The `throw` statements that were recently inserted into these tests have
an observable impact on the parsing behavior: they causes the `"use
strict"` token sequence to be interpreted as a string literal instead of
a directive prolog, which in turn effects how the tests are interpreted.
Remove the new `throw` statements from these tests and rely on
previously-existing statements that serve the same purpose without
impacting program strictness.
The templates are being used for many tests reusing the same available function forms.
The format they are provided allow us to extend tests with cases for other tests relying
in the same formats.
After @rwaldron's feedback:
The purpose of the `!` operator is to evaluate an UnaryExpression,
coerce the result to a boolean value and then return the negated
value of that operation. But that's not what you're trying to do at
all—you just want to evaluate the expression to the right of the
operator, nothing more, nothing less. In this specific case, you
don't even really care about the evaluation, the goal is write
valid (or invalid, as the case may be) syntax that is will be
parsed according to a specific grammar rule that requires some
operator to signal that the thing is an expression and not a Block
Statement.
Tests doesn't use async functionality and don't call $DONE, so remove
"async" flag:
- src/params/error/async-gen-named-func-expr.template
- test/language/expressions/async-generator/params-named-dflt-abrupt.js
- test/language/expressions/async-generator/params-named-dflt-ref-later.js
- test/language/expressions/async-generator/params-named-dflt-ref-self.js
Intl.PluralRules.prototype is no longer a Intl.Prototype instance:
- test/intl402/PluralRules/prototype/prototype.js
Intl.PluralRules throws an error when called as a function:
- test/intl402/PluralRules/undefined-newtarget-throws.js
Module namespace objects call OrdinaryDelete for symbol properties:
- test/language/module-code/namespace/internals/delete-non-exported.js
Async generators no longer retrieves "done" property twice:
- src/async-generators/yield-star-async-next.case
- src/async-generators/yield-star-async-return.case
- src/async-generators/yield-star-async-throw.case
Minor units of CLF is 4, so we need to test with maximumFractionDigits=3
to get an error:
- test/intl402/NumberFormat/dft-currency-mnfd-range-check-mxfd.js
DateTimeFormat.prototype.formatToParts length property was changed from
0 to 1:
- test/intl402/DateTimeFormat/prototype/formatToParts/length.js
minimumSignificantDigits and maximumSignificantDigits properties are
only retrieved once:
- test/intl402/NumberFormat/11.1.1_32.js
test/annexB/built-ins/Date/prototype/setYear/time-clip.js
test/built-ins/Date/prototype/setFullYear/new-value-time-clip.js
test/built-ins/Date/prototype/setMonth/new-value-time-clip.js
- Don't try to test time-clip at the end points, because this is near
impossible to get right (needs to consider time zone offset, dst, local
mean time because of Africa/Monrovia, etc.).
test/built-ins/DataView/prototype/setFloat64/detached-buffer-after-toindex-byteoffset.js
test/built-ins/DataView/prototype/setInt16/detached-buffer-after-toindex-byteoffset.js
- Wasn't update to expect RangeError
test/built-ins/Function/internals/Construct/derived-this-uninitialized-realm.js
- Change ClassDeclaration -> ClassExpression to get completion value
test/built-ins/Function/prototype/toString/AsyncFunction.js
- Add missing \n in expected string
- Also fixed in gh-847
test/built-ins/global/global-object.js
- Add 'var' to make test pass in strict-mode
test/language/block-scope/syntax/redeclaration-in-block/attempt-to-redeclare-function-declaration-with-function-declaration.js
- This is allowed in sloppy mode when Annex B is implemented
test/language/expressions/async-generators/expression-yield-as-statement.js
- Fix calls to then()
test/language/module-code/namespace/internals/own-property-keys-binding-types.js
test/language/module-code/namespace/internals/own-property-keys-sort.js
- Tests weren't updated after removal of @@iterator from module
namespace objects
test/language/module-code/namespace/internals/set-prototype-of-null.js
- Fix syntax error
test/language/statements/async-function/early-errors-no-async-generator.js
- No longer valid now that async iteration proposal is at stage 3
The HTML spec requires browsers define a non-configurable property
of the global (window) object named "top". This makes it
impossible for a browser to successfully run a test in strict mode
if said test attempts to write to a global variable named "top".
