AMBER Archive (2009)

Subject: Re: [AMBER] Installation directory

From: David A. Case (case_at_biomaps.rutgers.edu)
Date: Thu May 21 2009 - 13:19:28 CDT

On Thu, May 21, 2009, Thomas Spraggins wrote:
>
>
> I want to build amber 10 in directory "a", but ultimately move it to
> directory "b". When I do the configure, config.h contains "a", and when
> I ultimately do the make, some of the files in the bin directory contain
> "a" (e.g., xleap and tleap, but also some of the non-human-readable
> binaries). It's easy enough to modify the scripts, but what about the
> binaries?

The config.h file has "location of the installation" at the top, and I think
you can edit this before compiling to get what you want. But I haven't
actually done this myself. [I'm assuming there is a good reason that you
cannot just configure and compile in directory "b" in the first place.]

The config_amber.h file has the same directory names, but now at the bottom
of the file. I'll change that for the next release to put it at the top.

[A lot of Amber developers think that these paths should be determined at run
time, via an environment variable. That would make it easier to move the
entire installation somewhere else, but would offer an extra way for things to
fail for more inexperienced users (i.e. if the environment variable was set
incorrectly, or not passed to all nodes of a cluster, etc.) It's still not
clear how we will handle this in future releases.]

...dac

_______________________________________________
AMBER mailing list
AMBER_at_ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber