunate and bad assumption on
the part of the testsuite.  That is *exactly* why I added that failure.
The testsuite definitely should not have assumed that all targets utilize
and/or need -fpic/-fPIC as a specific variant. 

	AIX always is position-independent, but too many people were
invoking GCC on AIX with -fpic or -fPIC -- especially to create shared
libraries.  They obviously did not understand AIX and did not know what
they were doing.  I was getting extremely concerned about GCC's internal
use of flag_pic and how a user enabling it on a port where its value is
ignored and meaningless could cause problems.

	After seeing too many confused questions in the gcc and
comp.unix.aix newsgroups with people asking about -fpic or incorrectly
*recommending* that people use -fpic, I felt that I needed to do
something.  Hopefully this will stop Solaris users from answering AIX
questions when they don't know what they are talking about.

David