[Bug testsuite/108152] New: gcc.dg/pr71558.c fails for LLP64

[Bug testsuite/108152] New: gcc.dg/pr71558.c fails for LLP64

[nightstrike at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108152

            Bug ID: 108152
           Summary: gcc.dg/pr71558.c fails for LLP64
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: testsuite
          Assignee: unassigned at gcc dot gnu.org
          Reporter: nightstrike at gmail dot com
  Target Milestone: ---

gcc.dg/pr71558.c has the following lines:

__SIZE_TYPE__ strlen (const char *);
void *malloc (__SIZE_TYPE__);
__SIZE_TYPE__ b = strlen (a);


These are good in that they support LLP64 through size_t, however they fail
because this test case is compiled with -ansi.  When using -ansi, the following
warnings appear:

gcc.dg/pr71558.c:6:1: error: ISO C90 does not support 'long long' [-Wlong-long]
gcc.dg/pr71558.c:7:1: error: ISO C90 does not support 'long long' [-Wlong-long]
gcc.dg/pr71558.c:14:1: error: ISO C90 does not support 'long long'
[-Wlong-long]


And of course, on LLP64 systems, __SIZE_TYPE__ is in fact a long long. 
Hopefully, -ansi can just be removed from the compile options, although I don't
know how to do that.

[Bug testsuite/108152] gcc.dg/pr71558.c fails for LLP64

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108152

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
{ dg-options "" }

[Bug testsuite/108152] gcc.dg/pr71558.c fails for LLP64

[nightstrike at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108152

--- Comment #2 from nightstrike <nightstrike at gmail dot com> ---
(In reply to Andrew Pinski from comment #1)
> { dg-options "" }

That would remove every option, no?  Do others matter, like -pedantic, or
whatever else is there?

[Bug testsuite/108152] gcc.dg/pr71558.c fails for LLP64

[nightstrike at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108152

nightstrike <nightstrike at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |10walls at gmail dot com

--- Comment #3 from nightstrike <nightstrike at gmail dot com> ---
(In reply to Andrew Pinski from comment #1)
> { dg-options "" }

So just to be pedantic (pun intended!), the options currently in use are:

-fdiagnostics-plain-output -ansi -pedantic-errors


If we use dg-options "", both -ansi and -pedantic-errors are removed. 
-fdiagnostics-plain-output stays.  Is this ok?