catala/compiler/shared_ast/typing.ml

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

1082 lines
40 KiB
OCaml
Raw Normal View History

(* This file is part of the Catala compiler, a specification language for tax
and social benefits computation rules. Copyright (C) 2020 Inria, contributor:
Denis Merigoux <denis.merigoux@inria.fr>
Licensed under the Apache License, Version 2.0 (the "License"); you may not
use this file except in compliance with the License. You may obtain a copy of
the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
License for the specific language governing permissions and limitations under
the License. *)
(** Typing for the default calculus. Because of the error terms, we perform type
inference using the classical W algorithm with union-find unification. *)
2022-11-21 12:46:17 +03:00
open Catala_utils
module A = Definitions
type resolving_strategy = LeaveAny | ErrorOnAny
module Any =
2022-11-21 12:46:17 +03:00
Uid.Make
(struct
type info = unit
let to_string _ = "any"
2022-11-21 12:46:17 +03:00
let format fmt () = Format.fprintf fmt "any"
let equal _ _ = true
let compare _ _ = 0
end)
(struct
let style = Ocolor_types.(Fg (C4 hi_magenta))
end)
()
type unionfind_typ = naked_typ Mark.pos UnionFind.elem
2023-06-13 11:48:21 +03:00
(** We do not reuse {!type: A.typ} because we have to include a new [TAny]
variant. Indeed, error terms can have any type and this has to be captured
by the type sytem. *)
and naked_typ =
| TLit of A.typ_lit
2023-01-11 13:23:08 +03:00
| TArrow of unionfind_typ list * unionfind_typ
| TTuple of unionfind_typ list
| TStruct of A.StructName.t
| TEnum of A.EnumName.t
| TOption of unionfind_typ
| TArray of unionfind_typ
2023-10-31 13:58:54 +03:00
| TDefault of unionfind_typ
| TAny of Any.t
2023-06-15 12:11:56 +03:00
| TClosureEnv
let rec typ_to_ast ~(leave_unresolved : resolving_strategy) (ty : unionfind_typ)
: A.typ =
2022-12-13 18:06:36 +03:00
let typ_to_ast = typ_to_ast ~leave_unresolved in
let ty, pos = UnionFind.get (UnionFind.find ty) in
match ty with
| TLit l -> A.TLit l, pos
| TTuple ts -> A.TTuple (List.map typ_to_ast ts), pos
| TStruct s -> A.TStruct s, pos
| TEnum e -> A.TEnum e, pos
| TOption t -> A.TOption (typ_to_ast t), pos
2023-01-11 13:23:08 +03:00
| TArrow (t1, t2) -> A.TArrow (List.map typ_to_ast t1, typ_to_ast t2), pos
| TArray t1 -> A.TArray (typ_to_ast t1), pos
2023-10-31 13:58:54 +03:00
| TDefault t1 -> A.TDefault (typ_to_ast t1), pos
| TAny _ -> (
match leave_unresolved with
| LeaveAny -> A.TAny, pos
| ErrorOnAny ->
(* No polymorphism in Catala: type inference should return full types
without wildcards, and this function is used to recover the types after
typing. *)
Message.raise_spanned_error pos
"Internal error: typing at this point could not be resolved")
2023-06-15 12:11:56 +03:00
| TClosureEnv -> TClosureEnv, pos
let rec ast_to_typ (ty : A.typ) : unionfind_typ =
let ty' =
match Mark.remove ty with
| A.TLit l -> TLit l
2023-01-11 13:23:08 +03:00
| A.TArrow (t1, t2) -> TArrow (List.map ast_to_typ t1, ast_to_typ t2)
| A.TTuple ts -> TTuple (List.map ast_to_typ ts)
| A.TStruct s -> TStruct s
| A.TEnum e -> TEnum e
| A.TOption t -> TOption (ast_to_typ t)
| A.TArray t -> TArray (ast_to_typ t)
2023-10-31 13:58:54 +03:00
| A.TDefault t -> TDefault (ast_to_typ t)
| A.TAny -> TAny (Any.fresh ())
2023-06-15 12:11:56 +03:00
| A.TClosureEnv -> TClosureEnv
in
UnionFind.make (Mark.copy ty ty')
2020-12-14 20:09:38 +03:00
(** {1 Types and unification} *)
let typ_needs_parens (t : unionfind_typ) : bool =
2020-12-30 00:26:10 +03:00
let t = UnionFind.get (UnionFind.find t) in
