http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57359
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|RESOLVED |WAITING
Last reconfirmed| |2013-05-31
CC| |rguenth at gcc dot gnu.org
Resolution|INVALID |---
Ever confirmed|0 |1
--- Comment #7 from Richard Biener <rguenth at gcc dot gnu.org> ---
(In reply to Dara Hazeghi from comment #5)
> Sorry to ask again on this, but after re-reading, I'm not sure I understand
> the type-punning argument here:
>
> **ppll = ll; // write to u.ll
> *k = 0; // write to u.i
>
> j = *ppa; // u not touched
> ia[0][0] = *j; // u not touched
>
> printf("%d\n", u.i); // read from u.i
>
>
> From what I can tell, this is in fact reading the last member written to (so
> not falling under the unspecified behaviors listed in annex J / 6.2.6.1).
> Perhaps there is an additional restriction I am missing?
If the above is what really happens (the testcase is quite obfuscated ;))
then the testcase is valid.
Can you try to un-obfuscate it somewhat?
