OK I'm putting my hand up.
It seems to me that this thing should now be doable. So let's get it
done.
First I am no great shakes as a programmer. Hence I am not volunteering
to do serious code hacking.
However, I am prepared to do documentation and coordination work.
Below I set out a final objective and some ideas on process to get
there.
This draws heavily from the many helpful comments of others on this
list.
My apologies in advance for not acknowledging them specifically.
Please feel free to comment on or disagree with my approach.
Objective:
To develop a Swarm package that can be easily installed on MacOS X and
fully integrate with Apple's compiler and other development tools such
as
Project Builder (at least from the Jaguar release). OS X should be the
best environment for Swarm.
To me this implies having Tcl/Tk, BLT, hdf5 and Swarm as standard
Darwin/OS X frameworks which do not require XFree86.
Current state:
Workable gcc3.1 compilers can be built from Apple sources.
Tcl/Tk 8.3.x, BLT and hdf5 can all be compiled and installed on OS X
10.1.x in the standard UNIX fashion using XFree86 as the display. hdf5
appears to pass almost all of its test suite. The other packages have
various errors which may or may not be important, but they are
basically
functional.
Swarm code, as of the last snapshot, can be mostly compiled. There
are a
lot of patches needed, information about which is spread over a number
of
places. There are still problems particularly with avcall and possibly
other libraries.
Displacing XFree86 has a number of' show stopping' issues. Mainly the
current Aqua Tcl/Tk port uses 8.4aX sources. This code is flakey in
even
in its XFree86 builds and incompatible with current versions of BLT. It
does not seem to be worth pursuing this further until BLT catches up
and
this does not seem likely until after Tcl/Tk 8.4 (hopefully with OS X
support) is officially released.
How to move forward:
1. Make full information on how to build gcc3 and the other components
easily available so that new participants can get started more easily.
2. Maintain a depository of Swarm code containing the patches to date
which can be accessed by anyone willing to help.
3. Concentrate on getting the Swarm code to compile against the
existing
UNIX/XFree style libraries.
4. Once it compiles testing can start and bugs be nailed. At some
point
during this step the code becomes useful and might hopefully be
incorporated in standard Swarm snapshots.
5. At some advanced stage in the previous step, start work on a PB
project to compile Swarm as a framework.
5a. Compile hdf5 as a framework - this one looks quite simple.
6. Review the Tcl/Tk progress. There might be some point in building
an
Aqua Tcl/Tk 8.3 framework so that BLT can be built without waiting for
its
authors to update to 8.4.
7. Enjoy (No really)
Next a few Questions:
What is the intended function of the Swarm site on savannah.gnu.org?
Is the cvs there intended to be current?
If so, would it be possible to make a cvs branch for MacOS X so that the
current OS X code would always be available there?
Culd we use the Savannah facilities to run this project?
Comments, objections, improvements please.
Bill Northcott
School of Banking and Finance
UNSW
==================================
Swarm-Support is for discussion of the technical details of the day
to day usage of Swarm. For list administration needs (esp.
[un]subscribing), please send a message to <address@hidden>
with "help" in the body of the message.