On Sat, 1 Dec 2012, David Edelsohn wrote: > Richard, > > The testcases assume default signed char and fail on systems with > different semantics. I believe that both testcases need to declare c > as signed char to consistently test the desired behavior, right?
Fixed as follows. Richard. 2012-12-03 Richard Biener <rguent...@suse.de> * gcc.dg/torture/pr35634.c: Use signed char. * g++.dg/torture/pr35634.C: Likewise. Index: gcc/testsuite/gcc.dg/torture/pr35634.c =================================================================== --- gcc/testsuite/gcc.dg/torture/pr35634.c (revision 194077) +++ gcc/testsuite/gcc.dg/torture/pr35634.c (working copy) @@ -14,6 +14,6 @@ void foo (int i) int main () { - char c; + signed char c; for (c = 0; ; c++) foo (c); } Index: gcc/testsuite/g++.dg/torture/pr35634.C =================================================================== --- gcc/testsuite/g++.dg/torture/pr35634.C (revision 194077) +++ gcc/testsuite/g++.dg/torture/pr35634.C (working copy) @@ -14,6 +14,6 @@ void foo (int i) int main () { - char c; + signed char c; for (c = 0; ; c++) foo (c); }