gnupic: Re: [gnupic] XWisp2 with gcc patch
Subject:
Re: [gnupic] XWisp2 with gcc patch
From:
Rob Hamerling ####@####.####
Date:
14 Sep 2005 10:00:08 +0100
Message-Id: <4327E695.601@hccnet.nl>
Helo Daniƫl,
I have read your patches for xwisp2, and have some comments/questions.
Please have some patience with me, I have hardly any experience with
using Linux and even less with programming for Linux... On my Linux on
my notebook (SuSE 9.3) there is no GCC installed, I'll see if I can find
and install it!
Daniel Serpell wrote:
> xwisp2.170.linux.unwarn.patch :
> This fixes some of the warnings on the code. This is not
> complete, and some of the casts are really important if you
> want xwisp2 to work under 64bit arches.
If not complete, what would be missing?
> +#if !defined(__LINUX__) && !defined(__linux)
May I suppose that '__linux' is defined by the Linux flavour of GCC?
> -#include <sys\types.h>
> -#include <sys\stat.h>
The forward- in stead of back-slash works for the Open Watcom C compiler
and for the IBM Visual Age C compiler under OS/2 too! So I can safely
change that!
> + #define _searchenv(a,b,c) (*c=0)
Is there no _searchenv() functionality in Linux?
> - sprintf(szBuffer, "WbusNext failed, rc %lu", rc);
> + sprintf(szBuffer, "WbusNext failed, rc %lu", (unsigned long)rc);
When 'rc' is declared as 'int', wouldn't is be better to use:
sprintf(szBuffer, "WbusNext failed, rc %d", rc);
> -extern int main(int argc, unsigned char **argv) {
> +extern int main(int argc, char **argv) {
I never understood the difference between a pointer to a signed or
unsigned character, unless there are to me unknown architectures which
have a different length for these. Never mind.....
Regards, Rob.
--
Rob Hamerling, Vianen, NL phone +31-347-322822
homepage: http://www.robh.nl/