#1595: Is *) allowed as upper boundary for a parameter range? -----------------------+---------------------------------------------------- Reporter: eschnett | Owner: sbrandt Type: defect | Status: assigned Priority: minor | Milestone: ET_2015_05 Component: Cactus | Version: development version Resolution: | Keywords: postrelease -----------------------+----------------------------------------------------
Comment (by rhaas):
Quoted strings are widely used in the ET. Even making it a warning to
use them will generate a ton of > output. I am not worried about quoted ''strings'' (in fact we advertise using quotes for them) I am worried about surrounding the ranges (which are a set of numbers) by quotes and we forbid surrounding numbers by quotes for parameter values, ie: {{{ Cactus::cctk_final_time = "10.0" }}} is invalid and produces {{{ WARNING level 0 from host t1650-403463.aei.mpg.de process 0 while executing schedule bin (none), routine (no thorn)::(no routine) in thorn Cactus, file tov.par:66: -> Invalid assignment: Attempting to set a variable of type REAL with (STRING)"10.0" }}}
I wasn't introducing ":", that was already supported and used. See, for
example, ADMMass, CartGrid3D, > etc. Oh joy. Those entries are not my favorite parameter declarations :-). We will have to continue to support it (and the usage is actually backed up by the documentation it seems).
Good point about the "::". That has to be avoided. However, it seems it
was caught earlier on. The
match never sees it.
Well I would suspect that the range of type: {{{ 1::2 "blah" }}} gets split into a range of '1' and a description of '2 "blah"'. Eg setting {{{ REAL xmin "Coordinate minimum in x-direction" { :: :: "Anything" } -1.0 }}} gives {{{ CST error 1: -> Description of range for parameter xmin has misplaced quotes ( :: "Anything") in param.ccl for thorn CartGrid3D }}}
The regex doesn't need to handle a single star or number, because that's
also handled by code prior to > this line. ok, just wanted to make sure.
I think we can make this an error instead of a warning. The ET passes
with the above expression. It always did (or at least has done so since we introduced the warning), Ian's concern seems to be for third party thorns.