Gcc 4.7.2

[Armin]

​http://gcc.gnu.org/ml/gcc-announce/2012/msg00005.html

[bdubbs@…]

Built this and there are some regressions:

                === gcc Summary ===

# of expected passes            87527
# of unexpected failures        32

All the failures are in gcc.dg/guality/vla-{1,2}.c with various build options. Checking the gcc-testresults archive, this seems to be fairly common.

Going into the test and adding some printf statements and running that seems to give the right answers, so this may be a test harness issue.

[Armin]

That is one rather strange failure.

All tests passed when I compiled gcc with temporary gcc. [1] vla-{1,2}.c tests failed when I compiled gcc with running gcc (upgraded). [2]

I don't know how or why.

[1] http://gcc.gnu.org/ml/gcc-testresults/2012-08/msg02184.html
[2] http://gcc.gnu.org/ml/gcc-testresults/2012-07/msg01980.html

Those are snapshots of gcc 4.7.2 I played with. Strangely, same thing happened with some more of them:

http://gcc.gnu.org/ml/gcc-testresults/2012-05/msg02092.html (*)
http://gcc.gnu.org/ml/gcc-testresults/2012-06/msg01336.html
http://gcc.gnu.org/ml/gcc-testresults/2012-07/msg02121.html

(*) I am not sure, but I think that one was built with temporary gcc, and my temporary toolchain has gnat included.

I'm sorry for the code blocks, but I think trac doesn't like me today that much.

[Matthew Burgess]

Fixed in r10010.

[Matthew Burgess]

Replying to matthew@…:

Fixed in r10010.

That should, of course, be r10009.