C++ Parser testcase failure on AIX

[Kaveh R. Ghazi]

 > From: David Edelsohn <dje at watson dot ibm dot com> 
 > 
 > Now that the bulk of the issues with the new C++ Parser have been
 > fixed, there is a new regression in check-g++ that was introduced by
 > the new C++ Parser.  The testcase is g++.dg/ext/typeof2.C.  This
 > failure occurs for powerpc-ibm-aix4.3.3.0.
 > 
 > The testcase uses __asm ("barX") to change the name of a symbol "fooX"
 > to "barX".  With the introduction of the new parser, not all
 > references to fooX are replaced.
 > 
 > The first two cases which *do not* use __typeof (fooX) are converted,
 > but the latter two cases with __typeof (fooX) retain fooX in the
 > symbol name.

FWIW, I get the same exact problem (wrt __typeof) on mips-sgi-irix6.5.
E.g. http://gcc.gnu.org/ml/gcc-testresults/2003-01/msg00031.html

I also get some other (different) g++ testsuite regressions which I'll
report on seperately.

Are you going to file a PR for typeof2.C ?  If yes, then I won't bother.

--
Kaveh R. Ghazi			ghazi@caip.rutgers.edu