Patch Workflow and Next Steps

Patch Workflow and Next Steps

Patch workflow

Continue using one small Python patch per logical change.

Each patch should:

- support --dry-run and --apply
- create timestamped backups under .backups/YYYYMMDD_HHMMSS/
- fail loudly when anchors are missing or ambiguous
- be idempotent where practical
- print a clear summary of changed/skipped files
- list build and validation next steps
  1. Keep Patch 019 profiling updates.
  2. Keep Patch 020 combined Cantera thermo-sync optimization.
  3. Inspect git diff.
  4. Build release and debug.
  5. Run energy-disabled, constant-property energy, Cantera thermo without species, and Cantera thermo with species cases.
  6. Compare profiling output before/after thermo-sync optimization.
  7. Only then proceed to a radiation coupling stub, manufactured-qrad patch, or lower-risk diagnostics such as Cantera cache hit/miss counters.

Current design decision

The solver uses Option A.

When species transport changes Y before energy transport, preserve transported enthalpy:

h_after_species = h_before_species
T_after_species = T(h_after_species, Y_new, p0)

Do not reset h from the old temperature after species changes:

do not use: h = h(T_old, Y_new, p0)

The current Cantera thermo-sync optimization preserves this convention. It may combine temperature recovery and property refresh into one Cantera pass, but it must not alter the transported h field.