[sbryngelson]

Developed with AI assistance (Claude Code).

Contribution Policy

All contributions are expected to demonstrate:

• A clear understanding of the codebase
• Alignment with product direction
• Thoughtful reasoning behind changes
• Evidence of real-world usage or hands-on experience with the problem

Summary

Small fixes and tooling found while preparing #2116 and the AMR PR, split out so each can be reviewed on its own.

• bc_type dummy shape (bug): bc_type is allocated (1:num_dims, 1:2) but simulation/m_data_output.fpp declared
  the dummies (1:num_dims, -1:1). Now (1:num_dims, 1:2) everywhere.
• Serial boundary files every save: serial readers (post_process, restart) need bc_type.dat and bc_buffers.dat in
  each step directory, not only step 0. s_write_serial_boundary_condition_files now takes the packed buffers, so pre and
  sim share it.
• format.sh: ffmt reports unmatched block structure as a warning and exits 0; the script now shows stderr and fails
  on those warnings (the source would not compile).
• Cray debug builds get -h bounds, as the other compilers' debug builds already check bounds.
• Run template: passes --oversubscribe to Open MPI when ranks exceed cores (small CI runners refused -np 4).
• Source lint: rejects per-executable preprocessor guards (MFC_PRE_PROCESS, ...), which the sources no longer use.
• Validator: derives cfl_dt from cfl_adap_dt/cfl_const_dt like the Fortran; rejects MHD with Euler bubbles and a
  non-positive rburn%pign.

Verification

Frontier amdflang OpenMP full regression (as part of the branch these came from): all master goldens unchanged. CI runs
here on this branch alone.

[github-actions]

Lines of Code

File	Lines	Diff
src/common/m_boundary_io.fpp	319	-3
src/simulation/m_data_output.fpp	1581	+2
src/pre_process/m_data_output.fpp	677	+1

Directory	Lines	Diff
common	10541	-3
pre_process	5023	+1
simulation	28582	+2
total	47630	+0

[github-actions]

Claude Code Review

Head SHA: 7526aaf

Files changed:

• 11
• cmake/GPU.cmake
• docs/documentation/contributing.md
• src/common/m_boundary_io.fpp
• src/pre_process/m_data_output.fpp
• src/simulation/m_data_output.fpp
• toolchain/bootstrap/format.sh
• toolchain/mfc/case_validator.py
• toolchain/mfc/lint_source.py
• toolchain/mfc/test_lint_source.py
• toolchain/templates/default.mako

Findings:

• medium cmake/GPU.cmake: The Cray Debug branch adds "SHELL: -h bounds". As far as I know, -h bounds is a Cray C/C++ option, and the Cray Fortran compiler (ftn) enables bounds checking with -R b. If ftn rejects or ignores -h bounds, Cray Debug builds will fail or silently get no bounds checking, which contradicts the new comment. I am not certain, because I could not check the CCE version in use. Please check against the Cray Fortran documentation, and use -R b if needed.
• medium toolchain/templates/default.mako: --oversubscribe is now always passed whenever the launcher is Open MPI, even when ranks <= available cores. This is harmless for correctness. However, nproc reports the cores allowed by affinity or cgroup, which can differ from the slot count Open MPI uses. The warning is only printed when nproc is smaller than the rank count, so it may be misleading under some schedulers. Confidence is low, and this is a minor point.