Open cindytsai opened 1 year ago
Thank you @cindytsai for reporting this For completeness, did you check all frontends or just reported the ones you already found to have this problem ? I'm not sure it's a real problem as far as the AMRVAC frontend is concerned because I don't know of any way to inspect AMRVAC data for an ongoing simulation, but I can still take a look and patch it to avoid a false positive
Hi @neutrinoceros, thanks for replying!
I have checked all the frontends, and only these frontends listed above have this problem.
They are not a real problem except Enzo in-memory frontend (IOHandlerInMemory
), cause the other frontends listed read data from files.
DataIO Not Allocating New Buffer in Some Frontends
Summary
Some frontends didn't allocate new buffer in function
_read_fluid_selection
(io.py
) when usingGridSelector
.When running this line, it writes changes to the view of the data array read in.
This is not a problem for post-processing, cause they create a data buffer for data read from file. But this is a serious problem for in-situ analysis (Enzo in-memory and libyt frontend), as the above line tries to alter the ongoing simulation data.
Frontends that didn't allocate new buffer
IOHandlerInMemory
(Get error)IOHandlerPacked2D
IOHandlerYTNonspatialhdf5
IOHandlerYTGridHDF5