* Re-organize coverage for YieldExpression
* Extend test coverge for YieldExpression
* fixup! Extend test coverge for YieldExpression
Remove unused variables
A recent change to the specification [1] introduces parse-time errors
for certain usages of `super` within eval code. Modify all tests that
are affected by this change:
- Update the test bodies to accurately enforce the new semantics
- Rename files to better reflect the section of the specification that
they enforce
- Update test meta-data
- Change the `esid` meta-data to reflect the location of the relevant
specification text
- Remove the `es6id` meta-data as the behavior is no longer relatable
to that specification
- Introduce the `features` meta-data in cases where the test file's
new location no longer reflects all required language features
[1] "Normative: Clarify rules around super inside eval"
https://github.com/tc39/ecma262/pull/685
* Add tests for prototype realm inference
* Add tests for miscellaneous realm concerns
* Add tests for realm of spec-created Errors
In some cases, Error objects produced by the specification are
observable from ECMAScript code. Among these cases, some are further
differentiated in that they occur outside of any built-in function and
may be triggered through syntactic production directly. The current
realm record is commonly interpreted incorrectly under these
circumstances.
Add tests asserting that the expected realm record is used when
constructing such Error objects.
* Add tests for realm use in ArraySpeciesCreate
* Add tests for function realm retrieval
* Add tests for cross-realm behaviors of Symbols
* Add tests for GetValue and PutValue
* Add tests for realm of spec-created Arrays
In some cases, Arrays produced by CreateArrayFromList are observable
from ECMAScript code. Among these cases, two occur outside of any
built-in function and may be triggered through syntactic production
directly. The current realm record is commonly interpreted incorrectly
under these circumstances.
Add tests asserting that the expected realm record is used when
constructing arrays.
* Add test for spec-created object
* fixup! Add tests for realm of spec-created Errors
* fixup! Add tests for realm of spec-created Errors
* fixup! Add tests for prototype realm inference
* fixup! Add tests for miscellaneous realm concerns
Because expectations regarding error "phase" are now expressed via test
meta-data, the test runner may now enforce this requirement on negative
tests.
Remove the "NotEarlyError" from the project source. This reduces the
amount of domain knowledge required to author tests and lessens the
potential for inconsistencies between tests.
Authored via the following command:
$ find test -type f -print0 | \
xargs -0 sed \
-i 's/^\(\s*\)negative:\s*SyntaxError\s*$/\1negative:\n\1 phase: early\n\1 type: SyntaxError/g'
The expected errors in these tests cannot be asserted with the
`assert.throws` helper function for various reasons. Re-format their
meta-data according to the latest design in order to more precisely
describe test expectations.
The file previously named `values-binding-types_.js` is not intended to
be interpreted as a test. Therefor (in accordance with the project's
`INTERPETING.md` file), its name should include `_FIXTURE` as a suffix.
This test is being added because the committee is considering changing
this evaluation order (as discussed at the May 2016 Munich meeting). The
consequences of this spec change will be more clear as a test change
than a simple test addition.
* Assert creation of 'arguments' object
Ensure that the 'arguments' object is created in cases where it is not
required by the body but is required by the parameters.
* Add tests for cases that disable "arguments" map
* Add tests for NewTarget override of bound function
* Add test for properties of exotic String objects
ECMA262 allows for an arbitrary number of "ScriptJob"s to run in a given realm.
Although there is no standard mechanism for enqueuing these jobs, many
implementations offer this functionality through custom APIs. In those hosts,
the semantics describing script interactions are directly observable.
In order to guarantee conformance to the specification in advance of a
standardized API, Test262 now requires that hosts provide a $.evalScript
function whose behavior is defined in the project's "INTERPRETING.md" file.
Use this host-provided API to ensure that implementations correctly observe the
specification text that dictates script interactions.
(In writing these tests, I noticed some gaps in coverage that are observable
from a single script execution. This patch includes a dedicated commit for
these tests that do not require $.evalScript.)
* Improve coverage of GlobalDeclarationInstantiation
* Add tests for script interactions
Use the host-provied `$.evalScript` method to assert conformance to the
specification text that defines script interactions.
