As some have noticed, PDCurses 2.6 doesn't build cleanly with some recent
compilers. I've made some patches to fix this, which I'll be posting over
the next few messages.
First, here's a patch to add HAVE_STRING_H for Watcom, MinGW and
Borland/Turbo C++. This supresses warnings about memcpy(), etc., and in
the case of one compiler (I forget which) is actually necessary to get it
to build. It should probably be added for some of the others, too.
Also, while HAVE_UNISTD_H was specified for MinGW, I found it didn't like
it, so it's removed by this patch.
Tested with Open Watcom 1.3, Borland C++ 5.5 (free version), and MinGW
3.2.0-rc-3.
--
William McBrine <[EMAIL PROTECTED]>
*** curses-org.h 2005-05-26 19:18:02.107425600 -0400
--- curses.h 2005-05-26 19:32:25.536164480 -0400
***************
*** 238,243 ****
--- 238,246 ----
# ifndef HAVE_STDARG_H
# define HAVE_STDARG_H /* have <stdarg.h> */
# endif
+ # ifndef HAVE_STRING_H
+ # define HAVE_STRING_H /* have <string.h> */
+ # endif
# ifndef HAVE_MEMMOVE
# define HAVE_MEMMOVE /* have memmove() */
# endif
***************
*** 552,559 ****
# ifndef HAVE_STDARG_H
# define HAVE_STDARG_H /* have <stdarg.h> */
# endif
! # ifndef HAVE_UNISTD_H
! # define HAVE_UNISTD_H /* have <unistd.h> */
# endif
# ifndef HAVE_MEMMOVE
# define HAVE_MEMMOVE /* have memmove() */
--- 555,562 ----
# ifndef HAVE_STDARG_H
# define HAVE_STDARG_H /* have <stdarg.h> */
# endif
! # ifndef HAVE_STRING_H
! # define HAVE_STRING_H /* have <string.h> */
# endif
# ifndef HAVE_MEMMOVE
# define HAVE_MEMMOVE /* have memmove() */
***************
*** 653,658 ****
--- 656,664 ----
# ifndef HAVE_STDARG_H
# define HAVE_STDARG_H /* have <stdarg.h> */
# endif
+ # ifndef HAVE_STRING_H
+ # define HAVE_STRING_H /* have <string.h> */
+ # endif
# ifndef HAVE_MEMMOVE
# define HAVE_MEMMOVE /* have memmove() */
# endif