[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: binary packages
From: |
Per Persson |
Subject: |
Re: binary packages |
Date: |
Wed, 18 Feb 2004 17:34:39 +0100 |
On Feb 15, 2004, at 07:14, Paul Kienzle wrote:
Mac OS X 10.2, 10.3:
Similar to windows, there are presently packages for fink,
for darwinports and standalone in various states of
repair. My own preference is again for a standalone
binary which doesn't depend on much else so that getting
someone up and running is simply a matter of dropping
the binary package somewhere in the Applications directory
and dragging it to the dock.
While I agree that a completely self-contained Octave.app that could be
drag'n drop installed anywhere the user choose would be the ultimate
Mac experience, I also think that it would be too much work right now.
FWIW, I think such a goal could be achieved in the future, and without
any changes to octave itself.
(If you're interested I have a screenshot of something I hacked up some
time ago (2001?) here http://homepage.mac.com/persquare/oxide1.jpg and
http://homepage.mac.com/persquare/oxide2.jpg )
So, for now I think that the best option is to provide a standard
installer that installs a octave, octave-forge, gnuplot and all other
deps.
It is also simple to provide an octave "wrapper", a shell script that
behaves like an application in that is has a custom icon, may be
dragged to the doc, can be installed anywhere etc. (Just append
.command to an executable script, and you'll see.)
Such an installer may also provide post-install options to set up
paths, editor etc.
(I think a lot of people would like to use BBEdit which works really
well)
I certainly wouldn't miss the opportunity to add AquaTerm to the mix;-)
(AquaTerm also supports both PGPLOT and PLplot)
I'd be able to help on packaging octave et al as long as we keep it a
reasonably long term goal.
Again a lot of work needs to be
done on sundries to make sure the OS X package can use
qhull, ginac, fftw, hdf5, etc.
Yes, I'm playing with hdf5 and fftw3 right now.
I don't know if we need separate
binaries for 10.2 and 10.3.
Most likely not.
Another decision is the level of architecture support we'd want. Simple
option is to generate code that runs on all ppc's.
Even if we had to differ between 10.2 and 10.3 and decide to support G4
and G5 extensions, the libs and executables may be created "fat" which,
if I understand correctly, means that they can host code for different
targets. (Even ix86 if necessary...)
I don't know if veclib is always
available and sufficient,
Available and sufficient as of 10.2
or if we need to support a variety
of Atlas-enhanced libraries.
No.
I don't know if there is a standard
place where the community can post such binaries, or if
octave-forge is the best place to host them.
There are several link sites, including Apples own third-party
listings, that would host info and links to a binary at sourceforge.
/Per
-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.
Octave's home on the web: http://www.octave.org
How to fund new projects: http://www.octave.org/funding.html
Subscription information: http://www.octave.org/archive.html
-------------------------------------------------------------
- binary packages, Paul Kienzle, 2004/02/15
- binary packages, John W. Eaton, 2004/02/16
- Re: binary packages,
Per Persson <=
- More on binary packages and FFTW/ATLAS, Quentin Spencer, 2004/02/19
- Re: More on binary packages and FFTW/ATLAS, Per Persson, 2004/02/19
- Re: More on binary packages and FFTW/ATLAS, Dirk Eddelbuettel, 2004/02/19
- Re: More on binary packages and FFTW/ATLAS, Dmitri A. Sergatskov, 2004/02/19
- Re: More on binary packages and FFTW/ATLAS, John W. Eaton, 2004/02/20
- Re: More on binary packages and FFTW/ATLAS, Per Persson, 2004/02/20
- Re: More on binary packages and FFTW/ATLAS, John W. Eaton, 2004/02/20
- Re: More on binary packages and FFTW/ATLAS, Paul Kienzle, 2004/02/20
- Re: More on binary packages and FFTW/ATLAS, Quentin Spencer, 2004/02/20
- Re: More on binary packages and FFTW/ATLAS, Dirk Eddelbuettel, 2004/02/20