http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54481
Richard Guenther <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |NEW
Last reconfirmed| |2012-09-05
Component|c |tree-optimization
Ever Confirmed|0 |1
Severity|normal |enhancement
--- Comment #1 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-09-05
09:30:04 UTC ---
The issue is that a.fn(&a) is _not_ pure. foo is, but to see that a.fn is
always foo we'd have to know a.fn is always foo ... (we are not optimistically
assuming this which isn't easily possible within any of our memory optimization
frameworks).
Confirmed anyway, in theory it should be possible.
