On Oct 29, 2012, at 6:48 PM, Chandler Carruth <[email protected]> wrote:
> When Ted originally checked in -Weverything it was with the express
> intent of *not* being for general use, but being primarily for Clang
> developers or folks exploring the total set of warnings available.
Interesting. I have to look back at what I exactly said, but certainly the
"folks exploring the total set of warnings available" does fall under the
category of "general use" for some class of users. -Weverything was originally
motivated by users who wanted to discover all compiler warnings, and then turn
off the ones they don't want to use.
>
> There is a *long* path toward changing that state of the world
> including contradictory warnings and duplicate warnings which are
> often needed to be compatible with other compilers, different code
> bases' expectations of different compilers, and a long legacy of
> warning (mis)use. Essentially, I don't think we're in a good position
> to use "-Weverything isn't user-friendly" as a justification for
> changing warnings.
Agreed. I don't think that's a good justification either.
> That said, there may well be very good reasons to
> get rid of this warning. For example…
Right, I prefer we focus on those.
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
