[sbryngelson]

What

s_cbc computes the characteristic boundary fluxes only when
recon_type == WENO. With MUSCL the routine copies the fluxes in and out
unchanged, so a characteristic BC (-5 to -12) silently behaves like
extrapolation (-3): a MUSCL run with bc = -5 or -6 is bit-identical to
the same run with bc = -3.

Prohibit the combination in check_muscl_simulation, sharing the
characteristic-BC scan with the existing IGR check, and add unit tests.

2D_shockdroplet_muscl and 3D_shockdroplet_muscl used bc_x = -6 with
MUSCL; switch them to -3, which they were already getting. Their
outputs are unchanged (bit-identical p_all and D on the 2D test case;
tests BE796E10 and 23573861 pass against the existing goldens).

Verification

Testing. Built with GPU offload (OpenMP, AMD) and ran the full test suite with this fix and seven other independent fixes applied together: 729 tests in 4 shards, 0 failures (two shards on MI300A, two on MI250X).

Prepared with Claude Code (AI-assisted); I reviewed the change and ran the tests.

[github-actions]

Claude Code Review

Head SHA: b85b0ad

Files changed:

• 4
• examples/2D_shockdroplet_muscl/case.py
• examples/3D_shockdroplet_muscl/case.py
• toolchain/mfc/case_validator.py
• toolchain/mfc/test_case_validator.py

Findings:
No issues found. Checked in step 4: toolchain/mfc/case_validator.py (the new check sits after the recon_type != 2 early return in check_muscl_simulation, and the helper keeps the IGR check's -12..-5 range), plus the two example case.py files and test_case_validator.py. I did not verify the claim that s_cbc handles WENO only, because the Fortran source is outside the review scope.