http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53531
Jonathan Wakely <redi at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |jason at gcc dot gnu.org
--- Comment #2 from Jonathan Wakely <redi at gcc dot gnu.org> 2012-07-10
11:01:40 UTC ---
It's now rejected on trunk:
c.cc:3:15: error: template argument 1 is invalid
int i = S<,,,,>::undefined;
^
c.cc:3:15: error: template argument 2 is invalid
c.cc:3:15: error: template argument 3 is invalid
c.cc:3:15: error: template argument 4 is invalid
c.cc:3:15: error: template argument 5 is invalid
(I might have been using an older 4.8 snapshot when I reported the bug, not
sure now.)
Jason, any idea which patch fixed these cases and if we need anything in the
testsuite to prevent regressions before closing it?
