When traversing a tree parsing for TypeScript syntax and hitting a `TSFunctionType` node, the `typeParameters` and `typeAnnotation` fields are correctly visited but the `parameters` field isn't. As a result visitors by default don't visit `foo` below:
```
var x: (foo) => void; // foo is never visited
```
```
module.exports = function() {
return {
visitor: {
Identifier(path) {
if (path.node.name == "foo") {
// Never hit because it's nested within TSFunctionType.parameters
path.node.name = "bar";
}
}
}
};
}
```
It appears to be a bug in babel-types/src/definitions/typescript.js which omits `parameters` in the visitors list for `fnOrCtr`. Fixed by adding it.
* Add parser createParenthesizedExpressions option …
When set to `true` we create `ParenthesizedExpression` nodes instead of
setting `extra.parenthesized`.
* Also update babel-parser.d.ts
* add partial application syntax and some tests
* remove unnecessary error message and hasPartial function from parseNewArguments
* add types for PartialExpression
* Update the tests
* rename PartialExpression to Partial
* move Partial from expressions to types and rename to ArgumentPlaceholder
* add tests for ArgumentPlaceholder in babel-generator
* rename Partial to ArgumentPlaceholder
* update the tests
* remove alias from the type and undo changes in generated folder
* adds a nice error message
* better definition for the type
* auto-generated files
* update the conditional for allowPlaceholder message and tests
* update CallExpression definition to accept ArgumentPlaceholder
* change description
* clean up
* indent ArgumentPlaceholder entry and revert unwanted changes
* fix: strip type imports used in Enums and object types
* chore: update failing snapshot
* docs: correct TSPropertySignature comment
* fix: enum value should be considered a reference
* chore: add tests for TSPropertySignature and TSEnumMember
In https://github.com/babel/babel/issues/9511 (and #9495 is another symptom), @PavelKastornyy reported a node crash becaue the JavaScript heap run out of memory. The problem was that their code was adding enumerable properties to `Object.prototype`: it is something that shouldn't be done, but Babel shouldn't make node crash if someone adds them.
I reduced down the problem to `for...in` loops in `@babel/traverse` that grew the memory consumption exponentially because of that unexpected properties.
* Add parser createParenthesizedExpressions option …
When set to `true` we create `ParenthesizedExpression` nodes instead of
setting `extra.parenthesized`.
* Also update babel-parser.d.ts
* babel-types is* type checks accept null | undefiend as value TS type
* Use | null | undefined instead of any, make Flow type nullable
* Also change flow types of generated src
* Remove types on program exit in typescript transform
* Revert changes to typescript transform
* Correctly add visitors for param decorators
* Fix plugin for node 6
| Q | A
| ------------------------ | ---
| Fixed Issues? | #7747 (partly)
| Patch: Bug Fix? |
| Major: Breaking Change? |
| Minor: New Feature? | Yes
| Tests Added + Pass? | Yes
| Documentation PR |
| Any Dependency Changes? |
| License | MIT
@JamesHenry This changes the AST format. CC @DanielRosenwasser for review.
Supports parsing type arguments on tagged template calls.
Should wait on Microsoft/TypeScript#23430 to be merged so we're sure we have the final syntax.
Fixes#8163
The `isNodesEquivalent()` algorithm incorrectly assumes that any object properties must themselves be AST nodes, which is not the case here, causing the error.
| Q | A <!--(Can use an emoji 👍) -->
| ------------------------ | ---
| Fixed Issues? | Fixes#8272 <!-- remove the (`) quotes to link the issues -->
| Patch: Bug Fix? | Yes
| Major: Breaking Change? |
| Minor: New Feature? |
| Tests Added + Pass? | Yes
| Documentation PR | <!-- If so, add `[skip ci]` to your commit message to skip CI -->
| Any Dependency Changes? |
| License | MIT
The visitor for `typeArguments` in function calls wasn't set in the definition, so type annotations added there weren't visited.
This is required to fix https://github.com/babel/babel-eslint/issues/644
<!--
Before making a PR please make sure to read our contributing guidelines
https://github.com/babel/babel/blob/master/CONTRIBUTING.md
For issue references: Add a comma-separated list of a [closing word](https://help.github.com/articles/closing-issues-via-commit-messages/) followed by the ticket number fixed by the PR. It should be underlined in the preview if done correctly.
