Solaris Binary Installation Report/Gripe

[Barry McMullin]

This is the first time I tried a binary installation,
so I thought I should share the experiences.

It was mainly positive - a significantly more
straightforward exercise then my previous experience
with installing source distributions. Problems were
relatively minor:

o I found the documentation for the binary
distribution a bit confusing: the README appears to
suggest I need to do a make regardless of whether the
distribution is binary or source. I suggest the README
in the binary distributions should be clarified!

o Having compiled heatbugs (what else!) I ran into the
standard run time problem of init.tcl not being found
in the "standard" places.  I knew this *was* a standard
problem, so I checked the FAQ but couldn't find
anything there; tried search on the Swarm site for
"init.tcl" - nothing; tried search on swarm-support for
"init.tcl" - nothing; finally tried search on
swarm-support for "init tcl" and got the answer at:

http://www.santafe.edu/projects/swarm/archive/list-archive.9706/0016.html

For the record, I needed to set (and export)
environment vars roughly as follows:

    TK_LIBRARY=~alife/swarm/swarm/lib
    export TK_LIBRARY
    TCL_LIBRARY=~alife/swarm/swarm/lib
    export TCL_LIBRARY

Now I expect this to be a fairly common problem with
binary distributions, because then, precisely, there is
no necessity for tcl/tk to be already installed, so
there's every possibility that init.tcl and its
brethern won't be in "standard" places (even if they
were, we probably prefer to grab the versions in the
binary distribution to avoid any possible version
conflict).  So I suggest this issue should be added to
the FAQ, and the README (at least in the binary
distributions).


But all in all, a pretty straightforward install.

==============================

<gripe>

OK that was the bouquet, now for the brickbats:

o Why are there no 1.0.2 binaries?

o Corollary (as I shall explain...): why does the
binary distribution involve a single big swarm.x that
is apparently *not* a library, but more like a single
big object file?

Let me explain.  I sometimes need to replace files in
the swarm distribution with my own variants - e.g. to
temporarily correct a bug, until the fix feeds into a
full release (but also sometimes to actually adjust the
functionality, though that is more nebulous).  If I
was working with a full source distribution, I could
simply replace that source and rebuild - but (as far as
I can tell - corrections welcome!) I can't do that with
a binary release.  Now I would have expected that I
could still have local variants of individual swarm
object files, explicitly linked into my apps, which
would override the versions in swarm.x.  But when I try
it I get something like this:

  gcc -g -L../../swarmhome/lib  -L/usr/openwin/lib  -o scl main.o 
SCLModelSwarm.o
  SCLDisplay.o CellEditor.o WorldManager.o LatticeAgent.o AgentManager.o 
ParticleM
  anager.o ParameterManager.o VNneighborhood.o MooreNeighborhood.o Lattice2d.o 
Par
  ticle.o Hole.o Substrate.o Link.o Catalyst.o Bond.o Version.o SCLGlobals.o 
PRNG.
  o Agent.o UserCellEditor.o Bernoulli.o FHPneighborhood.o Coord2d.o 
PickNeighborC
  ode.o  UniformDouble.o Raster.o ZoomRaster.o ../../swarmhome/lib/swarm.x 
-lobjc
  -lX11 -lm  -lsocket -ldl -lnsl
  ld: fatal: symbol `__objc_class_name_Raster' is multiply defined:
          (file Raster.o and file ../../swarmhome/lib/swarm.x);
  ld: fatal: symbol `_GLOBAL_.I.UniformDouble.m' is multiply defined:
          (file UniformDouble.o and file ../../swarmhome/lib/swarm.x);
  ld: fatal: symbol `_GLOBAL_.I.ZoomRaster.m' is multiply defined:
          (file ZoomRaster.o and file ../../swarmhome/lib/swarm.x);
  ld: fatal: symbol `__objc_class_name_UniformDouble' is multiply defined:
          (file UniformDouble.o and file ../../swarmhome/lib/swarm.x);
  ld: fatal: symbol `_GLOBAL_.I.Raster.m' is multiply defined:
          (file Raster.o and file ../../swarmhome/lib/swarm.x);
  ld: fatal: symbol `__objc_class_name_ZoomRaster' is multiply defined:
          (file ZoomRaster.o and file ../../swarmhome/lib/swarm.x);
  ld: fatal: File processing errors.  No output written to scl
  make: *** [scl] Error 1


(:-(

I wouldn't mind so much except that some of the
variants I'm trying to link in here are to fix things
that are already fixed in 1.0.2 ... except there is no
binary distribution for that yet...

OK, I know the real answer to this may be to do a
source install.  But I'd appreciate some insight into why
swarm.x isn't a searchable library rather than
(apparently) a fixed set of objects with no choices.

</gripe>
====================================

Having said that, I want to end by saying that I still
very much support the idea of binary distributions, and
they really can make swarm significantly more
accessible.

Cheers,

- Barry.

PS: What the heck *is* a "brickbat" anyway?

-- 

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
| Barry McMullin,                            |    address@hidden |
| School of Electronic Engineering,          |  Voice: +353-1-704-5432 |
| Dublin City University, Dublin 9, IRELAND. |  FAX:   +353-1-704-5508 |
| http://www.eeng.dcu.ie/~mcmullin           |                         |
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

                  ==================================
   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.
                  ==================================