* Add error recovery support to @babel/parser
* Update @babel/parser tests to always recover from errors
* Update this.raise usage in @babel/parser:
- expression.js
- lval.js
- statement.js
- estree.js
- flow.js
- jsx/index.js
- tokenizer/index.js
* Update @babel/parser fixtures with recovered errors
* Fix tests out of @babel/parser
* Do not use try/catch for control flow
* Update invalid fixtures
* Do not report invalid lhs in toAssignable
* Do not validate function id multiple times
* Dedupe reserved await errors
* Remove duplicate errors about strict reserved bindings
* Remove duplicated error about yield/await inside params
* Don't error twice for methods in object patterns
* Don't report invalid super() twice
* Remove dup error about reserved param for expr arrows
* Remove double escapes in migrated tests
* Dedupe errors about invalid escapes in identifiers
* Remove duplicated error about decorated constructor
* Remove duplicated error about spread in flow class
* Don't throw for invalid super usage
* Don't fail for object decorators with stage 2
* Fix flow inexact type errors
* Fix flow
* Fix errors about escapes in keywords (ref: #10455)
* Update after rebase
* Fix todo
* Remove duplicated error when using += for defaults
* Remove unnecessary throw
* Nit: use ??
Instead of using `convert-source-map`'s `removeComments` method before
parsing the file, we can first parse the file with `@babel/parser` and then
analyze the comments.
This is needed because it is not possible to reliabily detect comments in
JavaScript without fully parsing the file:
https://github.com/thlorenz/convert-source-map/issues/63
These are now enabled by default:
- objectRestSpread (2018)
- asyncGenerators (2018)
- optionalCatchBInding (2019)
- jsonStrings (2019)
TODO (after this commit):
- [ ] Deprecate the `@babel/plugin-syntax-*` packages.
- [ ] Deprecate the `@babel/plugin-proposal-*` packages.
- [ ] Create the `@babel/plugin-transform-*` packages.
* Add sourcemap examples for identifiers in code generator.
* Generate sourcemaps with exact mappings for identifiers, even if leading/trailing tokens also use that location.
* Revert "Move subclass inheritance to end (#7772)"
This reverts commit f8ab9466d331871a90f458af40b14e8d831e0c29.
* Only use getPrototypeOf if setPrototypeOf is implemented
* Update fixtures
* Helpers updates
* Update fixtures
* Fall back to getPrototypeOf
* Update fixtures
We were using `Object.create` to setup the prototype chain at the start of the class definition, which lead to #7771.
I was a bit worried about a speed hit, but it seems everyone optimizes the two patterns the same way.
https://jsbench.github.io/#f9fca52407643d96458a35763b201215Fixes#7771.
* Update decorators parsing
This commit introduces three changes:
1) Class properties can be decorated
2) Decorators can contain arbitrary expressions, using @(...)
3) The Decorator node type has a new property, "arguments". This
makes it possible do distinguish @dec() and @(dec()), which have
different behaviors because @(dec()) is equivalent to @(dec())().
* Rename Decorator#expression to Decorator#callee
* Add test for @dec()()
* #7349
- Add "use strict" directive to transformed class bodies
* Update tests (#7349)
* Add test for program with use strict directive (#7349)
* Update tests (#7349)
* Update source mapping (#7349)
* Add test for parent blockStatement with use strict directive (#7349)
* Update tests (#7349)
* Update tests (#7349)
* Update tests (#7349)
* Add test for constructor-only class (#7349)
* Constructor only classes are strict, too
But constructor only that use non-simple parameters must use a strict function wrapper.
* Fix test
* Wrapper not needed if class is already strict
* Revert change to lerna.json
