IOSS reader fails with different multifile datasets with same name
The Exodus reader will read two datasets, with the same name, but different file counts. For instance, using can.exo, it would properly read can.e.4.[0-3] if can.e.2.[0-1] were in the same directory. This can easily happen as an analyst is ramping up the size of a simulation.
The IOSS reader fails under these conditions.
Super easy way to replicate.
- Copy g1s1.* into a sandbox directory.
- cp g1s1-10fps.e.16.00 g1s1-10fps.e.2.0
- cp g1s1-10fps.e.16.01 g1s1-10fps.e.2.1
- master, Linux, builtin server.
- Open g1s1-10fps.e.16. It will fail as follows:
ERROR: In /snip/5.10.0-alpha3/ParaView5/VTK/IO/Ioss/vtkIossReader.cxx, line 611 vtkIossReader (0x5422ce0): Filenames specified use inconsistent naming schemes. 'g1s1-10fps.e.2.0' has incorrect processor-count (2), '16' was expected. ERROR: In /snip/5.10.0-alpha3/ParaView5/VTK/Common/ExecutionModel/vtkExecutive.cxx, line 753 vtkPVCompositeDataPipeline (0x5663550): Algorithm vtkIossReader(0x5422ce0) returned failure for request: vtkInformation (0x5210d30) Debug: Off Modified Time: 385946 Reference Count: 1 Registered Events: (none) Request: REQUEST_DATA FROM_OUTPUT_PORT: 0 ALGORITHM_AFTER_FORWARD: 1 FORWARD_DIRECTION: 0