Store parsed string delimiters in Lambda constants

Give Lambda Const_string the same structured payload used by Lam, and move processed-delimiter decoding into External_arg_spec so Lambda producers can store the final delimiter directly.

Remove the inline-constant delimiter round-trip, cover every processed delimiter encoding, and consolidate the branch changelog into review-facing entries.

Signed-off-by: Cristiano Calcagno <cristianoc@users.noreply.github.com>

Author: cristianoc

CHANGELOG.md

@@ -70,19 +70,8 @@
 - Represent optional parameters with defaults structurally, removing downstream name-based detection and producing more consistent JavaScript parameter names. https://github.com/rescript-lang/rescript/pull/8580
 - Sync the platform npm package's compiler binaries (`packages/@rescript/<platform>/bin`) via dune promotion on every `dune build`, instead of Makefile/CI copy steps that only ran when make did: a plain `dune build` can no longer leave `cli/*.js` and the test harnesses running a stale compiler. https://github.com/rescript-lang/rescript/pull/8560
 - Remove unused compiler IR definitions, modules, helpers, error variants, and Typedtree fields. https://github.com/rescript-lang/rescript/pull/8551 https://github.com/rescript-lang/rescript/pull/8555
-- Remove the unused OCaml pipe primitives `%revapply`/`%apply` and the `Ploc` Lambda constructor. `__LOC__` and friends still compile to location constants in `translcore`.
-- Lower exception packing in `translcore` instead of convert, drop unused `raise_kind` / reraise tracking, and emit `RE_EXN_ID` string equality from matching instead of `Pextension_slot_eq`.
-- Remove the `Pwrap_exn` / `%wrap_exn` primitive. Exception packing, `Promise.catch`, and `JsExn.anyToExnInternal` call `Primitive_exceptions.internalToException` as a normal module function.
-- Split `let rec` groups into actual recursive clusters when Lambda is produced (`Lambda_scc.bind_rec`). Convert `Lletrec` is identity; `Lam_scc` is removed. JS compile only sorts functions before values for dummy/`updateDummy` init.
-- Remove the dense int-switch-to-add peephole (`happens_to_be_diff`) from Lambda-to-Lam conversion. Convert `Lswitch` is identity.
-- Remove the switcher-offset peephole from Lambda-to-Lam conversion. Convert `Llet` is identity.
-- Convert `Lstaticcatch` / `Lstaticraise` are identity. Drop exit aliasing (`exit_map`); `Lam_pass_exits` already inlines a catch whose handler is a tiny `Lstaticraise`.
-- Drop the unused `exports` argument from Lambda-to-Lam conversion.
-- Represent `%identity` / `%ignore` / unary `+` as `Peliminated` on both Lambda and Lam. `Lambda.mk_prim` expands them so they never appear as `Lprim` nodes; Lam matches `assert false`.
-- Use a single `int32` integer constant on Lambda and Lam (`Const_int of int32`). Assert-false is a dedicated constant (`Const_assertfalse`), not a tagged `0`.
-- Drop dead typedtree constants `Const_int32` and `Const_int64`. Integer literals are `Const_int` (native `int`) or `Const_bigint`.
-- Put char, string, float, and bigint constants on Lambda and drop `Const_base`.
-- Drop Lambda `Const_immstring`; location primitives use `Const_string`, and matching can share equivalent string-valued actions.
+- Make Lambda-to-Lam conversion structural for lets, switches, static exits, recursive binding groups, exception packing, and eliminated identity operations. Semantic rewrites now happen during Lambda production or in named Lam passes; obsolete conversion state and `Lam_scc` are removed.
+- Remove obsolete Lambda and Lam primitives and align their scalar constant representations. Lambda and Lam now use `int32` integers and matching char, string, float, and bigint cases; Lambda strings carry their parsed output delimiter, assert-false is distinct from integer zero, and dead typedtree integer variants are removed.
 - Make locally abstract value constraints (`let f: type a. t = value`) structural in the parsetree, remove the obsolete `Pexp_newtype` and `Texp_newtype` wrapper metadata, and keep the old encoding confined to the frozen external-PPX bridge. The CMT magic number is bumped to `Caml1999T024`. https://github.com/rescript-lang/rescript/pull/8575
 - Eliminate the `Pjs_fn_make`/`Pjs_fn_make_unit` arity-adjustment primitives and the `unsafe_adjust_to_arity` machinery: with structural arity, functions are constructed at their final arity, so the enforcement layer (and the active-pattern currying split it compensated for) is deleted. Generated code improves: no adapter closures for patterns on mutable fields, better constant propagation and name preservation, and recursive modules whose members are plain functions compile statically without the runtime bootstrap. https://github.com/rescript-lang/rescript/pull/8570
 - Cleanups enabled by structural arity: remove the unreachable `Too_many_arguments` error and the `?in_function` threading through the type checker that existed only to decorate it; remove the dead `function$`-vs-arrow unification bridge, `Ctype.arity`, and the unused parsetree arity helpers; deduplicate the analysis arrow-flattening helpers. https://github.com/rescript-lang/rescript/pull/8569

compiler/core/lam_constant_convert.ml

@@ -26,9 +26,7 @@ let rec convert_constant (const : Lambda.structured_constant) : Lam_constant.t =
   match const with
   | Const_int i -> Const_int i
   | Const_char i -> Const_char i
-  | Const_string (s, opt) ->
-    let delim = Ast_utf8_string_interp.parse_processed_delim opt in
-    Const_string {s; delim}
+  | Const_string {s; delim} -> Const_string {s; delim}
   | Const_float i -> Const_float i
   | Const_bigint (sign, i) -> Const_bigint (sign, i)
   | Const_pointer (Pt_constructor {name = "()"}) ->