match Mark.remove t with TArrow _ | TArray _ -> true | _ -> false
let with_color f color fmt x =
(* equivalent to [Format.fprintf fmt "@{<color>%s@}" s] *)
Format.pp_open_stag fmt Ocolor_format.(Ocolor_style_tag (Fg (C4 color)));
f fmt x;
Format.pp_close_stag fmt ()
let pp_color_string = with_color Format.pp_print_string
2021-01-14 02:17:24 +03:00
let rec format_typ
(ctx : A.decl_ctx)
~(colors : Ocolor_types.color4 list)
2021-01-14 02:17:24 +03:00
(fmt : Format.formatter)
(naked_typ : unionfind_typ) : unit =
2021-01-14 02:17:24 +03:00
let format_typ = format_typ ctx in
let format_typ_with_parens
~colors
(fmt : Format.formatter)
(t : unionfind_typ) =
if typ_needs_parens t then (
Format.pp_open_hvbox fmt 1;
pp_color_string (List.hd colors) fmt "(";
format_typ ~colors:(List.tl colors) fmt t;
Format.pp_close_box fmt ();
pp_color_string (List.hd colors) fmt ")")
else Format.fprintf fmt "%a" (format_typ ~colors) t
2020-12-30 00:26:10 +03:00
in
let naked_typ = UnionFind.get (UnionFind.find naked_typ) in
match Mark.remove naked_typ with
| TLit l -> Format.fprintf fmt "%a" Print.tlit l
| TTuple ts ->
Format.fprintf fmt "@[<hov 2>%a%a%a@]"
(pp_color_string (List.hd colors))
"("
2021-01-14 02:17:24 +03:00
(Format.pp_print_list
~pp_sep:(fun fmt () -> Format.fprintf fmt "@ *@ ")
(fun fmt t -> format_typ fmt ~colors:(List.tl colors) t))
2020-12-03 22:11:41 +03:00
ts
(pp_color_string (List.hd colors))
")"
| TStruct s -> A.StructName.format fmt s
| TEnum e -> A.EnumName.format fmt e
| TOption t ->
Format.fprintf fmt "@[<hov 2>option %a@]"
(format_typ_with_parens ~colors:(List.tl colors))
t
| TArrow ([t1], t2) ->
Format.fprintf fmt "@[<hov 2>%a@ →@ %a@]"
(format_typ_with_parens ~colors)
t1 (format_typ ~colors) t2
2020-12-30 00:26:10 +03:00
| TArrow (t1, t2) ->
Format.fprintf fmt "@[<hov 2>%a%a%a@ →@ %a@]"
(pp_color_string (List.hd colors))
"("
2023-01-11 13:23:08 +03:00
(Format.pp_print_list
~pp_sep:(fun fmt () -> Format.fprintf fmt ",@ ")
(format_typ_with_parens ~colors:(List.tl colors)))
t1
(pp_color_string (List.hd colors))
")" (format_typ ~colors) t2
| TArray t1 -> (
match Mark.remove (UnionFind.get (UnionFind.find t1)) with
| TAny _ when not Cli.globals.debug -> Format.pp_print_string fmt "list"
| _ -> Format.fprintf fmt "@[list of@ %a@]" (format_typ ~colors) t1)
2023-10-31 13:58:54 +03:00
| TDefault t1 ->
Format.pp_print_as fmt 1 "";
format_typ ~colors fmt t1;
Format.pp_print_as fmt 1 ""
| TAny v ->
if Cli.globals.debug then Format.fprintf fmt "<a%d>" (Any.hash v)
else Format.pp_print_string fmt "<any>"
2023-06-15 12:11:56 +03:00
| TClosureEnv -> Format.fprintf fmt "closure_env"
let rec colors =
let open Ocolor_types in
blue :: cyan :: green :: yellow :: red :: magenta :: colors
let format_typ ctx fmt naked_typ = format_typ ctx ~colors fmt naked_typ
exception Type_error of A.any_expr * unionfind_typ * unionfind_typ
2022-07-11 12:32:23 +03:00
(** Raises an error if unification cannot be performed. The position annotation
of the second [unionfind_typ] argument is propagated (unless it is [TAny]). *)
2021-01-14 02:17:24 +03:00
let rec unify
(ctx : A.decl_ctx)
(e : ('a, 'm) A.gexpr) (* used for error context *)
(t1 : unionfind_typ)
(t2 : unionfind_typ) : unit =
2021-01-14 02:17:24 +03:00
let unify = unify ctx in
(* Message.emit_debug "Unifying %a and %a" (format_typ ctx) t1 (format_typ
ctx) t2; *)
let t1_repr = UnionFind.get (UnionFind.find t1) in
let t2_repr = UnionFind.get (UnionFind.find t2) in
let raise_type_error () = raise (Type_error (A.AnyExpr e, t1, t2)) in
let () =
match Mark.remove t1_repr, Mark.remove t2_repr with
