A. Environment:
System consists of Sun Ultra 10 with Solaris 7.
Compiler is fully patched Sun WorkshopPro 5.0.
OpenGL is ogl13_rt32_64.shar with the Sun patches 113886-20 and
113887-20 installed (order of installation is important).
It should be noted that manufacturer's OpenGL releases are reputed
to have code optimized for the manufacturer's machines. Supposedly,
on the Sun, some of the functions were coded in assembler. In contrast,
Mesa would not have such optimizations. Thus my insistence upon
use of the manufacturer's OpenGL release.
glut compiled from sources is 3.7 and gives libglut.a only because
the author does not wish to deal with the differences in shared
object compilations with different platforms.
python is 2.3.4 compiled with the modification to socketmodule.c
as noted on the sourceforge python bugtracker, item 972724 for lack
of AF_INET6 and INET_ADDRSTRLEN in the Solaris 7 headers. Note that
the suggestion for python 2.3.3c1, item 854823, does not produce
the same functionality. Numeric 23.3 is also present in this python.
tcl/tk is 8.3.5 with tix 8.1.4
X11, GL, and libglut.a are under /usr/openwin
python, tcl, and tk are under /usr/local
B. Installation:
Over several previous beta releases, installCode.py had compensated
for my use of python-2.3 by requesting the python version. With
this release, that functionality was lost. Instead, installCode.py
checked my python directory, correctly determined that i did not have
python-2.2, and requested a python directory. After breaking out of
this loop, I edited all references to python-2.2 to make them python-2.3.
With this edited installCode.py, the build went to completion.
Of course, as with the more recent beta releases, the command
source linkSharedObjs
is run in sh shell because that is the shell called by runCmds() with
Solaris 7. Consequently, it will not source. Perhaps converting
linkSharedObjs into an executable script with a #!/bin/sh at the top,
followed by chmod to u+x and use of ./linkSharedObjs the command to be
run instead of trying to source linkSharedObjs would work better.
Before I could run analysis, I manually changed into the directories
~/ccpnmr/ccpnmr1.0/python/ccpnmr/analysis, and
~/ccpnmr/ccpnmr1.0/python/ccpnmr/clouds, and executed
source linkSharedObjs
C. Test:
After this, I was able to run the tutorial. The problem of persistent
cursor images, first reported with b17 remains, along with a bouncing
or flickering of the contour plot probably an artifact produced by the
persistent cursor images. Regrettably, this makes analysis-1.0.1 unusable
for my various users. The rgb window assertion error inherited from python,
Assertion failed in module 'ffb_window.c' at line 271
!win_info->is_rgba
persists, and probably will until someone in python development decides
to correct this known and long documented python problem, one of the
consequences of which is an inability to use most of the colors listed
for contours.
An interesting sidelight upon running this version of analysis that I
have not seen previously is that it apparently alters keyboard assignments
so that after quitting analysis, the backspace key no longer deletes one
character to the left, but emits ^H (the backspace character, control H)
instead. Thus, the exit is not clean because it does not restore prior
settings.
Sincerely,
--
Bruce D. Ray, Ph.D.
Associate Scientist, and Operations Director
NMR Center
IUPUI
Physics Dept.
402 N. Blackford St.
Indianapolis, IN 46202-3273
|