boingGraph available in anarchy (was Re: compiling DynamicGraph-1.2)

[Alex Lancaster]

>>>>> "ST" == Sven N Thommesen <address@hidden> writes:

ST> At 09:13 AM 7/13/1998 -0700, Marcus wrote:
>>>>>>> "ST" == Sven N Thommesen <address@hidden> writes:
>>
ST> Re-compiling Swarm with -fno-inline did not do it,
>>  Turn off optimization entirely, then.

ST> And that seemed to do it, until I tried clicking on the
ST> 'ringDistribute' or the 'boingDistribute' buttons. Coredumped
ST> again.

ST> HOWEVER, inspecting the source code (GraphLib-1.2/DiGraph.m) I
ST> found that the two calls to 'pow(something, 0.5)' could well be
ST> replaced by 'sqrt(something)'. AND THAT DID IT.

ST> Suggest you amend the source for GraphLib accordingly.

ST> Cheers, --Sven

Regarding source to GraphLib:

Paul Girard <address@hidden> from SAIC has brought to our
attention a problem with the algorithm used to position the nodes on
the Canvas graph in GraphLib (of the two demos - this feature is
*only* used in DynamicGraph, and only affects the position on the
Canvas and not the underlying data structure).  We are currently
working on merging this change into GraphLib itself.  Paul Girard has
also kindly provided us with a corrected version that fixes the
incorrectly formulated algorithm available at:

 ftp://ftp.santafe.edu/pub/swarm/users-contrib/anarchy/boingGraph.tar.gz

I too, am loathe to modify a library to compensate for a compiler bug,
especially when it's using a call to a nice well-behaved, well-tested
library like math.h.  As has been pointed out by Paul et. al. it looks
like this particular issue with RedHat can be fixed by upgrading to:

 glibc-devel-2.0.7-13

My initial reaction was rather like Sven's, i.e. to just fix the
problem in GraphLib. However, the *good thing* about fixing these
kinds of problems at the level of the OS, is that it fixes the problem
for *all* apps that use `pow' or other such mathlib routines.

I would be a little concerned about publishing any simulation results
if I knew that there was a bug in such a fundamental library as
mathlib - since where there is one visible bug - there often lurk
three in hiding...

Regards,

 --- Alex

-- 
  Alex Lancaster         |   e-mail: address@hidden
  Swarm Developer        |      web: http://www.santafe.edu/~alex
  Santa Fe Institute     |      tel: +1-(505) 984-8800 (ext 242)
------------------------------------------------------------------

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