* fixup! Improve coverage of GlobalDeclarationInstantiation
* Add tests for early errors in functions
* Improve tests for class accessors
Use the `propertyHelper.js` utility in order to functionally test the
property descriptors of class methods.
* Remove redundant tests
The semantics of an IdentifierReference as a PropertyDefinition within
an object initializer are exhaustively tested by the files in this
directory whose name match the pattern `prop-def-id-*.js`.
Delete the redundant tests in favor of the more descriptively-named and
more exhaustive alternatives.
* Rename tests
* Update test names to be more descriptive
* Add tests for property descriptors of accessors
* Add tests for runtime error during method dfn
* Add test for observable iteration
* Extend tests for class "name" inference
Ensure that when a class defines a static "name" method, that method
definition prevents the "name" inference behavior.
* Re-generate tests
This is a pure miscellaneous change to use global values of NaN and Infinity instead of their namespaced equivalents.
Ref 9ae0567a50 (r69049894)
cc @suwc
This re-factors some existing Sputnik tests to be more targeted and to use a
pattern that can be generalized to other forms. We could test these all day,
but I've limited myself to forms introduced in ES2015, specifically
YieldExpression and new.target. Note that SpiderMonkey incorrectly throws a
SyntaxError for these.
I thoughtlessly wrote ReferenceError tests for yield = 1 until I realized
that such productions are not actually recognized by the grammar, so the early
errors do not apply. Instead, I've added a negative syntax test for that case.
* Refactor test for valid cover
* Add tests for ValidSimpleAssignmentTarget
Ensure that constructs introduced in ES2015 are disallowed as assignment
targets, with or without a "cover" grammar.
* Add test for grammar precedence of YieldExpression
Fixes gh-696
As pointed in gh-696, the ArrayBuffer ctor won't throw a RangeError when
invoked without arguments.
Instead of fixing the invalid assertion, this commit removes it as the
coverage for subclassing is already satisfied by the use of slice.
* Add missing test for early error
* Add missing test for WithBaseObject
* Improve coverage for `new.target`
* Add test for deletion of SuperReference
* Add tests for `in` keyword restrictions
* fixup! Improve coverage for `new.target`
The global isNaN is not precise at all, and Number.isNaN is an ES6 feature that makes it preferrable to use assert's sameValue for NaN values, as it handles it internally using the comparison.
This test was intended to assert the semantics of the `for-in`
statement, but it was mistakenly authored to use a `for-of` statement.
Update the statement under test, and improve the testing methodology to
correctly assert the creation of distinct bindings.
Ensure that early errors restricting labelled function declarations
within WithStatement and IfStatement are honored. Rename existing tests
to match the specification's spelling.
* Improve test coverage for `super` keyword
Add tests for SuperCall and SuperProperty, organized together in the
`test/language/expressions/super/` directory. For SuperProperty, include
tests for usage from within Object initializers and class bodies because
a different set of semantics are observable from each context.
The prior version of this test asserted only the property's
configurability. It was also limited to the RegExp object as returned
from the RegExp constructor.
Extend the test to verify all values of the property descriptor.
Duplicate these assertions in a separate file dedicated to the RegExp
object as created from a RegExp literal.
The specification contains an explicit informative NOTE explaining that
the 'm' flag does not effect the behavior of the '^' assertion. Add a
test to verify this.
ES2015 reads:
> RegExpUnicodeEscapeSequence :: u{ HexDigits }
>
> - It is a Syntax Error if the MV of HexDigits > 1114111.
Use the MV 0x110000 to assert the lower boundary of values which are
expected to produce the SyntaxError.
A subtle aspect of the for-of iteration protocol concerns abrupt
completions that do *not* trigger iterator closing. Although this detail
is implicit in the current structure of the specification text, some
hosts may violate the protocol by closing the iterator because later
steps *do* specify that behavior.
The V8 engine is one such host--as of this writing, it incorrectly
closes the iterator when accessing the `value` property of the iterator
result produces an abrupt completion.
Add tests verifying that the iterator protocol is not violated in this
way for abrupt completions during the semantics of for-of evaluation.
