------- Comment #11 from bangerth at gmail dot com 2010-02-18 20:53 ------- (In reply to comment #10) > I'm not sure to fully understand, Wolfgang: you mean, we should change that > line in the library instead of dealing with a possible C++ issue here? That > would be easy to do, just tell me the exact form you would like to see, whose > correctness you are 101% sure about, and let's close this PR!
I think there are two issues: 1/ A possible bug in the C++ front end, though from what I read Nathan felt unsure about it. 2/ A poorly written typedef that, depending on the outcome of 1/ may or may not be invalid, but in any case isn't easy to understand to the human reader. Ideally, there would be separate PRs about these two issues. I was just addressing the second point, which could be solved by replacing typedef typename ctype::mask mask; by typedef typename __ctype_abstract_base<_CharT>::mask mask; which makes it abundantly clear that ctype::mask is not meant to be a reference to the typedef we are currently declaring. Best W. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=9990