| TLit tl1, TLit tl2 -> if tl1 <> tl2 then raise_type_error ()
2023-02-22 14:11:42 +03:00
| TArrow (t11, t12), TArrow (t21, t22) -> (
2022-09-16 19:15:30 +03:00
unify e t12 t22;
2023-02-22 14:11:42 +03:00
try List.iter2 (unify e) t11 t21
with Invalid_argument _ -> raise_type_error ())
| TTuple ts1, TTuple ts2 -> (
try List.iter2 (unify e) ts1 ts2
with Invalid_argument _ -> raise_type_error ())
| TStruct s1, TStruct s2 ->
if not (A.StructName.equal s1 s2) then raise_type_error ()
| TEnum e1, TEnum e2 ->
if not (A.EnumName.equal e1 e2) then raise_type_error ()
| TOption t1, TOption t2 -> unify e t1 t2
| TArray t1', TArray t2' -> unify e t1' t2'
2023-10-31 13:58:54 +03:00
| TDefault t1', TDefault t2' -> unify e t1' t2'
2023-06-15 12:11:56 +03:00
| TClosureEnv, TClosureEnv -> ()
| TAny _, _ | _, TAny _ -> ()
| ( ( TLit _ | TArrow _ | TTuple _ | TStruct _ | TEnum _ | TOption _
2023-10-31 13:58:54 +03:00
| TArray _ | TDefault _ | TClosureEnv ),
_ ) ->
raise_type_error ()
2020-12-30 03:02:04 +03:00
in
ignore
@@ UnionFind.merge
(fun t1 t2 -> match Mark.remove t2 with TAny _ -> t1 | _ -> t2)
t1 t2
let handle_type_error ctx (A.AnyExpr e) t1 t2 =
2022-07-11 12:32:23 +03:00
(* TODO: if we get weird error messages, then it means that we should use the
persistent version of the union-find data structure. *)
let t1_repr = UnionFind.get (UnionFind.find t1) in
let t2_repr = UnionFind.get (UnionFind.find t2) in
2023-11-06 18:52:01 +03:00
let e_pos = Expr.pos e in
let t1_pos = Mark.get t1_repr in
let t2_pos = Mark.get t2_repr in
2023-11-06 18:52:01 +03:00
let pos_msgs =
if e_pos = t1_pos then
2023-11-07 20:25:57 +03:00
[
( (fun ppf ->
Format.fprintf ppf "@[<hv 2>@[<hov>%a@ %a@]:" Format.pp_print_text
"This expression has type" (format_typ ctx) t1;
if Cli.globals.debug then Format.fprintf ppf "@ %a@]" Expr.format e
2023-11-07 20:25:57 +03:00
else Format.pp_close_box ppf ()),
e_pos );
( (fun ppf ->
Format.fprintf ppf
"@[<hov>Expected@ type@ %a@ coming@ from@ expression:@]"
(format_typ ctx) t2),
t2_pos );
]
else
[
( (fun ppf ->
Format.fprintf ppf "@[<hv 2>@[<hov>%a:@]" Format.pp_print_text
"While typechecking the following expression";
if Cli.globals.debug then Format.fprintf ppf "@ %a@]" Expr.format e
else Format.pp_close_box ppf ()),
e_pos );
( (fun ppf ->
Format.fprintf ppf "@[<hov>Type@ %a@ is@ coming@ from:@]"
(format_typ ctx) t1),
t1_pos );
( (fun ppf ->
Format.fprintf ppf "@[<hov>Type@ %a@ is@ coming@ from:@]"
(format_typ ctx) t2),
t2_pos );
2023-11-06 18:52:01 +03:00
]
in
Message.raise_multispanned_error_full
(List.map (fun (a, b) -> Some a, b) pos_msgs)
2023-06-07 19:10:50 +03:00
"@[<v>Error during typechecking, incompatible types:@,\
2023-06-19 18:29:51 +03:00
@[<v>@{<bold;blue>@<3>%s@} @[<hov>%a@]@,\
@{<bold;blue>@<3>%s@} @[<hov>%a@]@]@]" "" (format_typ ctx) t1 ""
2023-06-19 18:29:51 +03:00
(format_typ ctx) t2
2022-07-11 12:32:23 +03:00
let lit_type (lit : A.lit) : naked_typ =
match lit with
| LBool _ -> TLit TBool
| LInt _ -> TLit TInt
| LRat _ -> TLit TRat
| LMoney _ -> TLit TMoney
| LDate _ -> TLit TDate
| LDuration _ -> TLit TDuration
| LUnit -> TLit TUnit
(** [op_type] and [resolve_overload] are a bit similar, and work on disjoint
sets of operators. However, their assumptions are different so we keep the
functions separate. In particular [resolve_overloads] requires its argument
types to be known in advance. *)
let polymorphic_op_type (op : Operator.polymorphic A.operator Mark.pos) :
unionfind_typ =
let open Operator in
let pos = Mark.get op in
let any = lazy (UnionFind.make (TAny (Any.fresh ()), pos)) in
let any2 = lazy (UnionFind.make (TAny (Any.fresh ()), pos)) in
