[Bug c/64186] New: Conversion of signed to unsigned of same rank not performed

[Bug c/64186] New: Conversion of signed to unsigned of same rank not performed

[michael.chapman at cortus dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64186

            Bug ID: 64186
           Summary: Conversion of signed to unsigned of same rank not
                    performed
           Product: gcc
           Version: 4.9.2
            Status: UNCONFIRMED
          Severity: minor
          Priority: P3
         Component: c
          Assignee: unassigned at gcc dot gnu.org
          Reporter: michael.chapman at cortus dot com

Created attachment 34192
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=34192&action=edit
Test case which should return 0, but returns 1

I believe the attached test program should return 0 from main. However, it
returns 1.

This seems to be a common front end problem which applies to many versions of
the compiler.

Tested on:-
   gcc-4.6 (Ubuntu/Linaro 4.6.4-1ubuntu1~12.04) 4.6.4
and
   aps-gcc (GCC) 4.9.2 20141128 (Cortus)

Adding an explicit cast of (unsigned short) just before x (as indicated in the
comment in the attached test case) in the divide produces the correct result. 

My interpretation of the C99 conversion rules is that the conversion to
unsigned short is what should happen implicitly. The code generated seems to
suggest that an implicit conversion of x to signed int was performed.

[Bug c/64186] Conversion of signed to unsigned of same rank not performed

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64186

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
>My interpretation of the C99 conversion rules is that the conversion to unsigned short

You are incorrect, the conversion rules never convert to anything smaller than
int.

[Bug c/64186] Conversion of signed to unsigned of same rank not performed

[michael.chapman at cortus dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64186

--- Comment #2 from Michael Chapman <michael.chapman at cortus dot com> ---
I think you are right. Geez you need to be a lawyer to understand this
sometimes.

[Bug c/64186] Conversion of signed to unsigned of same rank not performed

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64186

Marek Polacek <mpolacek at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
                 CC|                            |mpolacek at gcc dot gnu.org
         Resolution|---                         |INVALID

--- Comment #3 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
Invalid.