OCaml 5.4 syntax support #2720
OCaml 5.4 syntax support #2720
Conversation
- update vendored parsers to mirror upstream at 5.4: * introduce locations for Longident.t components * distinguish (module M:S) and ((module M):(module S)) for expressions - support for new syntaxes: * bivariance * labelled tuples
Octachron
force-pushed
the
OCaml_5.4_support
branch
from
eabfcee to
201949e
Compare
vendor/parser-extended/parser.mly
Outdated
| @@ -3185,10 +3316,12 @@ type_variance: | |||
| | MINUS BANG { [ mkvarinj "-" $loc($1); mkvarinj "!" $loc($2) ] } | |||
| | BANG MINUS { [ mkvarinj "!" $loc($1); mkvarinj "-" $loc($2) ] } | |||
| | INFIXOP2 | |||
| { if ($1 = "+!") || ($1 = "-!") then [ mkvarinj $1 $sloc ] | |||
| { if $1 = "+!" || $1 = "-!" || $1 = "+-"|| $1 = "-+" | |||
There was a problem hiding this comment.
Do you think we could accept anything starting in +, - or ! here ? The standard parser will catch errors there anyway.
There was a problem hiding this comment.
From the parser point of view, we could accept any INFIXOP2. And I agree that it probably better to accept this when building the CST.
lib/Fmt_ast.ml
Outdated
| @@ -2822,13 +2863,36 @@ and fmt_expression c ?(box = true) ?(pro = noop) ?eol ?parens | |||
| in | |||
| let outer_wrap = has_attr && parens in | |||
| let inner_wrap = has_attr || parens in | |||
| let with_label (lbl, exp) = | |||
There was a problem hiding this comment.
Instead handling special cases for ( ~foo, ~(bar : t) ), etc.. do you think we could represent the concrete syntax in the parsetree ?
I mean something like type tuple_label = Tl_pun of string loc | Tl_constraint of string loc * core_type | Tl_normal of string loc
There was a problem hiding this comment.
Are case where people don't want to normalize ~x:x, ~y to ~x, ~y ? If yes, it is probably better to separate the two forms in the parsetree. Otherwise, I am not sure if this is that useful?
There was a problem hiding this comment.
I think it's useful to make the formatting code simpler and less buggy. Decoding the AST was a major source of bugs in the past.
We generally do the normalization by modifying the AST: https://github.com/ocaml-ppx/ocamlformat/blob/main/lib/Extended_ast.ml#L68
I'd be happy to do the changes if you don't have the time ?
There was a problem hiding this comment.
After experimenting a bit more, I agree that it feels better to not have a straight mapping of locations with punning. I will send my updated commits after some more tests.
Octachron
force-pushed
the
OCaml_5.4_support
branch
from
fced3c8 to
7cc3534
Compare
This PR updates the vendored parsers and related types to mirror the 5.4 version of the OCaml parser and adds a basic support for labeled tuples and bivariance.
The parser updates means that it is possible to locate comments between
Longident.tcomponent, and it is always possible to distinguish between(module M): (module S)at least for expressions.The support for labeled expressions is bit ad-hoc and hackish in this version due to the way that labeled tuple element interact with parenthesis, for instance inside:
and it could definitively be improved.