On Thu, Oct 2, 2014 at 5:59 PM, Larisse Voufo <[email protected]> wrote:
> > > On Thu, Oct 2, 2014 at 5:16 PM, Richard Smith <[email protected]> > wrote: > >> On Thu, Oct 2, 2014 at 4:13 PM, Larisse Voufo <[email protected]> wrote: >> >>> This qualifier seems to have been added accidentally. (It is the only >>> occurrence of basic_ostream within the same file that is qualified). >>> >>> I have been having a bit of trouble building libc++ successfully; and >>> even when I got that done, I have been unable to run the regression tests >>> successfully. So, I have not been able to verify that the changes I >>> suggested in the attached patch won't break anything. >>> >>> That said, it'd be great to have someone take a look and apply it. >>> >> >> Does this extra qualification break any conforming code, or just code >> that tries to forward-declare basic_ostream? >> > > I can't say for sure. > Actually scratch this (sort of). I ran into this problem when investigating if programs that are currently valid using libc++ and C++11 would still be valid with C++14. I found two main cases of failures, one of which was this. I cannot tell if fixing the other case would still break due to this case or not. > The test case I provide is what I was able to reduce a much larger build > failure down to. > I should add that the failure occurs when C++14 is enabled. > > >
_______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
