#85 uintptr_t on Solaris

[Poor Yorick]

on Solaris, it is probably necessary to add the following line to RPy.h:

#define HAVE_UINTPTR_T

[Gregory Warnes]

Logged In: YES
user_id=9316
Originator: NO

Hi Nathan!

This macro is never used in the rpy source code. Can you give me a bit more information on when this is needed?

-Greg

[Poor Yorick]

Logged In: YES
user_id=1046825
Originator: YES

Hi Greg, I'm opening this again just to make sure you see this comment. I hope I don't screw up your ticket stats!

This may well pertain to local configuration of gcc or python, but the issue is that for whatever reason, setup.py does not define HAVE_UINTPTR_T, causing an error when Rinterface.h gets included. From Rinterface.h:

#if !defined(HAVE_UINTPTR_T) && !defined(uintptr_t)
 typedef unsigned long uintptr_t;
#endif

Here is the error:

In file included from src/RPy.h:73,
 from src/rpymodule2031.c:43:
/path/to/R/lib/R/include/Rinterface.h:86: error: conflicting types for 'uintptr_t'
/usr/include/sys/int_types.h:111: error: previous declaration of 'uintptr_t' was here

--
Nathan