Anyone able to review this? It's minor since it doesn't seem to affect valid code, but still would like to get this fixed if my patch is OK....
Also, I've attached patch updated against the latest revision Thanks, Stephen On Tue, Mar 12, 2013 at 5:30 PM, Stephen Lin <[email protected]> wrote: > Hi, > > This is a bug fix for LLVM Bugzilla bug #15485 ( > http://llvm.org/bugs/show_bug.cgi?id=15485) > > Basically, llvm_unreachable is currently hit because CheckOverload assumes > that a new function declaration in record scope should trigger member > function hiding rules, but this is not the case for "friend" declarations. > > I believe the code causing the issue is illegal (the friend declaration > should not friend the function brought into scope by the using) and should > result in an error; this is what happens after the patch without any extra > effort (although the error could possibly be improved?). > > Please let me know if you have any comments. > > Best, > Stephen > > P.S. The patch is in git diff format; hope that's ok >
pr15485.patch
Description: Binary data
_______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
