VMD-L Mailing List

From: Axel Kohlmeyer (akohlmey_at_cmm.chem.upenn.edu)
Date: Sun May 17 2009 - 08:37:47 CDT

On Sat, 2009-05-16 at 23:00 -0400, Cameron Mura wrote:
> Hi John et al.,

hi cameron,

[...]

> no problem). What I don't get is that there's not a $VMDDEV/lib dir,
> even in a build tree pulled fresh from CVS... This is unlike the, say,
> pre-bundled 1.8.6 src, which contains

the pre-bundled VMD packages are assembled by make_distrib.

the lib directory is where all the custom compiled binaries and
libraries go. you have to create it and populate it with your binaries.
since you want to compile against native libraries, you can omit a lot
of those. to make the compilation work, you have two options:
1) populate the lib directory with symbolic links, that emulate the
structure of the lib tree like john uses it.
2) change the paths in configure (or in the created src/Makefile) to
reflect your installation.

> lib/{cave,fltk,Mesa,points,scripts,stride,surf,tachyon} ?... Of

you usally don't need cave, Mesa, points

> course the main build attempt (do 'make' in $VMDDEV/src/) then fails,
> as it doesn't know where to find the tachyon header files, etc.:
> > $] make veryclean && make
> > Compiling OpenGLDisplayDevice.C --> OpenGLDisplayDevice.o ...
> > Compiling OpenGLExtensions.C --> OpenGLExtensions.o ...
> > ...snip...
> > ...snip...
> > Compiling IMDSimThread.C --> IMDSimThread.o ...
> > Compiling CmdIMD.C --> CmdIMD.o ...
> > Compiling imd.C --> imd.o ...
> > Compiling LibTachyonDisplayDevice.C -->
> > LibTachyonDisplayDevice.o ...
> > In file included from LibTachyonDisplayDevice.C:27:
> > LibTachyonDisplayDevice.h:28:58: error: tachyon.h: No such file or
> > directory
> > LibTachyonDisplayDevice.C:35:2: error: #error
> > "LibTachyonDisplayDevice requires Tachyon version 0.98.4 or higher."
> > ....
> > ....
> probably missing something really simple ?

it is one of these things that is simple, once you get the hang of it.
i remember being similarly confused when i started building VMD off the
cvs, but with a little bit of experience in compiling codes that don't
use the more common GNU automake/autoconf toolchain, it is
straightforward.

i'm in the process of updating my vmd.spec files for building VMD as
rpm package under fedora and one of my machines is a f10 machine
with x86 architecture. perhaps that would be an alternative?

you may also need some patches to compile stride and actc under
a gcc-4.x compiler and the current cvs needs a tachyon snapshot
installation (actually, the snapshot that supports outline rendering
is not available yet and you need to comment out one line of code
to compile VMD).

i'll be depositing my stuff in the VMD biocore as soon it is finished.

cheers,
   axel.

> thanks again for any advice,
> cam
>
> 

-- 
=======================================================================
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.