http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47131
--- Comment #7 from John Marino <gnugcc at marino dot st> 2011-01-04 16:21:17 UTC --- When OpenBSD i386 finished the test suite, I was surprised to it passed gnat.dg pack9 unexpectedly. This is the "scan tree dump not optimized" test that was set to expect-fail a few weeks ago due to a change in the rtl optimization I believe (all this traces back to a PR). I'm wondering if this change to trans.c now causes SJLJ targets to pass the pack9 test as a side effect?