------- Comment #4 from paolo dot carlini at oracle dot com 2009-01-30 10:18 ------- The diagnostic issue is trivial, just add a check for error_mark_node on the return value of reshape_init_r, consistently with other calls.
However, I'm still analyzing whether we really want to reject. As data points, ICC doesn't, even in strict mode; on the other hand Comeau rejects the identifier "family" already... Mark, do you have a quick opinion? Thanks in advance... -- paolo dot carlini at oracle dot com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |mark at codesourcery dot com http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38928