Open eirnym opened 7 months ago
regress already has an extensive test suite which runs in its CI. Any difference you find between regress and EcmaScript syntax in Unicode mode (/u flag) is a bug.
Non-Unicode mode is still a WIP.
In addition to the internal tests, you can also look at the actual ECMAScript Test Suite tests ran on the boa
engine. We use regress
in boa
. There are some more things that are being tested there, since regress
only implements the regex engine and the ECMAScript tests sometimes have wider test boundaries.
You can take a look at those tests either in the test coverage on the boa website https://boajs.dev/conformance or via the independently run site https://test262.fyi where boa is compared to other engines like v8 and SpiderMonkey.
But to spare you from going trough those results I can already spoiler you. Except for 2 tests that are caused by the lack of Unicode 15.1 support, all relevent RegExp tests are passing. We actually wrote about this in our latest release notes: https://boajs.dev/blog/2024/03/07/boa-release-18#regexp
Based on that I think the claim in regress
stands on pretty solid ground ;)
The crate claims that
it targets EcmaScript (aka JavaScript) regular expression syntax
.Could you please add tests (with CI?) to test whole suite to know the exact difference between EcmaScript Regex and this library?