[luther@debian.org: Re: ocaml 3.10.0 packages news]
Forwarding on behalf of Sven.
----- Forwarded message from Sven Luther <luther@debian.org> -----
Date: Fri, 15 Jun 2007 09:22:27 +0200
From: Sven Luther <luther@debian.org>
To: debian-ocaml-maint@lists.debian.org,
Stefano Zacchiroli <zack@debian.org>
Subject: Re: ocaml 3.10.0 packages news
On Thu, Jun 14, 2007 at 07:50:37PM +0100, Stefano Zacchiroli wrote:
> I worked another bit on 3.10.0 packages, the main changes can be
This is great work, thanks to all of you for having taken over the ocaml
packages so effectively.
> summarized as follows:
> - re-enabled building of the ocaml-source package (but with a new
> methods, which is hopefully more maintainable in the future)
> - ported patches to the stable 3.10.0 tarball
>
> An upload of ocaml 3.10.0-1 to experimental has been done yesterday.
I would have uploaded to unstable directly, but hey ...
> Open issues (i.e. request for comments from other debian camlers):
>
> - the ocaml-source package now contains ocaml sources *after* debian
> specific patches have been applied. Is that what we want? (My answer
> is yes, but one can argue that a vanilla source tarball is desirable)
I like this approach. I am unsure if using a vanilla source tarball can
cause compatiblity problems between our ocaml packages and the stuff
built from these tarballs, but in doubt, it is better to stay
consistent.
> - the ocaml-nox .deb is now about 30 Mb and installs something like 9 Mb
> of camlp4 related executables (sizes on i386 arch). Shall we split out
> an ocaml-camlp4-extras with the various kind of camlp4 executables?
>
> - ocamlbuild is installed upstream as ocamlbuild.byte /
> ocamlbuild.native / ocamlbuild (the latter being a *copy* of the best
> executable among the former two). This breaks the convention of foo /
> foo.opt of other legacy ocaml tools. I thus renamed ocamlbuild so that
> the plain version is bytecode (and is in ocaml-nox) and so that
> ocamlbuild.opt is the optimized version (and is in
> ocaml-native-compilers). Do you like my choice?
Maybe these two points make it better to go for a new handling of the
native/byte solution.
Maybe we should split out the .opt packages from ocaml-nox, and have a
provides like in the bytecode case for providing the best compilers on
each architecture in a transparent way.
This worked well for pure bytecode tools (like the spamoracle test
case), so why not apply it here too, especially since this is apparently
the way favored by upstream.
What do you think ?
> - assuming the experimental packages are fine (please test them no
> matter the open issues above!) are we ready to upload to unstable? (Of
> course pending an approval of the RMs)
We are very very early in the lenny release cycle, do we really need RM
approval now ?
Friendly,
Sven Luther
----- End forwarded message -----
--
Stefano Zacchiroli -*- PhD in Computer Science ............... now what?
zack@{cs.unibo.it,debian.org,bononia.it} -%- http://www.bononia.it/zack/
(15:56:48) Zack: e la demo dema ? /\ All one has to do is hit the
(15:57:15) Bac: no, la demo scema \/ right keys at the right time
Reply to: