Generate Category Test Data

generate-category-test-data regenerates test/unittest/data/re/CategoryContainsTestData.txt from the verified Unicode snapshot under tool/data/ucd. It calculates regex category expectations independently of the library’s category implementation and requires no Python interpreter.

Check or Regenerate the Fixture

Run these commands from the repository root:

./erbsland-core-tool generate-category-test-data --check
./erbsland-core-tool generate-category-test-data

--check (-c) fails if the file is missing or differs, without writing files or creating directories. Normal generation preserves equivalent files and atomically replaces changed output. The header records the Unicode version and the command that produced the file.

The output_directory setting is in tool/conf/tools/generate_category_test_data.elcl. Use --output-directory (-o) to override it for one invocation; the destination must remain inside the project. The shared [unicode] configuration selects the source directory and explicit version pin. Source integrity and version checks must succeed before output is published.

Regenerate this fixture when updating Unicode or changing regex category rules. Run CategoryContainsTest and the native CategoryTestDataTest suite, including its FullRun tests. See Category Test Generator Reference for the classification and sampling contract.