Le 16/12/2011 13:39, Alain Frisch a écrit :
> 3. Binary packages are not created by casual users. It's not crazy to
> require, at least in the short term, a decent Unix-like environment
> (which includes a C compiler) in order to compile the libraries and
> create the binary packages. It would be nice to adapt all the OCaml
> libraries around so that they don't rely on external Unix tools, but
> this is simply not going to happen.

Binary packages could be cross-compiled. This is what has been done for
Coq dependencies [1]. Our goal while doing this was just to produce a
working Windows version of Coq, but maybe the idea could be extended to
provide a full development environment as envisioned in your point 1
(self-contained OCaml toolchain, with no C compiler). The "decent
Unix-like" environment you mention would then be an actual existing
(foreign) one.

Cross-compiling raises the issue of testing of binary packages. But if
we can get a working (and reasonably maintainable) OCaml toolchain on
Windows, pure OCaml test-suites could be used to test them natively.

I don't know about others, but the cross-compiling way would be my best
choice: I don't really care about Windows itself, but I do care about
Windows users of my software :-)

[1] http://ocaml.debian.net/mingw32/


Cheers,

-- 
Stéphane

-- 
Caml-list mailing list.  Subscription management and archives:
https://sympa-roc.inria.fr/wws/info/caml-list
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
Bug reports: http://caml.inria.fr/bin/caml-bugs

Reply via email to