From mboxrd@z Thu Jan  1 00:00:00 1970
Received: (from weis@localhost) by pauillac.inria.fr (8.7.6/8.7.3) id UAA01094 for caml-red; Mon, 21 Aug 2000 20:05:34 +0200 (MET DST)
Received: from concorde.inria.fr (concorde.inria.fr [192.93.2.39]) by pauillac.inria.fr (8.7.6/8.7.3) with ESMTP id LAA27714 for <caml-list@pauillac.inria.fr>; Mon, 21 Aug 2000 11:48:05 +0200 (MET DST)
Received: from pauillac.inria.fr (pauillac.inria.fr [128.93.11.35])
	by concorde.inria.fr (8.10.0/8.10.0) with ESMTP id e7L9fIr20245;
	Mon, 21 Aug 2000 11:41:18 +0200 (MET DST)
Received: (from xleroy@localhost) by pauillac.inria.fr (8.7.6/8.7.3) id LAA27638; Mon, 21 Aug 2000 11:41:18 +0200 (MET DST)
Message-ID: <20000821114118.58420@pauillac.inria.fr>
Date: Mon, 21 Aug 2000 11:41:18 +0200
From: Xavier Leroy <Xavier.Leroy@inria.fr>
To: Paul Stodghill <stodghil@cs.cornell.edu>, caml-list@inria.fr
Subject: Re: O'Caml 3.00 under Cygwin?
References: <yz18ztu8zup.fsf@MILHOUSE.cs.cornell.edu>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Mailer: Mutt 0.89.1
In-Reply-To: <yz18ztu8zup.fsf@MILHOUSE.cs.cornell.edu>; from Paul Stodghill on Fri, Aug 18, 2000 at 09:26:22AM -0400
Sender: weis@pauillac.inria.fr

> I have taken a first crack at getting O'Caml 3.00 to compile and run under
> Cygwin,

Funny you mention this, because I completed a Cygwin port of OCaml a
few weeks ago and merged it back in the working sources.
Interested parties can get the working sources from the CVS repository
(http://camlcvs.inria.fr/), although you should keep in mind that
this are alpha-quality code, not ready for release.

What works:
- "./configure; make world; make install".
- The native-code compiler.
- The replay debugger.
- Many of the external libraries (Unix, Str, Num, Dynlink, Bigarray).

What doesn't work yet:
- None of the threads library works; bytecode threads hang in select()
  apparently, and system threads don't compile because Cygwin doesn't
  implement a large enough subset of POSIX threads (e.g. they left out
  condition variables and thread-specific data).  Don't expect a fix soon.
- The configuration for LablTk fails to find Tcl/Tk, but this shouldn't
  be too hard to fix.
- No GUI for the toplevel.
- No Graphics library.

> I would be interested in comparing notes and patches with anyone who has
> tried this. I am particularly interested in getting the debugger to work
> reliably.

It took me a while, but I got it to work reliably (I think).
Initially, I attributed the crashes to bugs in the Cygwin socket code,
but actually Cygwin is innocent and this was just a text file vs
binary file issue (with rather strange consequences!).

- Xavier Leroy