tree: dbde0a89d0084dc31e1158e59afab3cf96b1877d [path history] [tgz]
  1. fowler/
  2. anchored.toml
  3. bytes.toml
  4. crazy.toml
  5. crlf.toml
  6. earliest.toml
  7. empty.toml
  8. expensive.toml
  9. flags.toml
  10. iter.toml
  11. leftmost-all.toml
  12. line-terminator.toml
  13. misc.toml
  14. multiline.toml
  15. no-unicode.toml
  16. overlapping.toml
  17. README.md
  18. regex-lite.toml
  19. regression.toml
  20. set.toml
  21. substring.toml
  22. unicode.toml
  23. utf8.toml
  24. word-boundary-special.toml
  25. word-boundary.toml
vendor/regex-1.10.6/testdata/README.md

This directory contains a large suite of regex tests defined in a TOML format. They are used to drive tests in tests/lib.rs, regex-automata/tests/lib.rs and regex-lite/tests/lib.rs.

See the regex-test crate documentation for an explanation of the format and how it generates tests.

The basic idea here is that we have many different regex engines but generally one set of tests. We want to be able to run those tests (or most of them) on every engine. Prior to regex 1.9, we used to do this with a hodge podge soup of macros and a different test executable for each engine. It overall took a longer time to compile, was harder to maintain and it made the test definitions themselves less clear.

In regex 1.9, when we moved over to regex-automata, the situation got a lot worse because of an increase in the number of engines. So I devised an engine independent format for testing regex patterns and their semantics.

Note: the naming scheme used in these tests isn't terribly consistent. It would be great to fix that.