From: Dae San Hwang <lists@...>
Date: 2008-12-04T13:40:39+09:00
Subject: [ruby-core:20291] Re: [Bug #805] Ruby 1.9.1 preview 2 : build failure on OpenSolaris

I traced back header inclusions:

"ruby-trunk/ext/curses/curses.c" conditionally includes "<curses.h>"

"/usr/include/curses.h" includes "<widec.h>"

"/usr/include/widec.h" includes "<euc.h>"

"/usr/include/euc.h" includes "<sys/euc.h>"

"/usr/include/sys/euc.h" defines "#define ISPRINT(c, wp)   
(wp._multibyte && !ISASCII(c) || isprint(c))"


Then "ruby-trunk/ext/curses/curses.c" defines function "curses_getch"  
which uses ISPRINT macro. (In Ruby 1.9.1 Preview 1, "curses_getch"  
didn't use ISPRINT macro.)


On Dec 4, 2008, at 12:39 PM, Nobuyoshi Nakada wrote:

> Hi,
>
> At Thu, 4 Dec 2008 12:15:03 +0900,
> Dae San Hwang wrote in [ruby-core:20284]:
>> I've checked system header files and found following macro  
>> definitions
>> of ISPRINT.
>>
>>
>> /* in "/usr/include/readline/chardefs.h" */
>>
>> #define ISPRINT(c)      (IN_CTYPE_DOMAIN (c) && isprint (c))
>>
>>
>> /* in "/usr/include/sys/euc.h" */
>>
>> #define ISPRINT(c, wp)  (wp._multibyte && !ISASCII(c) || isprint(c))
>
> Hmmm, they obviously conflict.
>
>
>> I guess the second definition of ISPRINT macro clashes with ruby?
>
> I guess so.  Could you tell how/from where those headers are
> included?
>
> -- 
> Nobu Nakada
>