René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"[parser] Flow types and syntax errors in function bodies","articleBody":"`lightscript@0.5.9`:\r\n```js\r\na(b?: T1): T2 -\u003e\r\n  switch\r\n```\r\n```js\r\nunknown: Unexpected token (1:4)\r\n\u003e 1 | a(b?: T1): T2 -\u003e\r\n    |     ^\r\n  2 |   switch\r\n  3 | \r\n```\r\nThe parser is misattributing the syntax error to the function arg list, rather than the body. Removing the return type annotation fixes this and gives the proper syntax error:\r\n```js\r\na(b?: T1) -\u003e\r\n  switch\r\n```\r\n```js\r\nunknown: Illegal newline. (2:8)\r\n  1 | a(b?: T1) -\u003e\r\n\u003e 2 |   switch\r\n    |         ^\r\n  3 | \r\n```\r\nI have traced the problem to this area:\r\nhttps://github.com/lightscript/babylon-lightscript/blob/master/src/parser/expression.js#L477\r\n\r\nI think the preceding code, when it tries to speculatively parse a function with a flowtype, is encountering an error in the body and eating it/resetting the state. Then because the `refShorthandDefaultPos` is set, the error appears outside the function body.\r\n\r\nNo suggested fix yet but I am looking into it.","author":{"url":"https://github.com/wcjohnson","@type":"Person","name":"wcjohnson"},"datePublished":"2017-06-20T21:47:14.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":1},"url":"https://github.com/38/lightscript/issues/38"}