You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I've always wondered why we don't do this already... is there a reason for it?
I'd also like to be able to read it from a directory defined in the input file.
I agree this would make more sense. The reason for the current system is that it's much more common to change the BOUT.inp than the mesh file: A set of simulations will all share the same mesh file, which can be in the root directory, with separate runs or restarted runs in subdirectories. I would however be happy to switch to the mesh file path being relative to the BOUT.inp file, so BOUT.inp could contain
[mesh]
file = "../some_mesh.nc"
to use subdirectories for separate runs.
It would be a breaking change, but probably not very difficult for people to adapt to.
One way to not make it a breaking change would be to first search (relative to?) the same directory as the input file, and then if not found, try the current working directory (or the other way round). That should probably not break things? Might be confusing?
Activity
mikekryjak commentedon Mar 4, 2026
I've always wondered why we don't do this already... is there a reason for it?
I'd also like to be able to read it from a directory defined in the input file.
bendudson commentedon Mar 4, 2026
I agree this would make more sense. The reason for the current system is that it's much more common to change the BOUT.inp than the mesh file: A set of simulations will all share the same mesh file, which can be in the root directory, with separate runs or restarted runs in subdirectories. I would however be happy to switch to the mesh file path being relative to the BOUT.inp file, so BOUT.inp could contain
to use subdirectories for separate runs.
It would be a breaking change, but probably not very difficult for people to adapt to.
ZedThree commentedon Mar 4, 2026
One way to not make it a breaking change would be to first search (relative to?) the same directory as the input file, and then if not found, try the current working directory (or the other way round). That should probably not break things? Might be confusing?
See also #3208
oparry-ukaea commentedon Mar 4, 2026
That makes sense to me. Searching cwd first would be guaranteed not to break anybody's current setup, presumably?