Hi,
On Fri, 18 Aug 2006, David A. Case wrote:
> Hi everyone:
>
> I am finally facing up to a problem that has existed for a while now. As you
> may know, Junmei has been distributing a stand-alone version of antechamber
> (see http://amber.scripps.edu/antechamber), which he updates from time to time
> with bug fixes and with upgrades.
>
> The problem, of course, is that this version can get out of sync with what is
> in Amber 10. Worse, people like Scott can put fixes to antechamber in Amber10
> that don't get included in Junmei's sources, creating a horrible mess.
>
> It seems clear that the "antechamber" CVS tree should be the master one, and
> that amber10 should mirror that. So, I have modified the repository files
> for amber10: amber10/src/antechamber and amber10/dat/antechamber now point
> (in the repository, not in users' working directories) to the same place.
> Hence, fixes made by Junmei in the antechamber CVS will automagically appear
> in amber10; likewise, changes made in the Amber10 tree will show up in
> Junmei's CVS.
>
> This is an experiment with these two directories only. I don't think anyone
> should have problems, but let me know if you do. It is possible that you
> will have to remove the above two directories in your working version, and
> do a "cvs update -d" to repopulate them, but I don't think that this is
> necessary. If this works well, we might try the same thing with ptraj, where
It was necessary for src/antechamber but not for dat/antechamber.
See the tail if u care about cvs details.
> there is a similar divergence between between the local Utah copy and what is
> in the amber10 repository.
>
> Two fairly minor points remain here:
>
> 1. Scott: can you please re-commit your recent patches for the prepi/prepc
> file length problem? I could try to do a merge, but since you understand
> the changes, I think it would be much easier for you to do that.
Committed.
> 2. We still have to get a good mechanism by which bugfixes to antechamber
> show up on the web page bugfixes for Amber 9. Junmei is learning how to
> do this, but others need to think about related problems: if you put code
> into amber10 that really should be a bugfix (not a functionality upgrade),
> please try to get bugfixes created. I'll try to post on the wiki detailed
> instructions (unless someone else beats me to it.)
We could use a script to create bugfix.all because there are a couple
of easily overlooked differences between bugfix.all and bugfix.x.
The clever part will be getting the file paths correct in bugfix.all
because there may be several files in the amber tree with the same name.
I have an idea, but dont plan to work on this until after SF ACS.
(Still even a script that eliminates some but maybe not all of the tedium
would help).
> Obviously, if you think this is all a bad idea, or that there is a better way
> to proceed, please speak up.
Having multiple CVS repositories for the same software is a pain.
If this works then maybe we can try something similar for DOCK.
Scott
>
> ...thanks...dac
>
Aug 18 2:57:40pm loyd.scripps.edu 983> /thr/loyd/sbrozell/10/src/antechamber cvsup
cvs update: Updating .
cvs update: GPL is no longer in the repository
cvs update: LICENSE is no longer in the repository
U Makefile
U ac.c
U alc.c
U am1bcc.c
U antechamber.c
U aromatic.c
U atom.h
U atomtype.c
U bondtype.C
U charge.c
U charmm.c
U charmmgen.c
cvs [update aborted]: could not find desired version 9.1 in /thr/loyd/case/cvsroot/amber10/src/antechamber/common.c,v
cvsdh prep.c parmchk.c
Index: prep.c
===================================================================
RCS file: /thr/loyd/case/cvsroot/amber10/src/antechamber/prep.c,v
retrieving revision 1.4
retrieving revision 9.2
cvs [diff aborted]: could not find desired version 9.2 in /thr/loyd/case/cvsroot
/amber10/src/antechamber/prep.c,v
Received on Sun Aug 20 2006 - 06:07:22 PDT
