Test262 is the implementation conformance test suite for the latest drafts (or most recent published edition) of the following Ecma specifications:
The goal of Test262 is to provide test material that covers every observable behavior specified in the ECMA-414 Standards Suite. Development of Test262 is an on-going process. As of October 2017, Test262 consisted of over 29272 individual test files covering the majority of the pseudo-code algorithms and grammar productions defined in the ECMA-414 Standards Suite. Each of these files contains one or more distinct test cases. This marks the most comprehensive ECMAScript test suite to date. While test coverage is broad, TC39 does not consider coverage to be complete and as with any software project there exists the possibility of omissions and errors. This project welcomes any contributions to Test262 that help make test coverage of existing features more comprehensive.
As defined in the TC39 Process, Stage 4 Entrance Criteria requires tests for new feature proposals to advance. Tests may be written by proposal champions, implementers, or any interested community member.
A proposal champion is someone that worked on the feature proposal and specification directly.
A community member is you, and we welcome you to contribute! If you're having trouble getting started, or even just want to ask a question, feel free to open an issue.
Guidance for contributing to Test262 can be found in CONTRIBUTING.md.
Authors of contributions from non-Ecma member organizations must sign the Test262 CLA
Guidance for running Test262 and explanations of how a test file is interpreted by a test runner is in INTERPRETING
Volunteer-maintained projects that may be used to execute Test262 in various ECMAScript hosts:
Test262 runs CI tests against every PR and commit. The only tests that are required to pass are visibly flagged as REQUIRED. The CI test results that are attributed to specific runs against specific engines should not be perceived as meaningful to anyone but the person that is reviewing the test material contained within the contributed changeset. These tests are almost always expected to fail, especially in the case of tests for new features. They may be helpful in determining whether or not a regression occurred, but that can only be determined by an actual human reviewing the results and comparing those outcomes to the expected outcomes of the tests.
It‘s been removed. If you need to access the code that contained in that directory, we’ve preserved it in a branch, available here.