On Sat, Aug 29, 2009, Ben Roberts wrote:
>
> >
> > I've also moved amber11/src/gleap/mortsrc/nablib to nablib.old. This is a
> > first step toward removing this subdirectory completely, since I think(?) it
> > is obsolete. If you are using code there, please let me know -- presumably
> > things will break if you are.
>
> Would that include the following response to a conventional build (well,
> attempted build) of AmberTools? And is the "objfun" code necessary for,
> presumably, LeAP?
>
> <snip>
> make[3]: Entering directory
> `/scr/wukong_2/roberts/apps/amber11/src/gleap/mortsrc/objfun'
> make[3]: *** No rule to make target `../nablib/sff.h', needed by
OK...I somehow missed the header dependency. The updated CVS tree should(!)
fix this. Thanks for the report.
Wei:
I wanted to remove unused code in nablib, and have a single, master version
of sff.h. So, I updated CVS to point to the sff.h file in amber11/src/sff.
But several points come up:
(1) You had added Scee and Scnb to the PARMSTRUCT_T structure, and these are
used in energee.cpp. But I don't really understand the code there. Is this
part of the update to allow different values of scee and scnb for different
torsions? Or is something else going on? I added these new variables to the
end of the structure (in amber11/src/sff/sff.h) rather than in the middle, as
you have it, but I'm not sure whether this will break anything in your code.
So you should look at the new sff.h and see if you think that will work.
(2) The test case in amber11/src/gleap/example/nab/runenergy.cpp won't
compile for me -- problem with the load_mdb() function. Does it work
for you? It doesn't look like anything I did should have changed the
compilation error...
(3) Are you using sff.h or libsff.a in any otherplaces than example/nab?
It looks like the code in mortsrc/enefrc is completely independent. Do you
have your own GB implementation?
(4) Obviously, if there is a good reason to keep the mortsrc/nablib directory,
we can undo the name change cited above. But we should have only a single
sff.h file, not two. I hope I didn't really break anything here.
(5) A number of the test cases in amber11/test/sleap are failing, but these
don't look(?) like things I did. Can you look at the tests there, and see
what is going on, updating if needed? Also, if we are going to support a mort
interface to sff, we should get that tested.
Hope things are going well for you...let me know. Thanks in advance for
looking at this....dac
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Wed Sep 02 2009 - 23:06:21 PDT
