Hi,
I am new to Amber and am struck by the rather non-canonical install procedure. It appears the working assumption is to run the package out of the source+build tree, with bin/ and lib/ created there. I'd rather separate those two, which its typically done by ./configure --prefix=… However, Amber's "./configure" script is decidedly different and not of the GNU autoconf variety.
I usually target an install dir that is separate from the source+build tree. For one thing, this permits different builds for libraries, optimizations, and so on, while providing the users with stable runtimes. Most packages are fine with a canonical workflow as follows:
./configure --prefix=/opt/soft/name-version-build [other options]
make 2>&1 | tee make.log
make install 2>&1 | tee install.log
Is there a way to coax Amber into this pattern or am I bumping against the old-school assumptions? For one thing, the auxiliary python gets hardwired with the source tree location.
[16:35] stern.login1:/home/share/soft/amber/amber11$ strings -a bin/python2.6 | grep -c $PWD
99
To be clear, /home/share/soft is my build repository, and /opt/soft my binary repository.
There's also catch-22 in that AmberTools wants to be unpacked into the same location as Amber, and built first, but it still requires $AMBERHOME.
Thanks,
Michael
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Nov 07 2011 - 15:30:02 PST
