[ET Trac] [Einstein Toolkit] #52: Ensure consistency between configurations on different systems

Einstein Toolkit trac-noreply at einsteintoolkit.org
Fri Aug 26 02:10:55 CDT 2011


#52: Ensure consistency between configurations on different systems
--------------------------+-------------------------------------------------
  Reporter:  anonymous    |       Owner:  mthomas
      Type:  enhancement  |      Status:  new    
  Priority:  minor        |   Milestone:         
 Component:  SimFactory   |     Version:         
Resolution:               |    Keywords:         
--------------------------+-------------------------------------------------

Comment (by hinder):

 At the moment I have configurations for both Damiana and Datura in the
 same source tree, but--as suggested I think by Erik--this would be better
 handled by having different source tree locations for the different
 machines (they share a head node and home filesystem).

 Most of the time I just use the default configuration name of "sim".  I
 would like this to use the default thornlist I have configured in
 defs.local.ini, which used to work but I believe is now broken (I should
 create a ticket for this).  However, I also use other configurations. For
 example I might have a small configuration for testing simfactory, another
 for testing Kranc examples, maybe one for Carpet-Git and another for
 Carpet-HG (I put the different Carpets in different arrangements in the
 same source tree).  So yes, I have multiple configurations.

 I think a mapping between thornlists and configurations makes sense, in
 the same way as I virtually always have a mapping between parameter files
 and simulations.  The only problem with this is that thornlist names tend
 to be long and the configuration name would have to be typed for every
 simfactory command.  This could be solved by supporting [http://www
 .debian-administration.org/articles/316 bash-completion] in simfactory so
 that you could tab-complete on the configuration name.  One could also use
 short names for thornlists instead of long names.

 So, I support the original suggestion and think it is a good idea.  It
 would be useful in the case where I have an old configuration and can't
 remember what thornlist I used for it.

-- 
Ticket URL: <https://trac.einsteintoolkit.org/ticket/52#comment:3>
Einstein Toolkit <http://einsteintoolkit.org>
The Einstein Toolkit


More information about the Trac mailing list