These tests have been re-factored to expand coverage of the "default
parameter" language feature and to more closely adhere to this project's
preferred file organization.
Ensure that when HasBinding of an Object environment record returns an
abrupt completion, that same completion is returned to the runtime.
Update the meta-data of related tests for consistency with this new
test.
The latest revision of ECMA262 makes special provisions for classes
which extend the `null` value [1]. Update the relevant tests
accordingly.
[1] https://github.com/tc39/ecma262/issues/543
Utilize the test generation tool to increase coverage of destructuring
assignment semantics. Previously, only destructuring assignment in the
AssignmentExpression position was tested. With this change applied, the
same tests will assert expected behavior for destructuring assignment in
`for..of` statements, as well.
A limited number of tests are applied to the `for..in` statement as
well, but due to the iteration protocol observed by that statement, many
destructuring tests are not relevant, and others cannot be automatically
generated from this format.
This change set does not include a test for restrictions relating to
template literals because such a test already exists in the project.
While a form of this test for string literals in strict mode code
existed previously, it is less precise and relies on unrelated
semantics. Remove the previous form and replace with a more direct
version.
Add tests that assert the management of the running execution context's
LexicalEnvironment and VariableEnvironment components, as created by the
following abstract operations:
- NewDeclarativeEnvironment
- NewObjectEnvironment
- NewFunctionEnvironment
Many tests require the use of non-strict direct eval, meaning they may
not be run in strict mode. This does not effect coverage because the
semantics in these cases are not observable from strict mode code.
Some situations require the creation of a binding, but this binding has
no relevance to the test itself. In these cases, use names consisting
solely of the underscore character (`_`).
Avoid the use of Block statements wherever possible, as these trigger
the creation of additional environments which may interfere with the
behavior under test.
Closes#572
Introduce tests for new semantics for ES2015 features such as
lexically-scoped bindings. Also add tests for semantics defined in prior
editions of the specification but not yet covered in this test suite.
Limit tests in `language/expressions/call/` and `built-ins/eval/` to
only assert details that are directly related to the invocation pattern.
Re-organize all other tests within the `language/eval-code/` directory,
further categorizing each as `direct` or `indirect` as appropriate.
Remove the extraneous "executable" flag from those files which carried
it.
This harness function is not necessary in the majority of cases in which
it is used. Remove its usage to simplify tests and decrease the amount
of domain-specific knowledge necessary to contribute to the test suite.
Persist the harness function itself for use by future tests for ES2015
modules (such a helper is necessary for tests that are interpreted as
module code).
Except for a small set of expected input/output pairs, both ES5 and
ES2015 define the expected return value of these methods in terms of an
"implementation-dependent approximation." This makes it inappropriate to
enforce expectations for specific values, even if expressed imprecisely.
The files in this patch are organized according to the following naming
scheme:
Prefix | Grammar production
-------------------|-------------------
`array-empty-` | ArrayAssignmentPattern : [ ]
`array-elision-` | ArrayAssignmentPattern : [ Elision ]
`array-rest-` | ArrayAssignmentPattern : [ Elisionopt AssignmentRestElement ]
`array-elem-` | ArrayAssignmentPattern : [ AssignmentElementList ]
`array-elem-trlg-` | ArrayAssignmentPattern : [ AssignmentElementList , Elisionopt AssignmentRestElementopt ]
Suffix | Intent
-------------------|-------
`-abpt-close-err` | The assignment evaluation returns an abrupt completion, and the iterator's `return` method throws an error
`-abpt-close-skip` | The assignment evaluation returns an abrupt completion, but IteratorClose is not invoked
`-abpt-close` | The assignment evaluation returns an abrupt completion, and IteratorClose is invoked as specified
`-get-err` | Abrupt completion returned from GetIterator
`-nrml-close-err` | The assignment evaluation completes, and the iterator's `return` method throws an error
`-nrml-close-null` | The assignment evaluation completes, and the iterator's `return` method returns a non-Object value (there is no corresponding `-abpt-` suffix because the algorithm does not reference the return value in those cases)
`-nrml-close-skip` | The assignment evaluation completes, but IteratorClose is not invoked
`-nrml-close` | The assignment evaluation completes, and IteratorClose is invoked as specified
Not all suffixes are appropriate for all productions. Suffixes have been
simplified in cases where less specificity is necessary to disambiguate
test cases.
