[ET Trac] [Einstein Toolkit] #2003: piraha_everywhere Zelmani stress test

[Einstein Toolkit]

#2003: piraha_everywhere Zelmani stress test
---------------------+------------------------------------------------------
  Reporter:  rhaas   |       Owner:  sbrandt            
      Type:  defect  |      Status:  assigned           
  Priority:  major   |   Milestone:                     
 Component:  Cactus  |     Version:  development version
Resolution:          |    Keywords:  piraha             
---------------------+------------------------------------------------------

Comment (by rhaas):

 yes the fixed ones are all passing since I added private at the beginning
 of each one. Please try the "original.tar.gz" ones for errors.

 The issue is that piraha fails to parse PNSHelper/param.ccl in the
 original.tar.gz archive unless one adds a "private" at the beginning. One
 gets
 {{{
 File:
 /home/rhaas/postdoc/gr/cactus/Zelmani/arrangements/Zelmani/PNSHelper/param.ccl;
 par key error:(PNSHELPER PRIVATE variables) new=(solve_after_recovery
 verbose
 update_GR_every
 update_GR_every_start
 extrapolation_order
 update_GR_switch_time
 nrad
 nrad_outer
 ntheta
 nphi
 rad_max
 rad_max_outer
 symm
 collect
 interpolator
 interpolator_options
 coordinate_system) old=() at
 /home/rhaas/postdoc/gr/cactus/Zelmani/lib/sbin/parameter_parser.pl line
 800.
         main::parse_param_ccl("PNSHelper", Group=HASH(0x55a3a81f6a08),
 "BOOLEAN solve_after_recovery \"solve for constraints after rec"..., "{",
 "} \"no\"", "BOOLEAN verbose \"Do you want me to talk?\"
 STEERABLE=ALWAYS", "{", "} \"no\"", ...) called at
 /home/rhaas/postdoc/gr/cactus/Zelmani/lib/sbin/parameter_parser.pl line 62
         main::create_parameter_database("HDF5",
 "/home/rhaas/postdoc/gr/cactus/Zelmani/arrangemernalLi"...,
 "ZelmaniWizard",
 "/home/rhaas/postdoc/gr/cactus/Zelmani/arrangements/Zelmani/Ze"...,olate2",
 "/home/rhaas/postdoc/gr/cactus/Zelmani/arrangements/Llama/Inte"...,
 "CarpetEvolutionMasme/rhaas/postdoc/gr/cactus/Zelmani/arrangements/Carpet/Car"...,
 ...) called at /home/rhaas/postdctus/Zelmani/lib/sbin/CST line 140
 }}}
 whatever that actual error inside of piraha is (maybe the message lets you
 guess), having to add a private to fix it does not seem the correct thing
 to do.

 I also noted that spelling "LANG" as "LABG" in a schedule.ccl gives me
 {{{
 CST error 1:
   -> Unrecognised statement in schedule block (decloop) in schedule.ccl
 for thorn SyncG2Test
 ""
 }}}

 The second part of the issue is that currently piraha's error messages are
 not yet as good as they could be I think.

 Generally I think the error messages should always list the line number in
 the ccl file in which the error occurred and not (at least eventually)
 output as much internals of the parser that are not very useful to an "end
 user".

-- 
Ticket URL: <https://trac.einsteintoolkit.org/ticket/2003#comment:4>
Einstein Toolkit <http://einsteintoolkit.org>
The Einstein Toolkit