[PATCH] Warning for main returning a bool.

[Hal Finkel via cfe-commits]

----- Original Message -----
> From: "Aaron Ballman" <aaron at aaronballman.com>
> To: "Hal Finkel" <hfinkel at anl.gov>
> Cc: "cfe-commits" <cfe-commits at lists.llvm.org>, "Joshua Hurwitz" <hurwitzj at google.com>
> Sent: Tuesday, November 15, 2016 4:42:05 PM
> Subject: Re: [PATCH] Warning for main returning a bool.
> 
> On Tue, Nov 15, 2016 at 5:22 PM, Hal Finkel <hfinkel at anl.gov> wrote:
> > ----- Original Message -----
> >> From: "Aaron Ballman via cfe-commits" <cfe-commits at lists.llvm.org>
> >> To: "Joshua Hurwitz" <hurwitzj at google.com>
> >> Cc: "cfe-commits" <cfe-commits at lists.llvm.org>
> >> Sent: Tuesday, November 15, 2016 12:17:28 PM
> >> Subject: Re: [PATCH] Warning for main returning a bool.
> >>
> >> On Fri, Oct 14, 2016 at 1:17 PM, Joshua Hurwitz via cfe-commits
> >> <cfe-commits at lists.llvm.org> wrote:
> >> > See attached.
> >> >
> >> > Returning a bool from main is a special case of return type
> >> > mismatch. The
> >> > common convention when returning a bool is that 'true' (== 1)
> >> > indicates
> >> > success and 'false' (== 0) failure. But since main expects a
> >> > return
> >> > value of
> >> > 0 on success, returning a bool is usually unintended.
> >>
> >> I am not convinced that this is a high-value diagnostic. Returning
> >> a
> >> Boolean from main() may or may not be a bug (the returned value is
> >> generally a convention more than anything else). Also, why Boolean
> >> and
> >> not, say, long long or float?
> >
> > I've seen this error often enough, but I think we need to be
> > careful about false positives here. I recommend that we check only
> > for explicit uses of boolean immediates (i.e. return true; or
> > return false;) -- these are often bugs.
> 
> I could get behind that.
> 
> > Aaron, I disagree that the return value is just some kind of
> > convention. It has a clear meaning.
> 
> For many hosted environments, certainly. Freestanding
> implementations?
> Much less so, but I suppose this behavior is still reasonable enough
> for them (not to mention, there may not even *be* a main() for a
> freestanding implementation).
> 
> > Furthermore, the behavior of the system can be quite different for
> > a non-zero exit code than otherwise, and users who don't
> > understand what's going on can find it very difficult to
> > understand what's going wrong.
> 
> That's a fair point, but my question still stands -- why only Boolean
> values, and not "return 0x1234567800000000ULL;" or "return 1.2;"?
> 
> Combining with your idea above, if the check flagged instances where
> a
> literal of non-integral type (other than Boolean) is returned from
> main(), that seems like good value.

Agreed.

FWIW, if we have a function that returns 'int' and the user tries to return '1.2' we should probably warn for any function.

 -Hal

> 
> ~Aaron
> 
> >
> > Thanks again,
> > Hal
> >
> >>
> >> ~Aaron
> >>
> >> >
> >> > _______________________________________________
> >> > cfe-commits mailing list
> >> > cfe-commits at lists.llvm.org
> >> > http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
> >> >
> >> _______________________________________________
> >> cfe-commits mailing list
> >> cfe-commits at lists.llvm.org
> >> http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
> >>
> >
> > --
> > Hal Finkel
> > Lead, Compiler Technology and Programming Languages
> > Leadership Computing Facility
> > Argonne National Laboratory
> 

-- 
Hal Finkel
Lead, Compiler Technology and Programming Languages
Leadership Computing Facility
Argonne National Laboratory