Mozart application bundle for MacOs X

[Torsten Anders]

Dear Gustavo,

> I will build mozart-1-3-2 to make another bundle, but some questions:
> 1) Which is the state of mozart-1-3-2 with respect to the trunk?, i
> mean, there should be bug fixes in the trunk that are not yet in
> mozart-1-3-2.

Sure, all fixes since its release are missing in Mozart 1.3.2 (e.g., your fixes concerning inconsitencies in property 'platform' in MacOs X).

> 2) Which contributions should i include in the new bundle?. Mozart
> contributions are built with explicit parameters and may use some
> extra libraries that should be put inside the bundle.

Mozart compiles with some default contributions (--enable-contrib), perhaps we should use just these. You would then need GNU gdbm 1.8.3 and GNU regex 0.12. When compiling gdbm "by hand" I ran into a problem of its Makefile generated by configure. This problem is addressed and solved here: http://www.perlmonks.org/?node_id=573038. Also, on MacOS I ran into problems compiling micq (implementing an ICQ application in Mozart), and I doubt anybody will miss it if you leave it out with --disable-contrib-micq.

Best
Torsten

--
Torsten Anders
Interdisciplinary Centre for Computer Music Research
University of Plymouth
http://strasheela.sourceforge.net
http://www.torsten-anders.de
________________________________________
From: hackers-bounces at mozart-oz.org [hackers-bounces at mozart-oz.org] On Behalf Of Gustavo Gutierrez [ggutierrez at atlas.puj.edu.co]
Sent: 20 September 2007 14:35
To: Mozart hackers
Subject: Re: Mozart application bundle for MacOs X

On 9/20/07, Raphael Collet <raphael.collet at uclouvain.be> wrote:
> Gustavo Gutierrez wrote:
> > Another thing that should goes in the documentation is that mozart is
> > not universal yet. And we need to provide another application like
> > this to run in powerpc's.
>
> Having a universal bundle would be very nice!
>

This is something i thought some months ago, ad with a little help
from hackers people shouldn't be very difficult. Just some issues:

1) We need to know which parts (files) of the source code take into
account processor endianess. As many f you may know, powerpc
architectures are big endian and intel ones are little endian.

2) Universal versions  of all the libraries needed to build mozart,
for instance libgmp and libz. Since some of them are not universal,
the easier way to make universal versions is to build them in
different architectures and  to use lipo to combine them.

3) Modify the corresponding makefiles to include appropriate options
when building in powerpc or intel macs to produce a universal binary
of the emulator.

The difficult part of having mozart to build universal is (1). To
produce universal binaries is to build some native code, to
crosscompile some other and put both together.


> > Sure, i think the same, is that intended to represent that is build
> > from the trunk?
>
> Yes, 1.3.99 is by convention the main development trunk.  You should
> bundle the directory tagged mozart-1-3-2 in the svn repository, which is
> the latest release.

I will build mozart-1-3-2 to make another bundle, but some questions:
1) Which is the state of mozart-1-3-2 with respect to the trunk?, i
mean, there should be bug fixes in the trunk that are not yet in
mozart-1-3-2.

2) Which contributions should i include in the new bundle?. Mozart
contributions are built with explicit parameters and may use some
extra libraries that should be put inside the bundle.

Cheers,
--
Gustavo Gutierrez
_________________________________________________________________________________
mozart-hackers mailing list                           mozart-hackers at mozart-oz.org
http://www.mozart-oz.org/mailman/listinfo/mozart-hackers