- Re-organize existing tests for identifiers
Name files according to their content (not their ES5 ID). Move tests
for IdentifierName and LabelIdentifier to more appropriate directories.
- Simplify and correct tests
Instead of asserting successful runtime evaluation using `eval`, rely
on the test runner's ability to detect syntax errors. Update the test
bodies to test the grammar referenced by their ES5 ID and
description--the IdentifierStart pattern.
- Use `negative` frontmatter to assert SyntaxErrors
- Remove redundant tests
- Use `assert.equal` helper function
- Add equivalent tests for literal unicode chars
- Add tests for variable-length unicode escape
The `test/language/generators/` directory contained a single file that
concerned generators derived both from syntactic form and from a
built-in function.
Refactor this test into two files and place each in the appropriate
directory.
- 15.10.2.15-3-1 and 15.10.2.15-3-2 are both invalid (U+002F (SOLIDUS) instead of U+005C (REVERSE SOLIDUS) was used to start an escape sequence).
- 15.10.4.1-2 and 15.10.2.2-1 are identical tests, delete the latter.
- Update tests to avoid using runTestCase() function.
- Update multiple test infos to point to correct algorithm step (CharacterRange step 6 instead of step 1).
Uniformly and exhaustively test the generic aspects of the static- and
instance Promise methods, including:
- function `length` properties
- function `name` properties
- method property descriptors
Some tests involving the directive prologue are invalidated by source
text transformations that insert executable code in the beginning of the
script. Implement a `raw` flag that allows these tests to opt-out of
this transformation. Update the relevant tests to use this flag (and
remove references to globals only available when code is injected).
Update the Python runner accordingly:
- Do not run tests marked as "raw" in strict mode
- Reject invalid test configurations
Update the browser runner accordingly:
- Do not modify the script body of tests marked as "raw"
Including:
- 21.2.4.2 get RegExp [ @@species ]
- 21.2.5.3 get RegExp.prototype.flags
- 21.2.5.4 get RegExp.prototype.global
- 21.2.5.5 get RegExp.prototype.ignoreCase
- 21.2.5.7 get RegExp.prototype.multiline
- 21.2.5.10 get RegExp.prototype.source
- 21.2.5.12 get RegExp.prototype.sticky
- 21.2.5.15 get RegExp.prototype.unicode
- 22.1.2.5 get Array [ @@species ]
- 23.1.2.2 get Map [ @@species ]
- 23.2.2.2 get Set [ @@species ]
- 24.1.3.3 get ArrayBuffer [ @@species ]
- 24.1.4.1 get ArrayBuffer.prototype.byteLength
- 24.2.4.1 get DataView.prototype.buffer
- 24.2.4.2 get DataView.prototype.byteLength
- 24.2.4.3 get DataView.prototype.byteOffset
- 25.4.4.6 get Promise [ @@species ]
- B.2.2.1.1 get Object.prototype.__proto__
- B.2.2.1.2 set Object.prototype.__proto__
Skipped TypedArray tests in order to avoid an early test strategy definition for them.
The "monkeyYaml" parser is intended to serve as a lightweight fallback
to Python's standard YAML parser in contexts where the latter is not
available. Any intentionally-simplified implementation will necessarily
exhibit non-standard behavior for different input, so not all input
accepted by the standard parser will be accepted by "monkeyYaml". If
loaded exclusively in fallback situations, these edge cases can only be
identified (and debugged) in the environments that require the fallback.
This has allowed developers to unknowingly author tests that cause
errors.
Update the test runner to use "monkeyYaml" in all cases, ensuring more
consistent behavior across contexts and precluding this class of
regression.
In ECMAScript 5, assignment to a non-reference value throws a runtime
ReferenceError. ECMAscript 6 specifies an early ReferenceError in these
cases. Tests for this behavior have been authored to pass in both cases.
Simplify these tests to describe and assert the early error.
