VMD-L Mailing List

From: Axel Kohlmeyer (akohlmey_at_cmm.chem.upenn.edu)
Date: Thu Nov 08 2007 - 06:44:28 CST

On Thu, 8 Nov 2007, Olaf Lenz wrote:

OL> -----BEGIN PGP SIGNED MESSAGE-----
OL> Hash: SHA1
OL>
OL> Hi!

hi olaf,

OL> As a developer of molfile reader plugins, I noticed two things, that
OL> make the task a bit tedious:
OL>
OL> 1. One problem occurs when I want to enhance a plugin that is already
OL> part of VMD (like, for example, to improve the vtf plugin). I have tried
OL> in various ways, but I was not able to load new version of the plugin
OL> into VMD. I even tried to delete the dynamical plugin from
OL> $VMDDIR/lib/..., but to no avail. Apparently, the "core" plugins are
OL> linked statically into the VMD code. Is that right? Or what did I do wrong?

you were probably loading the test files from the command line.
VMD uses a library of statically linked plugins and later
in the startup overrides them with dynamically loaded plugins
(if they have a higher minor version number). by the time the
load of newly compiled plugin commences, you are already loading
molecule files that are handed over at the command line.

solution: start vmd and then load your molecule with 'mol new'.
alternatively put this in a scripts

OL> 2. When I compile a C++-based plugin that uses exceptions, and I am
OL> using GCC 4 to compile it, the plugin is not compatible with the

arrrrgggghhhh, don't use exceptions. plugins are best written in
plain c. otherwise you will end up in the "dynamic linking hell"
that should be reserved for the developers of GCC and GLIBC.

cheers,
    axel.
OL> downloaded VMD binary version, which was apparently compiled with GCC 3.
OL> This is caused by a change of the ABI of exceptions between GCC 3 and 4
OL> (I think, the change might also have happened somewhere before).
OL> As compiling VMD manually is not really trivial, this actually makes it
OL> relatively hard to develop such plugins.
OL> Does anybody have a good idea or even know how to solve this problem?
OL>
OL> Best regards
OL> Olaf
OL> -----BEGIN PGP SIGNATURE-----
OL> Version: GnuPG v1.4.1 (GNU/Linux)
OL> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
OL>
OL> iD8DBQFHMslttQ3riQ3oo/oRAla7AJ0Sq67woG47FCL5BWsul+tbObth4QCbBDHZ
OL> ZR2gfP7AOYbHgzgwy8P8y8k=
OL> =O7Pj
OL> -----END PGP SIGNATURE-----
OL> 

-- 
=======================================================================
Axel Kohlmeyer   akohlmey_at_cmm.chem.upenn.edu   http://www.cmm.upenn.edu
   Center for Molecular Modeling   --   University of Pennsylvania
Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
tel: 1-215-898-1582,  fax: 1-215-573-6233,  office-tel: 1-215-898-5425
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.