[Bug testsuite/39537] overhaul printf formats and type casts in testsuite

["pinskia at gcc dot gnu dot org" <gcc-bugzilla at gcc dot gnu dot org>]

------- Comment #2 from pinskia at gcc dot gnu dot org  2009-03-24 06:54 -------
Let's look at the testcases:
g++.dg/ext/align1.C: a run testcase, the printf can be removed in this case.
g++.old-deja/g++.law/operators28.C: An assemble only testcase so the undefined
behavior is not invoked at runtime so it does not matter.
gcc.dg/matrix/matrix-2.c: Weird as it has two dg-do, I don't know which one
should be used here, but those changes look correct
g++.dg/opt/alias2.C: that looks obviously correct.
g++.old-deja/g++.abi/vbase1.C: Likewise.
g++.old-deja/g++.brendan/template8.C: Again it only assembles the testcase so
the undefined behavior does not matter.
g++.old-deja/g++.eh/ptr1.C:  that looks obviously correct.
g++.old-deja/g++.jason/access23.C: Just gets assembled so the undefined
behavior does not matter.
g++.old-deja/g++.law/cvt8.C: Likewise.
g++.old-deja/g++.mike/net35.C: that looks obviously correct.
g++.old-deja/g++.mike/offset1.C: Likewise.
g++.old-deja/g++.mike/p12306.C: That code is obviously dead code, it is if(0)
out so maybe it should be removed ...
g++.old-deja/g++.mike/p3579.C: : that looks obviously correct.
g++.old-deja/g++.mike/p3708a.C: Likewise.
g++.old-deja/g++.mike/p3708b.C: Likewise.
g++.old-deja/g++.mike/p3708.C: Likewise.
g++.old-deja/g++.mike/p646.C:  Just gets assembled so the undefined behavior
does not matter.
g++.old-deja/g++.mike/p710.C: Likewise.
g++.old-deja/g++.mike/p789a.C: that looks obviously correct.
g++.old-deja/g++.mike/pmf2.C: Likewise.
g++.old-deja/g++.mike/temp.C: Likewise.
g++.old-deja/g++.other/temporary1.C: Likewise.
g++.old-deja/g++.other/virtual8.C: Likewise.
g++.old-deja/g++.pt/memtemp23.C: This testcase just links so the undefined
behavior does not matter.
g++.old-deja/g++.pt/memtemp24.C: Likewise.
g++.old-deja/g++.pt/memtemp25.C: Likewise.
g++.old-deja/g++.pt/memtemp26.C: Likewise.
g++.old-deja/g++.pt/t39.C: Looks wrong, it should use %p for even ptr[2], the
printf itself does not matter.
g++.old-deja/g++.robertl/eb17.: that looks obviously correct.
gcc.dg/pch/inline-4.c: This does not matter as only the assembly is compared so
calling printf with a non string constant is ok.

In summary some of the changes are correct, others technically don't matter and
really are really changing for the better as the testcase changes itself and
changes what the testcase is testing and Ubuntu is broken for enabling those
warnings by default even when testing GCC.


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|enhancement                 |trivial


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39537