[ET Trac] [Einstein Toolkit] #1651: Problems with expansions in parameter files

Einstein Toolkit trac-noreply at einsteintoolkit.org
Thu Sep 11 10:24:26 CDT 2014


#1651: Problems with expansions in parameter files
---------------------+------------------------------------------------------
  Reporter:  hinder  |       Owner:  sbrandt            
      Type:  defect  |      Status:  accepted           
  Priority:  major   |   Milestone:                     
 Component:  Cactus  |     Version:  development version
Resolution:          |    Keywords:                     
---------------------+------------------------------------------------------

Comment (by knarf):

 For what it is worth, most examples brought up use generic variable names
 'A', 'B', ect. - while in practice a reader would most often "correctly
 guess" from the variable name what is done here:
 {{{
 CarpetRegrid2::radius_1[1] = CoordBase::xmax / 2
 CarpetRegrid2::radius_1[1] = CarpetRegrid2::radius_1[2] / 2
 }}}
 strongly suggests that CoordBase::xmax is set somewhere in the file,
 simply because that is something you would expect a user to always set. I
 am not arguing that it would not be possible to construct confusing
 examples - but rather that most cases would "intuitively understood" by
 the context.

 I would suggest to leave the position of an assignment free, but to
 provide a facility in Cactus that, given a parameter file, does not run
 past the parameter file parser (especially does not allocate large
 memory), and simply outputs the parameter file with all expressions
 replaced. This way a review of a parameter file would be made much simpler
 - there would be no surprises and replacements can easily be checked even
 by the person sending the parameter file for review, creating even less
 work for the reviewer. For convenience it might even provide the "diff"
 between the original and processed parameter file as well.

-- 
Ticket URL: <https://trac.einsteintoolkit.org/ticket/1651#comment:18>
Einstein Toolkit <http://einsteintoolkit.org>
The Einstein Toolkit


More information about the Trac mailing list