Add try..finally syntax #7903
Add try..finally syntax #7903
Conversation
rescript
@rescript/darwin-arm64
@rescript/darwin-x64
@rescript/linux-arm64
@rescript/linux-x64
@rescript/runtime
@rescript/win32-x64
commit: |
|
Maybe add a |
| attach t.trailing expr.pexp_loc after_expr; | ||
| walk_list (cases |> List.map (fun case -> Case case)) t rest | ||
| (* unary expression: todo use parsetreeviewer *) | ||
| | Pexp_try (expr, cases, finally_expr) -> ( |
There was a problem hiding this comment.
why was this case not handled at all before?
I would expect a change that only deals with finally, here
| else if | ||
| Parser.lookahead p (fun state -> | ||
| match state.Parser.token with | ||
| | Lident "finally" -> true |
There was a problem hiding this comment.
what's finally in the grammar? just an identified?
Not a keyword I guess
| } catch { | ||
| | InnerError => Console.log("Inner caught") | ||
| } | ||
| } catch { |
There was a problem hiding this comment.
Is this the indentation we want? Does TS look like this when formatted?
| let finally_typed = | ||
| type_expect ~context:None env expr Predef.type_unit | ||
| in | ||
| (* Finally blocks must return unit *) |
| *) | ||
| | Texp_try of expression * case list | ||
| (** try E with P1 -> E1 | ... | PN -> EN *) | ||
| | Texp_try of expression * case list * expression option |
There was a problem hiding this comment.
nit: get assistant to use a record in a hypothetical follow-up PR, if this goes ahead
| Sexp.list (map_empty ~f:case cases); | ||
| ] | ||
| | Pexp_try (expr, cases) -> | ||
| | Pexp_try (expr, cases, finally_expr) -> |
There was a problem hiding this comment.
I keep on forgetting why we have this debugger file. Gives maintenance burden but can't remember why we need it.
| | Lstaticraise of int * t list | ||
| | Lstaticcatch of t * (int * ident list) * t | ||
| | Ltrywith of t * ident * t | ||
| | Ltrywith of t * ident * t option * t option |
There was a problem hiding this comment.
opportunity: this file is entirely uncommented
a separate PR could assist in adding some explanation
More bigger / longer term assistance opportunity: why do we have both lambda and lam? Is that a historical artefact for what used to be backwards compatibility, or can they possibly be merged now?
| P.string f L.finally; | ||
| P.space f; | ||
| brace_block cxt f b)) | ||
| P.space f; |
There was a problem hiding this comment.
curious why fin was already here
Adds support for
try..finallyandtry..catch..finallysyntax to ReScript.No breaking changes / full backwards compatibility.
Resolves #6974.