regen-codegen

name	regen-codegen
description	Regenerates the openEHR spec + ITS layer from the vendored specs (emit / emit-xml / emit-rest) and checks for drift. Use after changing the openehr-codegen emitter or bumping a vendored spec, or when asked to regenerate the generated crates or run the codegen drift check.
allowed-tools	["Bash","Read"]
argument-hint	(none)

/regen-codegen

Regenerate everything openehr-codegen produces, in-session, and verify it is in sync with the vendored specs. Never hand-edit // @generated files — that is what this regenerates.

Steps

1. Regenerate all targets, in this order (the exact sequence scripts/check-codegen-drift.sh runs — spec crates first):

   cargo run -p openehr-codegen -- emit           # openehr-base/rm/am/term/lang (rm incl. src/model)
   cargo run -p openehr-codegen -- emit-xml        # openehr-its XML ToXml/FromXml
   cargo run -p openehr-codegen -- emit-json       # openehr-its canonical-JSON ToJson/FromJson codec
   cargo run -p openehr-codegen -- emit-rest       # openehr-its ITS-REST contract
   cargo run -p openehr-codegen -- emit-opt        # openehr-its opt14 (OPT 1.4 model + XML codec)
   cargo run -p openehr-codegen -- emit-rm-model   # openehr-rm src/model (static RM attribute/type model)
   cargo run -p openehr-codegen -- emit-validate   # openehr-rm src/validate/generated.rs (RM invariant cores)
   

   (emit already emits openehr-rm/src/model; emit-rm-model refreshes just that subtree and is byte-identical, so run order does not matter.)
2. Build + gate the affected crates:

   cargo build -q -p openehr-base -p openehr-rm -p openehr-am -p openehr-term -p openehr-lang -p openehr-its
   cargo clippy -q -p openehr-codegen -p openehr-its --all-targets
   cargo nextest run -p openehr-its        # fidelity gates: JSON round-trip + schema + XML round-trip + EHRbase XML
   

3. Drift check — the generated output must be a pure function of the specs + emitter, so a clean tree regenerates byte-identically:

   bash scripts/check-codegen-drift.sh
   

   (Also the CI codegen-drift job.) If it reports drift after your emitter change, that is expected — commit the emitter change and the regenerated output together.
4. Report what changed (which crates, file counts) and whether all gates + the drift check are green. If a generated diff appears with no emitter change, something hand-edited a @generated file — investigate, don't commit.

Notes

• Vendored inputs: BMM at tools/openehr-codegen/vendor/bmm/, XSD/OAS/JSON at crates/openehr-its/{schemas,vendor}/ (each with PROVENANCE.md).
• To change generated output, edit the emitter (tools/openehr-codegen/src/, the load/→analyze/→plan/→render/ pipeline stages) or a *_impl.rs sibling — see .claude/rules/codegen.md.