Re: xl, xxl, and mxl

--===============1017754969==
Content-Type: multipart/signed; boundary="=-=-=";
	micalg=pgp-sha1; protocol="application/pgp-signature"

--=-=-=
Content-Transfer-Encoding: quoted-printable


Hi Jeremy... and sorry for the late reply....

Jeremy Shaw <[email protected]> writes:

> Thanks for the detailed information, I think I now understand what you
> are trying to do.
>
> How easy will it be to cross the boundry between the DSL and scheme?
> For example, in this fragment:
>
> (c-library-rule libfoo (src foo.c bar.c) (flags -O3 -g))
>
> Would I be able to replace (src foo.c bar.c) with an arbitrary scheme
> expression that generated (src foo.c bar.c) ? Or would I be limited
> the primatives defined in the DSL ?
>

Not at all. One of the points of Conjure is that is build specs are
pure Scheme, and the ability to use arbitrary expressions almost
everywhere one of its main strengths (as i see it)

>
> The things I specifically have issue with are:
>
> (1) doing macro/template type stuff requires the use of a third party
>     macro language (m4/cpp/etc), which has different syntax, etc.
>
> (2) The use of whitespace seperated lists and no escaping mechanism
>     makes working with filenames that contains spaces very annoying.
>
> (3) The lack of safety features:=20
>
>     (a) very weak type system (everything is a string or list of
>     strings) does not allow you to say what you mean very accurately.
>
>     (b) no way to check for use of unitialized variables (important
>     for catching typos, and other common problems).
>

Absolutely.

> If I really wanted a better make, I think I would probably explore the
> possibilty of fixing the existing make before attempting to write a
> competing product.
>

From=20my point of view, doing so would only add more cruft to the
already kludgy means of abstraction that Make currently uses
(m4,shell...). I think it's better to start with a language powerful
enough...=20

> In any case, I think your DSL and xl could have a lot in common, so it
> should be interesting to see the results of making different choices
> along the way.
>

Sure. If i'm understanding you correctly, we ultimately have the same
goals, although we differ in the choice of language (xl vs. scheme).

jao
=2D-=20
If at first you don't succeed, try and try and try again. And then give up.
There's no point making a damn fool out of yourself.
		-- Dilbert's Rules of Work, slightly paraphrased

--=-=-=
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)

iD8DBQBBXGa3lX2PbVNDo+wRAhTtAJ9gO5nZFoL+8BlRyAp72ll5a6OofACcCKux
y9+gTyriD/7+SK/rtNQgnsI=
=eZdq
-----END PGP SIGNATURE-----
--=-=-=--



--===============1017754969==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

_______________________________________________
Pika-dev mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/pika-dev

--===============1017754969==--