From mboxrd@z Thu Jan 1 00:00:00 1970 Received: (from majordomo@localhost) by pauillac.inria.fr (8.7.6/8.7.3) id PAA32018; Wed, 26 Sep 2001 15:06:52 +0200 (MET DST) X-Authentication-Warning: pauillac.inria.fr: majordomo set sender to owner-caml-list@pauillac.inria.fr using -f Received: from nez-perce.inria.fr (nez-perce.inria.fr [192.93.2.78]) by pauillac.inria.fr (8.7.6/8.7.3) with ESMTP id PAA32015 for ; Wed, 26 Sep 2001 15:06:51 +0200 (MET DST) Received: from imo-r10.mx.aol.com (imo-r10.mx.aol.com [152.163.225.106]) by nez-perce.inria.fr (8.11.1/8.10.0) with ESMTP id f8QD6of00292 for ; Wed, 26 Sep 2001 15:06:50 +0200 (MET DST) Received: from CaptnJamesKirk@aol.com by imo-r10.mx.aol.com (mail_out_v31_r1.7.) id j.93.10ca38c3 (658); Wed, 26 Sep 2001 09:06:41 -0400 (EDT) From: CaptnJamesKirk@aol.com Message-ID: <93.10ca38c3.28e32ce0@aol.com> Date: Wed, 26 Sep 2001 09:06:40 EDT Subject: Re: [Caml-list] Initial port of ocaml for mingw (long) To: ayerkes@gmvnetwork.com CC: caml-list@inria.fr MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit X-Mailer: AOL 5.0 for Windows sub 40 Sender: owner-caml-list@pauillac.inria.fr Precedence: bulk In a message dated 9/25/2001 3:29:34 PM Central Daylight Time, ayerkes@gmvnetwork.com writes: > Actually, my thought was to build with cygwin itself at least in the > short term. The important thing (to me) is the ability to write nice > software and distribute it without the cygwin dll. It's a bonus too > that ocaml.exe works properly without it too, but the main thing was > the ability to run ocamlopt and get an exe out that you can pass > around easily. I built with cygwin's gcc (-mno-cygwin), tho. I think > that it may not be realistic to build ocaml otherwise given that it > uses unix tools to create the prims list, however it can easily be > used without cygwin. > > I should've made it more clear that you need cygwin gcc as yet to build. > The important point was that it's possible to get it away from > dependence on cygwin1.dll... After I fired off my last message, I started to wonder if that's what you did, so I tried it myself under cygwin. Copying libncurses.a to libpdcurses.a still works (though this may need to be changed to use the regular ncurses which is part of the standard cygwin distro). I ran into some other problems, and here's what happened. (first untarring the ocaml source, then patching, then untarring the boot stuff...) 1) make First break is at utils/misc.mli, where it says cannot open pervasives.cmi. The only thing that worked for me here was to... 2) make world which exits shortly with a cryptic (at least for me) error at [coldstart], but then 3) make seems to be happy until it gets to bigarray. It needs bigarray.cmi but doesn't have it since it doesn't have big_int.cmi. Making the 3 cmis you mention in otherlibs/num (int_misc.cmi, string_misc.cmi, and arith_flags.cmi) may be the first step, but it doesn't fix it. The only way I could build big_int.cmi was to cd to "otherlibs/num" and type "make big_int.cmi" there, after building nat.cmi as well. Two more tries indicated that "ratio.cmi" and "num.cmi" also have to be built in the "otherlibs/num" directory. THEN, we have to cd to "otherlibs/bigarray" and do "make bigarray.cmi" there. Then, 4) make goes all the way to ocamldebugger where it exits with "Uncaught exception: Not_found." I can't figure out how to fix it, so I do your next steps of "rm byterun/io.h", "make -C asmrun depend", and "make -C byterun depend". These last two have many warnings, most of which seem tied to redefinitions in fail.h. 5) make opt then says I need arith_status.cmi in otherlibs/num, but once that's built it finishes without further error. However, after doing "make install" and "make installopt" and trying to run ocaml.exe, I get "Cannot exec /usr/local/ocamlrun". In fact, all of the executables return that error. And now I don't know how to fix that... Oh, and as long as we're working on it, it would be VERY nice if we could get labltk to build as well. /John ------------------- Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/ To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr