Symptom: a case setting patch_ib(k) for 54,000 < k <= 2,050,000 passes validation, then fails (or worse) at the Fortran namelist read — the array is dimensioned by the smaller constant.
Mechanism: toolchain/mfc/params/definitions.py:34 reads num_ib_patches_max (= 2,050,000, src/common/m_constants.fpp:30) as the patch_ib max index, while the namelist array is dimensioned num_ib_patches_max_namelist (= 54,000, m_constants.fpp:32).
Introduced: #1524 (16fed9bde, 2026-06-03, @danieljvickers) flipped the constant name from num_ib_patches_max_namelist to num_ib_patches_max; before that (#1378 onward) the toolchain pointed at the namelist constant correctly. Tagging @danieljvickers for context.
Fix: #1552 repoints the bound at num_ib_patches_max_namelist.
Symptom: a case setting
patch_ib(k)for 54,000 < k <= 2,050,000 passes validation, then fails (or worse) at the Fortran namelist read — the array is dimensioned by the smaller constant.Mechanism:
toolchain/mfc/params/definitions.py:34readsnum_ib_patches_max(= 2,050,000,src/common/m_constants.fpp:30) as thepatch_ibmax index, while the namelist array is dimensionednum_ib_patches_max_namelist(= 54,000,m_constants.fpp:32).Introduced: #1524 (
16fed9bde, 2026-06-03, @danieljvickers) flipped the constant name fromnum_ib_patches_max_namelisttonum_ib_patches_max; before that (#1378 onward) the toolchain pointed at the namelist constant correctly. Tagging @danieljvickers for context.Fix: #1552 repoints the bound at
num_ib_patches_max_namelist.