Opened 6 years ago

Closed 5 years ago

#4969 closed defect (fixed)

Issue with NFCeval.evalBinarySub

Reported by: casella Owned by: adrpo
Priority: high Milestone: 2.0.0
Component: New Instantiation Version:
Keywords: Cc:

Description

Please check ThermoPower.Examples.HRB.Simulators.ClosedLoopDigitalSimulator. The following error is reported:

[Compiler/NFFrontEnd/NFCeval.mo:559:9-559:67:writable] Error: 
Internal error NFCeval.evalBinarySub failed to evaluate ‘j - 1‘

The error is so bad that even the encoding of the stuff which cannot be evaluated is wrong...

Change History (8)

comment:1 Changed 6 years ago by sjoelund.se

The error is so bad that even the encoding of the stuff which cannot be evaluated is wrong...

I fixed the server to return encoding=utf8 for .err and .sim files. It just says:

‘j - 1‘

(using utf8 pretty quotes)

Last edited 6 years ago by sjoelund.se (previous) (diff)

comment:2 Changed 6 years ago by perost

  • Resolution set to fixed
  • Status changed from new to closed

Fixed in 82c686da. The model now seemingly fails due to #4968 instead.

comment:3 Changed 5 years ago by casella

  • Resolution fixed deleted
  • Status changed from closed to reopened

Another failure case here: Modelica.Magnetic.QuasiStatic.FundamentalWave.Examples.BasicMachines.InductionMachines.IMC_Characteristics:

[OMCompiler/Compiler/NFFrontEnd/NFCeval.mo:752:9-752:67:writable]
Error: Internal error NFCeval.evalBinarySub failed to evaluate k - 1

comment:4 Changed 5 years ago by perost

  • Owner changed from perost to adrpo
  • Status changed from reopened to assigned

The issue is that we try to generate equality constraint equations during the typing, before for-loops have been unrolled. And as part of that we try to evaluate the subscripts used in connect equations, which in the case of the FundamentalWave models include iterators that can't be evaluated at that stage.

comment:5 Changed 5 years ago by casella

  • Resolution set to fixed
  • Status changed from assigned to closed

This particular problem is now solved, the model fails for other reasons, possibly related to #5163

comment:6 Changed 5 years ago by casella

  • Resolution fixed deleted
  • Status changed from closed to reopened

comment:7 Changed 5 years ago by casella

As well as 16 models in Modelica.Magnetic.QuasiStatic.FundamentalWave.Examples.BasicMachines

comment:8 Changed 5 years ago by perost

  • Resolution set to fixed
  • Status changed from reopened to closed

This was fixed some time ago, all the QuasiStatic models now seem to work as well as with the old frontend.

Note: See TracTickets for help on using tickets.