From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Original-To: caml-list@yquem.inria.fr Delivered-To: caml-list@yquem.inria.fr Received: from mail2-relais-roc.national.inria.fr (mail2-relais-roc.national.inria.fr [192.134.164.83]) by yquem.inria.fr (Postfix) with ESMTP id 05E6BBBAF for ; Mon, 22 Nov 2010 13:04:09 +0100 (CET) X-IronPort-Anti-Spam-Filtered: true X-IronPort-Anti-Spam-Result: An0FAP/q6UzUUiSF/2dsb2JhbACUUI52u1uFSwSFL4Uv X-IronPort-AV: E=Sophos;i="4.59,235,1288566000"; d="scan'208";a="80682140" Received: from one4vision-tunnel.absint.com (HELO mail.absint.com) ([212.82.36.133]) by mail2-smtp-roc.national.inria.fr with ESMTP; 22 Nov 2010 13:04:08 +0100 Received: from mail.absint.com (localhost [127.0.0.1]) by mail.absint.com (Postfix) with ESMTP id D1AFD11C01D6 for ; Mon, 22 Nov 2010 13:03:58 +0100 (CET) X-Spam-Checker-Version: SpamAssassin 3.3.1-spampd (2010-03-16) on mail.absint.com X-Spam-Level: Received: from absint.com (imp.absint.com [192.168.10.48]) by mail.absint.com (Postfix) with ESMTP for ; Mon, 22 Nov 2010 13:03:58 +0100 (CET) Received: by absint.com (Postfix, from userid 15028) id B942F1054BF; Mon, 22 Nov 2010 13:03:58 +0100 (CET) From: Christoph Cullmann Organization: AbsInt To: caml-list@yquem.inria.fr Subject: OCaml + mingw-w64 Date: Mon, 22 Nov 2010 13:03:58 +0100 User-Agent: KMail/1.13.2 (Linux/2.6.35-14-generic; KDE/4.4.2; x86_64; ; ) MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Message-Id: <201011221303.58505.cullmann@absint.de> X-Spam: no; 0.00; ocaml:01 ocaml:01 mingw:01 gdb:01 mingw:01 frisch:01 gcc:01 unices:01 bytecode:01 byterun:01 gesch:98 ferdinand:98 unix:01 compile:01 msvc:01 Hi, atm we compile out large ocaml application (Astree, http://www.astree.ens.f= r/)=20 with ocaml + MSVC64. I would like to go away from MSVC to mingw, to e.g. have gdb available and= =20 more unix like environment. Atm it seems only the normal mingw (32bit) is supported. I found some problem with using the 64bit version with flexlinker, which I= =20 reported to Alain Frisch to get that sorted out. Aside from that, I think a problem might be, that there is no matching=20 assembler for amd64 + mingw (windows calling convention). Is that correct and could somebody help me out with this? I think, given that gcc on unices with 64bit work fine here, it shouldn't b= e=20 too much effort to get mingw-w64 working too, enabling to have a nice (and = fast=20 for the bytecode interpreter) variant of ocaml for win64 using mostly open= =20 source components ;) Other feedback (which is valid for MSVC64, too): In ocaml 3.12, in byterun/major_gc.h, there is: intnat caml_major_collection_slice (long howmuch) whereas in the .c file that is: intnat caml_major_collection_slice (intnat howmuch) As intnat is with mingw-w64 long long, this doesn't match. Should not the header use intnat, too? Greetings Christoph =2D-=20 =2D------------------------------------- Christoph Cullmann --------- AbsInt Angewandte Informatik GmbH Email: cullmann@AbsInt.com Science Park 1 Tel: +49-681-38360-22 66123 Saarbr=FCcken Fax: +49-681-38360-20 GERMANY WWW: http://www.AbsInt.com =2D------------------------------------------------------------------- Gesch=E4ftsf=FChrung: Dr.-Ing. Christian Ferdinand Eingetragen im Handelsregister des Amtsgerichts Saarbr=FCcken, HRB 11234