X-Recipient: archive-cygwin AT delorie DOT com
X-SWARE-Spam-Status: No, hits=-0.6 required=5.0	tests=AWL,BAYES_00,DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,FREEMAIL_FROM,RCVD_IN_DNSWL_NONE,T_TO_NO_BRKTS_FREEMAIL
X-Spam-Check-By: sourceware.org
MIME-Version: 1.0
In-Reply-To: <4C654584.9090305@redhat.com>
References: <AANLkTimQfeALLK=x_UvJYLjusd6ygCtjUKDrdA91e7Ej AT mail DOT gmail DOT com>	<20100812225838 DOT GB15724 AT ednor DOT casa DOT cgf DOT cx>	<4C647D32 DOT 1090907 AT redhat DOT com>	<AANLkTikCc0gdoVTWfhp=xoNY_m=Up4XQiApLmm9K9da+@mail.gmail.com>	<AANLkTin_VGPC-qfVROBhyFqcU0DBD5B8RAxLCF9RsNDC AT mail DOT gmail DOT com>	<4C654584 DOT 9090305 AT redhat DOT com>
Date: Fri, 13 Aug 2010 12:39:46 -0300
Message-ID: <AANLkTiktHdbUiyUzdYRwhLXdWXeh-dCNGg575aVYe7Rt@mail.gmail.com>
Subject: Re: diff /usr/include/endian.orig.h /usr/include/endian.h > endian.h.diff
From: Pedro Izecksohn <pedro DOT izecksohn AT gmail DOT com>
To: cygwin AT cygwin DOT com
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
X-IsSubscribed: yes
Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm
Precedence: bulk
List-Id: <cygwin.cygwin.com>
List-Unsubscribe: <mailto:cygwin-unsubscribe-archive-cygwin=delorie DOT com AT cygwin DOT com>
List-Subscribe: <mailto:cygwin-subscribe AT cygwin DOT com>
List-Archive: <http://sourceware.org/ml/cygwin/>
List-Post: <mailto:cygwin AT cygwin DOT com>
List-Help: <mailto:cygwin-help AT cygwin DOT com>, <http://sourceware.org/ml/#faqs>
Sender: cygwin-owner AT cygwin DOT com
Mail-Followup-To: cygwin AT cygwin DOT com
Delivered-To: mailing list cygwin AT cygwin DOT com

--- Eric Blake <eblake@...> wrote:
>
> Umm - did you copy straight from glibc's endian.h? =C2=A0That's a no-no;
> cygwin generally doesn't want to borrow LGPL sources to avoid any
> licensing questions (borrowing from BSD is okay, on the other hand).
> You would have to implement things from scratch from a documentation
> page, or copy from a less-questionable source, rather than using glibc's
> implementation.
>
> I'm stopping right here, so I don't risk tainting myself. =C2=A0How about=
 you
> instead describe which macros you are missing, so someone can do a
> clean-room implementation of those macros.

  I did not copy my code from glibc's endian.h: I wrote my code
yesterday, while on Windows Vista for 32 bits using joe on Cygwin and
notepad. Now I'm using Ubuntu for x86-64. I'm also surprised that
glibc's code is similar to mine:

glibc's folks wrote:
#ifdef __USE_BSD
/* Conversion interfaces.  */
# include <bits/byteswap.h>

I wrote:
#ifdef _BSD_SOURCE

#include <byteswap.h>

  The man 3 page, from Linux and from the web, does not tell about __USE_BS=
D.

  The next line is equal, probably because the glibc's guy who wrote
it also was using x86-32. If I would be using Ubuntu for x86-64
yesterday I would have written the opposite.
#if __BYTE_ORDER =3D=3D __LITTLE_ENDIAN
  If you want, to swap the order of my macros is very easy.

  Then glibc continues, interleaving be with le, and a pair of hto
with another pair of toh:
#  define htobe16(x) __bswap_16 (x)
#  define htole16(x) (x)
#  define be16toh(x) __bswap_16 (x)
#  define le16toh(x) (x)
  It is very different from what I wrote. I interleaved the types:
#define htobe16(x) bswap_16(x)
#define htobe32(x) bswap_32(x)
#define htobe64(x) bswap_64(x)

  To use bswap to implement hto and toh macros is the most intelligent
approach. Fortunately, glibc's __bswap_16 differs from Cygwin's
bswap_16.

  Also: glibc's style is different from mine: They use a space between
the # and the define.

  After writing this message I'm proud that I'm as smart as the glibc's cod=
er.

  I need the 64 bits macros only. I wrote the other macros for
completeness. In my project, that can not be disclosed now, I use the
arpa/inet.h macros for the smaller types.

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple