[Bug target/11565] New: gcc.dg/i386-387-1.c test flawed due to assumptions made about arch

["gschafer at zip dot com dot au" <gcc-bugzilla at gcc dot gnu dot org>]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11565

           Summary: gcc.dg/i386-387-1.c test flawed due to assumptions made
                    about arch
           Product: gcc
           Version: 3.4
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: target
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: gschafer at zip dot com dot au
                CC: gcc-bugs at gcc dot gnu dot org
  GCC host triplet: i686-pc-linux-gnu

Since the introduction of the "--with-cpu" & --with-arch" configure switches, it
is now possible to build the compiler so that the default -march setting is no
longer i386

Unfortunately, gcc.dg/i386-387-1.c seems to not allow for this as evidenced by
the test results here:-

http://gcc.gnu.org/ml/gcc-testresults/2003-07/msg01076.html

in particular:-

FAIL: gcc.dg/i386-387-1.c scan-assembler call	_?sin
FAIL: gcc.dg/i386-387-1.c scan-assembler call	_?cos
FAIL: gcc.dg/i386-387-1.c scan-assembler call	_?sqrt
FAIL: gcc.dg/i386-387-1.c scan-assembler call	_?atan2
FAIL: gcc.dg/i386-387-1.c scan-assembler call	_?log
FAIL: gcc.dg/i386-387-1.c scan-assembler call	_?exp
FAIL: gcc.dg/i386-387-5.c scan-assembler call	_?atan

Note that the compiler in those test results was built with
"--with-arch=athlon-4" but the same holds true for "--with-arch=i686"

It should be possible to build with the new switches and not experience
"apparent" test suite regressions. The test needs to be made smarter but I
unfortunately do not really understand the testing framework and thus cannot
offer a patch (but it probably involves adding a -march switch to the dg-options
in the test).
Thanks