-->
| Q | A <!--(Can use an emoji 👍) -->
| ------------------------ | ---
| Fixed Issues? | <!-- remove the (`) quotes to link the issues -->
| Patch: Bug Fix? |
| Major: Breaking Change? | No
| Minor: New Feature? | Yes
| Tests Added + Pass? | Yes
| Documentation PR | <!-- If so, add `[skip ci]` to your commit message to skip CI -->
| Any Dependency Changes? |
| License | MIT
See eb815be907 for more information about this feature.
The proto modifier indicates that a property declared using `x: T` syntax is actually present on the prototype object of the class, rather than an own property.
The proto and static modifiers are mutually exclusive, as class declarations don't simultaneously define the static prototype object, as they do the instance prototype.
This syntax is only supported on declared classes, not object types, interfaces, or runtime class declarations, and as such should only appear in library definitions.
| Q | A <!--(Can use an emoji 👍) -->
| ------------------------ | ---
| Fixed Issues? | Fixes#8001
| Patch: Bug Fix? | Y
| Major: Breaking Change? | Not unless things were relying on a very specific AST structure as output
| Minor: New Feature? |
| Tests Added + Pass? | Yes
| Documentation PR | <!-- If so, add `[skip ci]` to your commit message to skip CI -->
| Any Dependency Changes? |
| License | MIT
The filed bug was for `-0`, but I also realize that `numericLiteral` also shouldn't be getting `-Infinity`/`Infinity` or `NaN` since those just get stringified with the JS-standard stringification logic which means we get a reference to `NaN` or `Infinity` identifiers, which could have been shadowed in a local scope and thus not be the right value. I've avoided that by generating `NaN` as `0/0` and the infinity values as `1/0` and `-1/0`.
Follow-up to https://github.com/babel/babel/pull/7404
There is no TypeParameter type in the Babylon TypeScript AST hierarchy.
Flow does have a TypeParameter type, but it should not be confused with
the TypeScript TSTypeParameter, since Flow !== TypeScript, and the types
have totally different fields. Instead, the .typeParameter.type of a
Babylon-parsed TSInferType node should be TSTypeParameter.
It would probably be fine to leave the declared type of the TSInferType
.typeParameter field as TSType instead of TSTypeParameter, but the more
specific TSTypeParameter type should be safe/correct, since the TypeScript
parseInferType function always uses SyntaxKind.TypeParameter:
66d6e5e6e0/src/compiler/parser.ts (L3006)
I noticed this typo because it has been causing ast-types test failures
lately, e.g. https://travis-ci.org/benjamn/ast-types/jobs/369634972
* Add all option to babel-plugin-syntax-flow
The Flow parser has some conditional logic based on whether types should
be parsed or not, which is based on either (a) the @flow pragma in the
docblock of the file of (b) the `all` configuration, provided via
command line or .flowconfig.
This commit adds the ability to provide the `all` configuration to
Babel as well, via the syntax-flow plugin. This should be set to `true`
if the project uses all=true.
* Parse @flow pragma
The Flow parser has some conditional logic based on whether types should
be parsed or not, which is based on either (a) the @flow pragma in the
docblock of the file of (b) the `all` configuration, provided via
command line or .flowconfig.
This commit parses the @flow (or @noflow) pragma from the first comment
in the source file. Directives are allowed to appear before the comment.
* WIP: add tests for explicit type arguments
This commit includes tests which have unexpected output, but will change
to the expected output as later commits add parsing support for various
features.
* Parse type arguments in new expressions
* Parse type arguments in call expressions
* Parse optional call expressions with explicit type args
* Add explicit type arguments to babel-types
Flow calls these typeArguments instead of typeParameters, which clearly
separates formal/actual parameters, and mirrors the existing arguments
key.
The existing definitions to support TypeScript also included Flow's
TypeParameterInstantiation node type, which I've moved to the the new
field.
* Add support for explicit type arguments to babel-generator
* Add test for explicit type args to transform-flow-strip-types plugin
* Oops. Forgot to regenerate the babel-types README.
* Fix Flow parser shouldParseTypes() function
I was looking at `options.all`, but the correct property ws
`options.flowAll`. Oops!
* Remove typeapp_call from whitelist of expected failures
Now that Babylon parses this syntax extension, we can remove the
typeapp_call tests from the list of expected differences.
Note that I am using the `flowAll` option, mirroring the behavior of the
Flow tests, which assume types without requiring the `@flow` pragma.
* Use Babylon plugin options instead of parser options
* Parse optional call expressions type arguments unambiguously
* Support internalSlots to babel-types and babel-generator
* Parsing support for internal slot properties
* Print internal slots in babel-generator
* Add whitespace before first internal slot property
