This is the mail archive of the cygwin-patches mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] Fix type inconsistencies in stdint.h


On Sat, Apr 11, 2009 at 07:17:54PM +0100, Dave Korn wrote:
>Christopher Faylor wrote:
>> On Sat, Apr 11, 2009 at 10:07:36AM +0200, Corinna Vinschen wrote:
>>> On Apr 11 04:51, Dave Korn wrote:
>>>> Corinna Vinschen wrote:
>>>>
>>>>> Good point, I guess.  So, if we all agree on that, I'd suggest to
>>>>> change Dave's patch to the one below.
>>>>   Two hunks went astray in the adjustment, the fixes for INTPTR_Mxx and
>>>> SIZE_MAX still apply because we didn't change their types.
>>>>
>>>>   Also, Joseph just introduced a new testcase in GCC SVN, and it shows up a
>>>> problem with our definition of WINT_MAX.
>>> What problem exactly?  UINT_MAX not defined?
>>>
>>>> 	* include/stdint.h (INTPTR_MIN, INTPTR_MAX):  Add 'L' suffix.
>>>> 	(WINT_MAX):  Add 'U' suffix.
>>> Applied.
>> 
>> These are now different than linux.  Didn't we want to make them the same?
>
>  If linux doesn't have them matching the types they define, the linux headers
>are going to need patching upstream as well.  We have
>
>typedef long intptr_t;
>
>therefore the constants must be long as well.  What does Linux have?

 From:
http://sourceware.org/cgi-bin/cvsweb.cgi/~checkout~/libc/sysdeps/generic/stdint.h?rev=1.11&cvsroot=glibc

/* Types for `void *' pointers.  */
#if __WORDSIZE == 64
# ifndef __intptr_t_defined
typedef long int		intptr_t;
#  define __intptr_t_defined
# endif
typedef unsigned long int	uintptr_t;
#else
# ifndef __intptr_t_defined
typedef int			intptr_t;
#  define __intptr_t_defined
# endif
typedef unsigned int		uintptr_t;
#endif

So the answer is "int".

cgf


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]