2024-01-24 17:36:51 +03:00
let any3 = lazy (UnionFind.make (TAny (Any.fresh ()), pos)) in
let bt = lazy (UnionFind.make (TLit TBool, pos)) in
2023-03-28 11:07:01 +03:00
let ut = lazy (UnionFind.make (TLit TUnit, pos)) in
let it = lazy (UnionFind.make (TLit TInt, pos)) in
let cet = lazy (UnionFind.make (TClosureEnv, pos)) in
let array a = lazy (UnionFind.make (TArray (Lazy.force a), pos)) in
let ( @-> ) x y =
2023-01-11 13:23:08 +03:00
lazy (UnionFind.make (TArrow (List.map Lazy.force x, Lazy.force y), pos))
in
let ty =
match Mark.remove op with
2023-01-11 13:23:08 +03:00
| Fold -> [[any2; any] @-> any2; any2; array any] @-> any2
| Eq -> [any; any] @-> bt
| Map -> [[any] @-> any2; array any] @-> array any2
2024-01-24 17:36:51 +03:00
| Map2 -> [[any; any2] @-> any3; array any; array any2] @-> array any3
2023-01-11 13:23:08 +03:00
| Filter -> [[any] @-> bt; array any] @-> array any
| Reduce -> [[any; any] @-> any; any; array any] @-> any
| Concat -> [array any; array any] @-> array any
| Log (PosRecordIfTrueBool, _) -> [bt] @-> bt
| Log _ -> [any] @-> any
| Length -> [array any] @-> it
2024-01-22 18:49:58 +03:00
(* The [HandleDefault] and [HandleDefaultOpt] need to be typed before and
after the Lcalc monomorphization which affects arrays and option types.
Because of that, we give the operators very lax typing rules with [any]
but it doesn't matter for unification because the concrete types on which
they will be instantiated are stored in the [EAppOp] node. *)
| HandleDefault -> [any2; [ut] @-> bt; [ut] @-> any] @-> any
| HandleDefaultOpt -> [any2; [ut] @-> bt; [ut] @-> any] @-> any
| ToClosureEnv -> [any] @-> cet
| FromClosureEnv -> [cet] @-> any
in
Lazy.force ty
let resolve_overload_ret_type
2022-12-13 18:06:36 +03:00
~leave_unresolved
(ctx : A.decl_ctx)
e
(op : Operator.overloaded A.operator)
tys : unionfind_typ =
let op_ty =
Operator.overload_type ctx
(Mark.add (Expr.pos e) op)
2022-12-13 18:06:36 +03:00
(List.map (typ_to_ast ~leave_unresolved) tys)
in
ast_to_typ (Type.arrow_return op_ty)
2020-12-14 20:09:38 +03:00
(** {1 Double-directed typing} *)
module Env = struct
type 'e t = {
structs : unionfind_typ A.StructField.Map.t A.StructName.Map.t;
enums : unionfind_typ A.EnumConstructor.Map.t A.EnumName.Map.t;
vars : ('e, unionfind_typ) Var.Map.t;
scope_vars : A.typ A.ScopeVar.Map.t;
scopes : A.typ A.ScopeVar.Map.t A.ScopeName.Map.t;
scopes_input : A.typ A.ScopeVar.Map.t A.ScopeName.Map.t;
toplevel_vars : A.typ A.TopdefName.Map.t;
}
2020-11-23 12:44:06 +03:00
let empty (decl_ctx : A.decl_ctx) =
(* We fill the environment initially with the structs and enums
declarations *)
{
structs =
A.StructName.Map.map
(fun ty -> A.StructField.Map.map ast_to_typ ty)
decl_ctx.ctx_structs;
enums =
A.EnumName.Map.map
(fun ty -> A.EnumConstructor.Map.map ast_to_typ ty)
decl_ctx.ctx_enums;
vars = Var.Map.empty;
scope_vars = A.ScopeVar.Map.empty;
scopes = A.ScopeName.Map.empty;
scopes_input = A.ScopeName.Map.empty;
toplevel_vars = A.TopdefName.Map.empty;
}
let get t v = Var.Map.find_opt v t.vars
let get_scope_var t sv = A.ScopeVar.Map.find_opt sv t.scope_vars
let get_toplevel_var t v = A.TopdefName.Map.find_opt v t.toplevel_vars
