Path: csiph.com!au2pb.net!usenet.blueworldhosting.com!feeder01.blueworldhosting.com!news.glorb.com!usenet.stanford.edu!not-for-mail
From: Richard Frith-Macdonald <richardfrithmacdonald@gmail.com>
Newsgroups: gnu.gnustep.bug
Subject: Re: msys2 patches
Date: Mon, 24 Aug 2015 10:35:57 +0100
Lines: 48
Approved: bug-gnustep@gnu.org
Message-ID: <mailman.13.1440425938.7866.bug-gnustep@gnu.org>
References: <CAKur+CUAtji+e8qeHcFovOaKfarEP+41kGJOzfWwQoBrQ-NbdA@mail.gmail.com>
NNTP-Posting-Host: lists.gnu.org
Mime-Version: 1.0 (Mac OS X Mail 8.2 \(2104\))
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Cc: bug-gnustep@gnu.org
To: Seong Gu Lee <sgleehd@gmail.com>
Envelope-to: bug-gnustep@gnu.org
In-Reply-To: <CAKur+CUAtji+e8qeHcFovOaKfarEP+41kGJOzfWwQoBrQ-NbdA@mail.gmail.com>
Precedence: list
Xref: csiph.com gnu.gnustep.bug:4048


> On 16 Aug 2015, at 12:15, Seong Gu Lee <sgleehd@gmail.com> wrote:
>=20
>  Dear gnustep developers,
>=20
>  It is first time for me to submit something here.
>=20
>  msys2 is more convinient and updated than msys,
> so I tried to build windows version of gnustep with it.
>=20
>  Some build tweaks are need to compile.
>=20
>  Though gnustep windows build seems to be rarely used,
> it would be good that anyone having interest check my tweaks.
>  There may be some violation of patch submitting rule such as format =
or file which I have not noticed.=20
>=20
>  Unfortunately, libojbc2 still cannot be build with msys2/clang.
>  All compiling was done with mingw32(gcc 5.2.0)

Thanks very much for your work.
Fred applied your Korean patches to gnustep-base, and I have been =
looking at the other issues in your patch.

Firstly, the winsock2.h change.
You added this to GSConfig.h.in =E2=80=A6 but that file is pretty much =
included everywhere, so including winsock2.h there means that the =
sockets header gets included everywhere, and in most places the code =
doesn=E2=80=99t need/want that.
So I think what=E2=80=99s actually needed is for the winsock2.h header =
to be included in only those places where it=E2=80=99s actually needed =
(ie in specific .m files and not in any .h file).
Looking at Report_winsock2.txt, I think the problem might just be with =
Source/win32/GSFileHandle.m, so I have tried including windows.h and =
winsock2.h at the start of that file.
Please let us know if that works for you.

Secondly, the BOOL change (Report_objc_bool.txt)
I may be wrong, but it looks to me like this change is intended to work =
around the problems caused by including winsock2.h in GSConfig.h.in
So, if we don=E2=80=99t include winsock2.h there, I suspect this change =
is not needed.
Again, please can you give it a try and let us know.

The simplest way of trying things would be to check out a fresh copy of =
gnustep-base from svn-trunk, and see if it builds on your system.