problem compiling with FLOATS

[ggerbi]

I am trying to run a simulation with the cpp option FLOATS, and I get a compilation error. Has anyone else encountered this? This is my first time with FLOATS. Are there other cpp options that I should be using? I am using revision 291.

The error messages are:

PGF90-S-0038-Symbol, flt_lagran, has not been explicitly declared (step_floats.f90)
PGF90-S-0038-Symbol, flt_isobar, has not been explicitly declared (step_floats.f90)
PGF90-S-0038-Symbol, flt_geopot, has not been explicitly declared (step_floats.f90)

greg

[m.hadfield]

Try compiling the FLT_TEST case, first without, then with SOLVE3D defined. It is working OK for me with revision 297. If the FLT_TEST case works for you, have a look at the differences between it and your run.

[ggerbi]

Thanks.
FLT_TEST compiles without SOLVE3D, but not with SOLVE3D.
Does it work both ways for you? It doesn't seem to depend on the version I am using.

[m.hadfield]

FLT_TEST compiles without SOLVE3D, but not with SOLVE3D. Does it work both ways for you? It doesn't seem to depend on the version I am using.

Yes it does work both ways for me. Can you update to 297 and try again? (Or have you already done so?)

If that doesn't work, you need to look at the .f90 files (preprocessed source) to see where the relevant variables are declared (or not). Here are the results of a (case-insensitive) search for "flt_lagran" on my machine, with file names and line numbers followed by the line where the match was found

mod_scalars.f90(67): ! flt_Lagran: 3D Lagrangian floats
mod_scalars.f90(71): integer, parameter :: flt_Lagran = 1
step_floats.f90(135): IF (Ftype(l).eq.flt_Lagran) THEN
step_floats.f90(256): IF (Ftype(l).eq.flt_Lagran) THEN

[ggerbi]

Thanks for checking that.
I get the same error with version 297.
I'll dig deeper into my files.

greg