Two generated inventories are stale in components outside the test list

Table of Contents

This page is a capture in the inbox bucket of the product backlog — a pre-sprint idea, not yet pulled into a sprint as a story.

1. What

Two generated inventories are stale on main, in components that are not in COMPONENTS_UNDER_TEST:

  • regenerate_cmake_component_files.py --all --check reports ores.synthetic/core, ores.synthetic/service, ores.utility/src, ores.workspace/api, ores.workspace/core and ores.workspace/service stale. The change each wants is the generated marker and the reworded header comment the Qt-removal work left behind, not a missing file.
  • regenerate_physical_space_inventories.py --check reports projects/ores.codegen/library/templates/ores.doc.http-recipe.org stale.

None is dq's, and dq's own four CMake lists are current.

2. Why

A component outside COMPONENTS_UNDER_TEST is nobody's gate. That is deliberate — listing a component claims responsibility for regenerating it — but it means a shared template change can leave a component's generated output stale indefinitely, and the only thing that notices is a whole-tree check nobody runs per component. The stale lists were found while bringing dq to the standard, and fixing them here would pull three unrelated components into a dq branch.

The repair is to regenerate them in the branch that owns each component, or to add a gate that fails when the generator's output differs from the tree for any component, not only the listed ones.

3. References

  • projects/ores.codegen/scripts/regenerate_cmake_component_files.py
  • projects/ores.codegen/scripts/regenerate_physical_space_inventories.py
  • projects/ores.codegen/scripts/component_registry.py — the list that decides which components are checked.

4. See also

Emacs 29.3 (Org mode 9.6.15)