Make scopes directly callable Quite a few changes are included here, some of which have some extra implications visible in the language: - adds the `Scope of { -- input_v: value; ... }` construct in the language - handle it down the pipeline: * `ScopeCall` in the surface AST * `EScopeCall` in desugared and scopelang * expressions are now traversed to detect dependencies between scopes * transformed into a normal function call in dcalc - defining a scope now implicitely defines a structure with the same name, with the output variables of the scope defined as fields. This allows us to type the return value from a scope call and access its fields easily. * the implications are mostly in surface/name_resolution.ml code-wise * the `Scope_out` struct that was defined in scope_to_dcalc is no longer needed/used and the fields are no longer renamed (changes some outputs; the explicit suffix for variables with multiple states is ignored as well) * one benefit is that disambiguation works just like for structures when there are conflicts on field names * however, it's now a conflict if a scope and a structure have the same name (side-note: issues with conflicting enum / struct names or scope variables / subscope names were silent and are now properly reported) - you can consequently use scope names as types for variables as well. Writing literals is not allowed though, they can only be obtained by calling the scope. Remaining TODOs: - context variables are not handled properly at the moment - error handling on invalid calls - tests show a small error message regression; lots of examples will need tweaking to avoid scope/struct name or struct fields / output variable conflicts - add a `->` syntax to make struct field access distinct from scope output var access, enforced with typing. This is expected to reduce confusion of users and add a little typing precision. - document the new syntax & implications (tutorial, cheat-sheet) - a consequence of the changes is that subscope variables also can now be typed. A possible future evolution / simplification would be to rewrite subscopes as explicit scope calls early in the pipeline. That could also allow to manipulate them as expressions (bind them in let-ins, return them...)
2022-10-21 16:47:17 +03:00
let get_subscope_out_var t scope var =
Option.bind (A.ScopeName.Map.find_opt scope t.scopes) (fun vmap ->
A.ScopeVar.Map.find_opt var vmap)
let add v tau t = { t with vars = Var.Map.add v tau t.vars }
let add_var v typ t = add v (ast_to_typ typ) t
let add_scope_var v typ t =
{ t with scope_vars = A.ScopeVar.Map.add v typ t.scope_vars }
let add_scope scope_name ~vars ~in_vars t =
2023-11-07 20:25:57 +03:00
{
t with
scopes = A.ScopeName.Map.add scope_name vars t.scopes;
2023-11-07 20:25:57 +03:00
scopes_input = A.ScopeName.Map.add scope_name in_vars t.scopes_input;
}
let add_toplevel_var v typ t =
{ t with toplevel_vars = A.TopdefName.Map.add v typ t.toplevel_vars }
let open_scope scope_name t =
let scope_vars =
2023-12-01 01:53:38 +03:00
A.ScopeVar.Map.disjoint_union t.scope_vars
(A.ScopeName.Map.find scope_name t.scopes)
in
{ t with scope_vars }
let dump ppf env =
let pp_sep = Format.pp_print_space in
Format.pp_open_vbox ppf 0;
(* Format.fprintf ppf "structs: @[<hov>%a@]@,"
* (A.StructName.Map.format_keys ~pp_sep) env.structs;
* Format.fprintf ppf "enums: @[<hov>%a@]@,"
* (A.EnumName.Map.format_keys ~pp_sep) env.enums;
* Format.fprintf ppf "vars: @[<hov>%a@]@,"
* (Var.Map.format_keys ~pp_sep) env.vars; *)
Format.fprintf ppf "scopes: @[<hov>%a@]@,"
(A.ScopeName.Map.format_keys ~pp_sep)
env.scopes;
Format.fprintf ppf "topdefs: @[<hov>%a@]@,"
(A.TopdefName.Map.format_keys ~pp_sep)
env.toplevel_vars;
Format.pp_close_box ppf ()
end
let add_pos e ty = Mark.add (Expr.pos e) ty
let ty : (_, unionfind_typ A.custom) A.marked -> unionfind_typ =
fun (_, A.Custom { A.custom; _ }) -> custom
2020-12-14 20:09:38 +03:00
(** Infers the most permissive type from an expression *)
let rec typecheck_expr_bottom_up :
type a m.
