Allow computed class properties #5206
Conversation
| @@ -222,6 +222,7 @@ grammar = | |||
| ObjAssignable: [ | |||
| o 'SimpleObjAssignable' | |||
| o '[ Expression ]', -> new Value new ComputedPropertyName $2 | |||
| o '@ [ Expression ]', -> new Value LOC(1)(new ThisLiteral $1), [LOC(3)(new ComputedPropertyName($3))], 'this' | |||
There was a problem hiding this comment.
Mostly copied this grammar rule from the rule for ThisProperty
| @@ -2007,7 +2007,11 @@ exports.ExecutableClassBody = class ExecutableClassBody extends Base | |||
| # The class scope is not available yet, so return the assignment to update later | |||
| assign = @externalCtor = new Assign new Value, value | |||
| else if not assign.variable.this | |||
| name = new (if base.shouldCache() then Index else Access) base | |||
| name = | |||
There was a problem hiding this comment.
This is the fix for the invalid JS for non-static computed class properties (extracted from #5205)
| @@ -118,14 +118,21 @@ exports.Rewriter = class Rewriter | |||
| # The lexer has tagged the opening bracket of an indexing operation call. | |||
| # Match it with its paired close. | |||
| closeOpenIndexes: -> | |||
| startToken = null | |||
There was a problem hiding this comment.
Allowing static computed class properties/methods was a bit more complex. In addition to the new grammar rule above, in the rewriter here we first want to make sure that the [/] in eg @[b]: ... are not tagged as INDEX_START/INDEX_END since that would be inaccurate and creates an ambiguous grammar
I believe this is safe because there's not a scenario where an actual INDEX_END would be immediately followed by a : token
| @@ -321,7 +328,12 @@ exports.Rewriter = class Rewriter | |||
| if tag is ':' | |||
| # Go back to the (implicit) start of the object. | |||
| s = switch | |||
| when @tag(i - 1) in EXPRESSION_END then start[1] | |||
| when @tag(i - 1) in EXPRESSION_END | |||
There was a problem hiding this comment.
Then, we have to make sure that for an implicit object (whether an actual implicit object or a class body) starting with a static computed property eg @[b]: ->, the start of the implicit object comes before the @
|
Thanks, this looks great. Seeing how complicated this got, I’m glad we split it off from the AST work. |
Fixes #5204
@GeoffreyBooth this PR contains two things (both raised in #5204):
I extracted this fix from #5205 to target it here against
master