compiler/frontend/ast_utf8_string_interp.ml

@@ -271,13 +271,6 @@ let transform_test s =
   List.rev cxt.segments
 
 module Delim = struct
-  let parse_processed = function
-    | None -> Some External_arg_spec.DNone
-    | Some "json" -> Some DNoQuotes
-    | Some "*j" -> Some DStarJ
-    | Some "bq" -> Some DBackQuotes
-    | _ -> None
-
   type interpolation =
     | BackQuotes (* string interpolation *)
     | Js (* simple double quoted string *)
@@ -333,4 +326,4 @@ let transform_pat (p : Parsetree.pattern) s delim : Parsetree.pattern =
     }
   | Unrecognized -> p
 
-let parse_processed_delim = Delim.parse_processed
+let parse_processed_delim = External_arg_spec.parse_processed_delim

compiler/ml/external_arg_spec.ml

@@ -26,6 +26,13 @@
 
 type delim = DNone | DStarJ | DNoQuotes | DBackQuotes
 
+let parse_processed_delim = function
+  | None -> Some DNone
+  | Some "json" -> Some DNoQuotes
+  | Some "*j" -> Some DStarJ
+  | Some "bq" -> Some DBackQuotes
+  | _ -> None
+
 type cst = Arg_int_lit of int | Arg_string_lit of string * delim
 
 type label_noname = Arg_label | Arg_empty | Arg_optional

compiler/ml/external_arg_spec.mli

@@ -24,6 +24,8 @@
 
 type delim = DNone | DStarJ | DNoQuotes | DBackQuotes
 
+val parse_processed_delim : string option -> delim option
+
 type cst = private Arg_int_lit of int | Arg_string_lit of string * delim
 
 type attr =

compiler/ml/lambda.ml

@@ -336,7 +336,7 @@ type pointer_info =
 type structured_constant =
   | Const_int of int32
   | Const_char of int
-  | Const_string of string * string option
+  | Const_string of {s: string; delim: External_arg_spec.delim option}
   | Const_float of string
   | Const_bigint of bool * string
   | Const_pointer of pointer_info
@@ -423,11 +423,14 @@ and lambda_switch = lambda switch
 *)
 let const_int (i : int) = Const_int (Int32.of_int i)
 
+let const_string s delim =
+  Const_string {s; delim = External_arg_spec.parse_processed_delim delim}
+
 let const_of_typed (c : Asttypes.constant) : structured_constant =
   match c with
   | Asttypes.Const_int i -> Const_int (Int32.of_int i)
   | Asttypes.Const_char i -> Const_char i
-  | Asttypes.Const_string (s, d) -> Const_string (s, d)
+  | Asttypes.Const_string (s, d) -> const_string s d
   | Asttypes.Const_float f -> Const_float f
   | Asttypes.Const_bigint (sign, i) -> Const_bigint (sign, i)
 

compiler/ml/lambda.mli

@@ -292,7 +292,7 @@ and value_kind = Pgenval
 type structured_constant =
   | Const_int of int32
   | Const_char of int
-  | Const_string of string * string option
+  | Const_string of {s: string; delim: External_arg_spec.delim option}
   | Const_float of string
   | Const_bigint of bool * string
   | Const_pointer of pointer_info
@@ -403,6 +403,7 @@ and lambda_switch = lambda switch
 val make_key : lambda -> lambda option
 
 val const_int : int -> structured_constant
+val const_string : string -> string option -> structured_constant
 val const_of_typed : constant -> structured_constant
 val const_unit : structured_constant
 val lambda_assert_false : lambda

compiler/ml/matching.ml

@@ -2702,9 +2702,8 @@ let partial_function loc () =
               Lconst
                 (Const_block
                    ( Blk_tuple,
-                     [
-                       Const_string (fname, None); const_int line; const_int char;
-                     ] ));
+                     [const_string fname None; const_int line; const_int char]
+                   ));
             ],
             loc );
       ],

compiler/ml/printlambda.ml

@@ -20,7 +20,7 @@ open Lambda
 let rec struct_const ppf = function
   | Const_int n -> fprintf ppf "%ld" n
   | Const_char i -> fprintf ppf "%s" (Pprintast.string_of_int_as_char i)
-  | Const_string (s, _) -> fprintf ppf "%S" s
+  | Const_string {s} -> fprintf ppf "%S" s
   | Const_float f -> fprintf ppf "%s" f
   | Const_bigint (sign, n) -> fprintf ppf "%sn" (Bigint_utils.to_string sign n)
   | Const_pointer (Pt_constructor {name}) -> fprintf ppf "`%s" name

compiler/ml/transl_recmodule.ml

@@ -15,11 +15,11 @@ let undefined_location loc =
   Lconst
     (Const_block
        ( Lambda.Blk_tuple,
-         [Const_string (fname, None); const_int line; const_int char] ))
+         [const_string fname None; const_int line; const_int char] ))
 
 let init_shape modl =
   let add_name x id =
-    Const_block (Blk_tuple, [x; Const_string (Ident.name id, None)])
+    Const_block (Blk_tuple, [x; const_string (Ident.name id) None])
   in
   let module_tag_info : Lambda.tag_info =
     Blk_constructor