leave_unresolved:resolving_strategy ->
A.decl_ctx ->
(a, m) A.gexpr Env.t ->
(a, m) A.gexpr ->
(a, unionfind_typ A.custom) A.boxed_gexpr =
2022-12-13 18:06:36 +03:00
fun ~leave_unresolved ctx env e ->
typecheck_expr_top_down ~leave_unresolved ctx env
2022-10-07 16:59:10 +03:00
(UnionFind.make (add_pos e (TAny (Any.fresh ()))))
e
2021-01-13 14:04:14 +03:00
(** Checks whether the expression can be typed with the provided type *)
and typecheck_expr_top_down :
type a m.
leave_unresolved:resolving_strategy ->
A.decl_ctx ->
(a, m) A.gexpr Env.t ->
unionfind_typ ->
(a, m) A.gexpr ->
(a, unionfind_typ A.custom) A.boxed_gexpr =
2022-12-13 18:06:36 +03:00
fun ~leave_unresolved ctx env tau e ->
(* Message.emit_debug "Propagating type %a for naked_expr :@.@[<hov 2>%a@]"
(format_typ ctx) tau Expr.format e; *)
let pos_e = Expr.pos e in
let () =
(* If there already is a type annotation on the given expr, ensure it
matches *)
match Mark.get e with
| A.Untyped _ | A.Typed { A.ty = A.TAny, _; _ } -> ()
| A.Typed { A.ty; _ } -> unify ctx e tau (ast_to_typ ty)
| A.Custom _ -> assert false
in
let context_mark = A.Custom { A.custom = tau; pos = pos_e } in
let mark_with_tau_and_unify uf =
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
(* Unify with the supplied type first, and return the mark *)
unify ctx e uf tau;
A.Custom { A.custom = uf; pos = pos_e }
2022-07-11 12:34:01 +03:00
in
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
let unionfind ?(pos = e) t = UnionFind.make (add_pos pos t) in
let ty_mark ty = mark_with_tau_and_unify (unionfind ty) in
match Mark.remove e with
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
| A.ELocation loc ->
let ty_opt =
match loc with
| DesugaredScopeVar { name; _ } | ScopelangScopeVar { name } ->
Env.get_scope_var env (Mark.remove name)
| SubScopeVar { scope; var; _ } ->
Env.get_subscope_out_var env scope (Mark.remove var)
2023-12-01 01:53:38 +03:00
| ToplevelVar { name } -> Env.get_toplevel_var env (Mark.remove name)
in
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
let ty =
match ty_opt with
| Some ty -> ty
| None ->
Message.raise_spanned_error pos_e "Reference to %a not found"
(Print.expr ()) e
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
in
Expr.elocation loc (mark_with_tau_and_unify (ast_to_typ ty))
| A.EStruct { name; fields } ->
let mark = ty_mark (TStruct name) in
let str_ast = A.StructName.Map.find name ctx.A.ctx_structs in
let str = A.StructName.Map.find name env.structs in
let _check_fields : unit =
let missing_fields, extra_fields =
A.StructField.Map.fold
(fun fld x (remaining, extra) ->
if A.StructField.Map.mem fld remaining then
A.StructField.Map.remove fld remaining, extra
else remaining, A.StructField.Map.add fld x extra)
fields
(str_ast, A.StructField.Map.empty)
in
let errs =
List.map
(fun (f, ty) ->
( Some (Format.asprintf "Missing field %a" A.StructField.format f),
Mark.get ty ))
(A.StructField.Map.bindings missing_fields)
@ List.map
(fun (f, ef) ->
let dup = A.StructField.Map.mem f str in
( Some
(Format.asprintf "%s field %a"
(if dup then "Duplicate" else "Unknown")
A.StructField.format f),
Expr.pos ef ))
(A.StructField.Map.bindings extra_fields)
in
if errs <> [] then
Message.raise_multispanned_error errs
"Mismatching field definitions for structure %a" A.StructName.format
name
in
let fields =
A.StructField.Map.mapi
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
(fun f_name f_e ->
