[pism/pism] 2afb5a: "std::isfinite()" is more appropriate than "not st...

[Constantine Khrulev]

Branch: refs/heads/dev
Home: https://github.com/pism/pism
Commit: 2afb5a5ee7eb63b4d54aad8b73641d92634ac506
https://github.com/pism/pism/commit/2afb5a5ee7eb63b4d54aad8b73641d92634ac506
Author: Constantine Khrulev <ckhr...@alaska.edu>
Date: 2021-11-17 (Wed, 17 Nov 2021)

Changed paths:
M src/geometry/grounded_cell_fraction.cc

Log Message:
-----------
"std::isfinite()" is more appropriate than "not std::isnan()"


Commit: 1137648170ce3d29b96433e84548abfe9a5c585c
https://github.com/pism/pism/commit/1137648170ce3d29b96433e84548abfe9a5c585c
Author: Constantine Khrulev <ckhr...@alaska.edu>
Date: 2021-11-18 (Thu, 18 Nov 2021)

Changed paths:
M src/coupler/ocean/Pico.cc

Log Message:
-----------
Work around a bug in PICO

This commit ensures that we don't divide by zero and so allows PISM to continue running,
but it *does not* fix the real bug: for some reason sometimes (depending on the number of
processes and the parallel domain decomposition) n_shelf_cells[s] is zero. This is not
right.


Compare: https://github.com/pism/pism/compare/19b400eccdfd...1137648170ce