http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56815
Manuel López-Ibáñez <manu at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |jason at gcc dot gnu.org
--- Comment #15 from Manuel López-Ibáñez <manu at gcc dot gnu.org> 2013-04-03
12:06:11 UTC ---
(In reply to comment #11)
> In that case we can have a pedwarn instead of a permerror in the C++ front-end
> and restore some consistency.
I think this is also a valid option. In fact, I don't understand why G++
rejects code that it is able to compile perfectly and without ambiguity, but
that has been the norm in the past.
So I see only two valid options:
1) Every void* arithmetic is a permerror in C++, so use emit_diagnostic with
the appropriate DK_ flag in the c-common.c code. This will break some code that
relies on this extension (like GCC). Remove the pedantic||warn_* check for the
permerrors.
2) Every void* arithmetic should be a pedwarn conditional on -Wpedantic |
-Wpointer-arith. This will break less code and people can still use
-pedantic-errors or -Werror=pointer-arith to force the errors.
At the end, this decision needs to be taken by the maintainers (Jason and/or
Gabriel), we can discuss all day the pros and cons of either choice, but they
will decide.