Some tests for `eval` assert the equality of key enumeration on the
global object, comparing the ordering when the object is referenced
through eval.
Based on the test bodies and the "info" field in their metadata, these
tests appear to have been written under the mistaken impression that
erroneous creation of an environment record would be observable through
the `this` value.
In reality, the value in such cases resolves to the global object. That
renders these tests redundant and overly complex--none of the
distinctions between each test actually demonstrates a different
behavior.
Remove the redundant tests and introduce three new tests asserting the
correct resolution of the `this` keyword for direct eval code from
within the global scope and from within function scope.
Files whose name ends in `_.js` are not themselves valid Test262 tests
and should not be interpreted as such by test runners.
---
The tests for the GetBindingValue method of the module Environment
Record are very minimal. This is because GetBindingValue is necessary to
assert any aspect of binding creation/initialization/mutation. In this
way, GetBindingValue is being implicitly tested by every test that
references a binding value.
Files whose name ends in `_.js` are not themselves valid Test262 tests
and should not be interpreted as such by test runners.
---
Because the tests in this patch concern declaration *instantiation*,
care has been taken to avoid asserting binding values following
evaluation. Because a given module's dependencies are evaluated prior to
the module itself, this is only observable in modules which import their
own bindings.
A separate patch dedicated to the evaluation of module code asserts the
behavior of bindings following evaluation.
---
For tests that concern the creation of a module namespace object, this
patch relies on the semantics of the `in` operator. The `in` operator
uses the [[HasProperty]] internal method and avoids testing unrelated
semantics concerning binding resolution. Those semantics should be
explicitly asserted with a separate set of tests dedicated to that
purpose.
---
One test case which is notably missing is error resulting from a cycle
due to an `import` declaration (under the current file naming scheme,
such a test might be named `instn-named-err-circular.js`). Due to the
recursive nature of ModuleDeclarationInstantiation, it is not
technically possible for a circular request to be found in step 12.c.i.
Cycles rely on at least 2 `export` declarations, and because these are
resolved *before* imports, any cycle would trigger failure prior to step
12.c.
---
One aspect of *module* resolution that makes ordering observable is the
fact that resolution can fail in two distinct ways (i.e. with a
SyntaxError or with a ReferenceError). This patch includes tests that
leverage this detail in order to assert that modules are resolved in the
correct sequence.
However, from the perspective of the ECMA-262 specification, the
ordering of *export* (e.g. binding) resolution is not observable. This
is due to restrictions on the grammar, where each export is independent.
When *export* resolution fails, it does so with instances of SyntaxError
alone, precluding the above strategy.
So while ModuleDeclarationInstantiation resolves the exports of the
module's dependencies in the following order:
1. "Indirect" exports, e.g.
- `export { x } from './y.js';`
- `export { x as z } from './y.js';`
- `import { x } from './y.js'; export { x };`
2. "Star" imports
- `import * as ns from './y.js';`
3. "Named" (my word) imports
- `import x from './y.js';`
- `import { x } from './y.js';`
- `import { x as z } from './y.js';`
Intentional failures cannot be used to discern resolution ordering.
The runtime semantics of this statement are host-defined and therefore
untestable, but the statement's affect on the formal grammar should be
consistent across all implementations.
V8 ran into an issue where the YAML parser our test setup is using
didn't understand the newline, and failed to parser the negative
test expectation below, causing the test to fail. This patch fixes
the issue.
Leo Balter
André Bargull
Isiah Meadows
Marshall Roch
Daniel Ehrenberg
Robin Templeton
Rick Waldron
Lucas Azzola
Georg Neis
Mike Pennisi
Kevin Gibbons
Caio Lima
Joseph Pecoraro
Jeff Walden
Caitlin Potter
Shu-yu Guo
⭐caitp⭐
Kevin Gibbons
arai-a
Adam Klein
Michael Ficarra
Josh Wolfe
dop251
Tim Disney
Tom Care
Yusuke Suzuki
Brian Terlson
Jeff Morrison
Mathias Bynens
Ian Halliday
Gorkem Yakin