let f_ty = A.StructField.Map.find f_name str in
typecheck_expr_top_down ~leave_unresolved ctx env f_ty f_e)
fields
in
Expr.estruct ~name ~fields mark
| A.EDStructAccess { e = e_struct; name_opt; field } ->
let t_struct =
match name_opt with
| Some name -> TStruct name
| None -> TAny (Any.fresh ())
in
let e_struct' =
2022-12-13 18:06:36 +03:00
typecheck_expr_top_down ~leave_unresolved ctx env (unionfind t_struct)
e_struct
in
let name =
match UnionFind.get (ty e_struct') with
| TStruct name, _ -> name
| TAny _, _ ->
Printf.ksprintf failwith
"Disambiguation failed before reaching field %s" field
| _ ->
Message.raise_spanned_error (Expr.pos e)
"This is not a structure, cannot access field %s (%a)" field
(format_typ ctx) (ty e_struct')
in
let str =
try A.StructName.Map.find name env.structs
with A.StructName.Map.Not_found _ ->
Message.raise_spanned_error pos_e "No structure %a found"
A.StructName.format name
in
let field =
let candidate_structs =
try A.Ident.Map.find field ctx.ctx_struct_fields
with A.Ident.Map.Not_found _ ->
Message.raise_spanned_error
(Expr.mark_pos context_mark)
"Field @{<yellow>\"%s\"@} does not belong to structure \
@{<yellow>\"%a\"@} (no structure defines it)"
field A.StructName.format name
in
try A.StructName.Map.find name candidate_structs
with A.StructName.Map.Not_found _ ->
Message.raise_spanned_error
(Expr.mark_pos context_mark)
"@[<hov>Field @{<yellow>\"%s\"@}@ does not belong to@ structure \
@{<yellow>\"%a\"@},@ but to %a@]"
field A.StructName.format name
(Format.pp_print_list
~pp_sep:(fun ppf () -> Format.fprintf ppf "@ or@ ")
(fun fmt s_name ->
2023-12-01 01:53:38 +03:00
Format.fprintf fmt "@{<yellow>\"%a\"@}" A.StructName.format
s_name))
(A.StructName.Map.keys candidate_structs)
in
let fld_ty = A.StructField.Map.find field str in
let mark = mark_with_tau_and_unify fld_ty in
Expr.estructaccess ~name ~e:e_struct' ~field mark
| A.EStructAccess { e = e_struct; name; field } ->
let fld_ty =
let str =
try A.StructName.Map.find name env.structs
with A.StructName.Map.Not_found _ ->
Message.raise_spanned_error pos_e "No structure %a found"
A.StructName.format name
in
try A.StructField.Map.find field str
with A.StructField.Map.Not_found _ ->
Message.raise_multispanned_error
[
None, pos_e;
( Some "Structure %a declared here",
Mark.get (A.StructName.get_info name) );
]
"Structure %a doesn't define a field %a" A.StructName.format name
A.StructField.format field
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
in
let mark = mark_with_tau_and_unify fld_ty in
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
let e_struct' =
2022-12-13 18:06:36 +03:00
typecheck_expr_top_down ~leave_unresolved ctx env
(unionfind (TStruct name)) e_struct
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
in
Expr.estructaccess ~e:e_struct' ~field ~name mark
| A.EInj { name; cons; e = e_enum }
when Definitions.EnumName.equal name Expr.option_enum ->
if Definitions.EnumConstructor.equal cons Expr.some_constr then
let cell_type = unionfind (TAny (Any.fresh ())) in
let mark = mark_with_tau_and_unify (unionfind (TOption cell_type)) in
let e_enum' =
typecheck_expr_top_down ~leave_unresolved ctx env cell_type e_enum
in
Expr.einj ~name ~cons ~e:e_enum' mark
else
(* None constructor *)
let cell_type = unionfind (TAny (Any.fresh ())) in
let mark = mark_with_tau_and_unify (unionfind (TOption cell_type)) in
let e_enum' =
typecheck_expr_top_down ~leave_unresolved ctx env
(unionfind (TLit TUnit)) e_enum
