From mboxrd@z Thu Jan 1 00:00:00 1970 Path: news.gmane.org!not-for-mail From: Eli Zaretskii Newsgroups: gmane.emacs.devel Subject: Re: Any posibility get non-dw2 libiconv2-dll and libintl-8.dll for gnutls/libxml2 emacs-compatible Date: Tue, 21 May 2013 19:45:07 +0300 Message-ID: <83hahwuvsc.fsf@gnu.org> References: Reply-To: Eli Zaretskii NNTP-Posting-Host: plane.gmane.org X-Trace: ger.gmane.org 1369154729 23214 80.91.229.3 (21 May 2013 16:45:29 GMT) X-Complaints-To: usenet@ger.gmane.org NNTP-Posting-Date: Tue, 21 May 2013 16:45:29 +0000 (UTC) Cc: emacs-devel@gnu.org To: Miguel Ruiz Original-X-From: emacs-devel-bounces+ged-emacs-devel=m.gmane.org@gnu.org Tue May 21 18:45:30 2013 Return-path: Envelope-to: ged-emacs-devel@m.gmane.org Original-Received: from lists.gnu.org ([208.118.235.17]) by plane.gmane.org with esmtp (Exim 4.69) (envelope-from ) id 1Uepgn-0005XO-JP for ged-emacs-devel@m.gmane.org; Tue, 21 May 2013 18:45:29 +0200 Original-Received: from localhost ([::1]:39457 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Uepgn-0006g9-8c for ged-emacs-devel@m.gmane.org; Tue, 21 May 2013 12:45:29 -0400 Original-Received: from eggs.gnu.org ([208.118.235.92]:56225) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Uepgf-0006fc-0B for emacs-devel@gnu.org; Tue, 21 May 2013 12:45:26 -0400 Original-Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1UepgZ-00051g-Rh for emacs-devel@gnu.org; Tue, 21 May 2013 12:45:20 -0400 Original-Received: from mtaout22.012.net.il ([80.179.55.172]:39042) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1UepgZ-00051O-JY for emacs-devel@gnu.org; Tue, 21 May 2013 12:45:15 -0400 Original-Received: from conversion-daemon.a-mtaout22.012.net.il by a-mtaout22.012.net.il (HyperSendmail v2007.08) id <0MN500100QDZC200@a-mtaout22.012.net.il> for emacs-devel@gnu.org; Tue, 21 May 2013 19:45:13 +0300 (IDT) Original-Received: from HOME-C4E4A596F7 ([87.69.4.28]) by a-mtaout22.012.net.il (HyperSendmail v2007.08) with ESMTPA id <0MN5000KKQJDR8B0@a-mtaout22.012.net.il>; Tue, 21 May 2013 19:45:13 +0300 (IDT) In-reply-to: X-012-Sender: halo1@inter.net.il X-detected-operating-system: by eggs.gnu.org: Solaris 10 X-Received-From: 80.179.55.172 X-BeenThere: emacs-devel@gnu.org X-Mailman-Version: 2.1.14 Precedence: list List-Id: "Emacs development discussions." List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: emacs-devel-bounces+ged-emacs-devel=m.gmane.org@gnu.org Original-Sender: emacs-devel-bounces+ged-emacs-devel=m.gmane.org@gnu.org Xref: news.gmane.org gmane.emacs.devel:159713 Archived-At: > Date: Tue, 21 May 2013 08:52:42 +0200 > From: Miguel Ruiz > > Based on > > http://gcc.gnu.org/wiki/WindowsGCCImprovements The small print near the end of that page says "last edited 2011-03-23". So this is more than 2 years old now, and clearly outdated, as the name of the libgcc_s_dw2-1.dll tells you (that "dw2" part does stand for DWARF 2"). > I am not aware of the very internals of emacs, but, as a conservative > behaviour, I build my w32 emacs production binaries with tdm-gcc-4.7.1-2 > (MinGW/sjlj) (http://tdm-gcc.tdragon.net/) and the crashes have completely > disappeared. I'm glad you solved your problem, but that cannot be the solution for everyone else. There are at least 3 different chains of GCC and related tools ported to Windows; supporting only one of them, and not the main one at that, cannot be good enough. FWIW, that's the first time I heard from anyone that they build Emacs with TDM; most use the tools provided by mingw.org, and a few reported they use MinGW64 tools. I have nothing against supporting TDM as well, but forcing users to use only that is a non-starter. > I follow the recent mails to the list and to gcc list ( > http://gcc.gnu.org/ml/gcc/2013-05/msg00180.html) about the problems in > relation to libgcc_s_dw2-1.dll. If you follow that, you already know that there are no firm conclusions reached in that discussion yet. Until the upstream GCC maintainers say their definitive word on this, raising the issue here is not very useful. > At this moment > > - libiconv-2.dll (in gnutls-3.0.9-w32-bin.zip) > - libintl-8.dll (in gnutls-3.0.9-w32-bin.zip) > > - libiconv-2.dll (in libxml2-2.7.8-w32-bin.zip) > > depend on libgcc_s_dw2-1.dll That's not true, at least not with the GnuTLS and libxml2 distros from my ports (on http://sourceforge.net/projects/ezwinports/files/). I'm quite sure that if you have those dependencies, you got libintl and libiconv from some other place, not from those zip files. Please double check. With the DLLs that I provided in the zip files, there's no dependency on libgcc DLL, and no crashes at all. (I will never ever produce any port that depends on libgcc as a shared library, for reasons that have nothing to do with the issue at hand here.) > Is there any possibility that you provide the gnutls and libxml2 packages, > and by extension, the named libraries, based on non-dw2 binaries, i.e., > built with a non-dw2 compiler like > tdm-gcc-4.7.1-2, for example, as I do in a regular basis with emacs builds? Sorry, but I will only use one toolchain, and currently that is the toolchain distributed by mingw.org. I cannot afford to have more than one development environment, and I cannot invest more effort I already do in porting optional libraries. Moreover, since the problem with libgcc_s_dw2-1.dll doesn't happen with my ports, I see no reason to switch to TDM toolchain in the first place. The problem which I'm trying to solve, and which prompted me to start a discussion on the GCC list, is that I'd like this to work with any GCC-based toolchain. So switching to TDM, even if I wanted to, is not the solution at all. Eventually, with help from the GCC maintainers, we will crack this problem, and find how to solve it. I really suggest that people wait for a little bit more; in the meantime, as already written in the other thread here that raised the issue, using the DLLs from my ports is a much easier work-around than switching to a completely different development environment.