in
Expr.einj ~name ~cons ~e:e_enum' mark
| A.EInj { name; cons; e = e_enum } ->
let mark = mark_with_tau_and_unify (unionfind (TEnum name)) in
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
let e_enum' =
2022-12-13 18:06:36 +03:00
typecheck_expr_top_down ~leave_unresolved ctx env
(A.EnumConstructor.Map.find cons (A.EnumName.Map.find name env.enums))
2022-09-26 19:19:39 +03:00
e_enum
in
Expr.einj ~e:e_enum' ~cons ~name mark
| A.EMatch { e = e1; name; cases }
when Definitions.EnumName.equal name Expr.option_enum ->
let cell_type = unionfind ~pos:e1 (TAny (Any.fresh ())) in
let t_arg = unionfind ~pos:e1 (TOption cell_type) in
let cases_ty =
ListLabels.fold_right2
[Expr.none_constr; Expr.some_constr]
[unionfind ~pos:e1 (TLit TUnit); cell_type]
~f:A.EnumConstructor.Map.add ~init:A.EnumConstructor.Map.empty
in
let t_ret = unionfind ~pos:e (TAny (Any.fresh ())) in
let mark = mark_with_tau_and_unify t_ret in
2023-03-30 12:02:35 +03:00
let e1' = typecheck_expr_top_down ~leave_unresolved ctx env t_arg e1 in
let cases =
A.EnumConstructor.Map.merge
(fun _ e e_ty ->
match e, e_ty with
| Some e, Some e_ty ->
Some
2023-03-30 12:02:35 +03:00
(typecheck_expr_top_down ~leave_unresolved ctx env
(unionfind ~pos:e (TArrow ([e_ty], t_ret)))
e)
| _ -> assert false)
cases cases_ty
in
Expr.ematch ~e:e1' ~name ~cases mark
| A.EMatch { e = e1; name; cases } ->
let cases_ty = A.EnumName.Map.find name ctx.A.ctx_enums in
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
let t_ret = unionfind ~pos:e1 (TAny (Any.fresh ())) in
let mark = mark_with_tau_and_unify t_ret in
2022-12-13 18:06:36 +03:00
let e1' =
typecheck_expr_top_down ~leave_unresolved ctx env (unionfind (TEnum name))
2022-12-13 18:06:36 +03:00
e1
in
let cases =
A.EnumConstructor.Map.mapi
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
(fun c_name e ->
let c_ty = A.EnumConstructor.Map.find c_name cases_ty in
(* For now our constructors are limited to zero or one argument. If
there is a change to allow for multiple arguments, it might be
easier to use tuples directly. *)
2023-01-11 13:23:08 +03:00
let e_ty = unionfind ~pos:e (TArrow ([ast_to_typ c_ty], t_ret)) in
2022-12-13 18:06:36 +03:00
typecheck_expr_top_down ~leave_unresolved ctx env e_ty e)
cases
in
Expr.ematch ~e:e1' ~name ~cases mark
| A.EScopeCall { scope; args } ->
let scope_out_struct =
(A.ScopeName.Map.find scope ctx.ctx_scopes).out_struct_name
in
let mark = mark_with_tau_and_unify (unionfind (TStruct scope_out_struct)) in
let vars = A.ScopeName.Map.find scope env.scopes_input in
let args' =
A.ScopeVar.Map.mapi
Make scopes directly callable Quite a few changes are included here, some of which have some extra implications visible in the language: - adds the `Scope of { -- input_v: value; ... }` construct in the language - handle it down the pipeline: * `ScopeCall` in the surface AST * `EScopeCall` in desugared and scopelang * expressions are now traversed to detect dependencies between scopes * transformed into a normal function call in dcalc - defining a scope now implicitely defines a structure with the same name, with the output variables of the scope defined as fields. This allows us to type the return value from a scope call and access its fields easily. * the implications are mostly in surface/name_resolution.ml code-wise * the `Scope_out` struct that was defined in scope_to_dcalc is no longer needed/used and the fields are no longer renamed (changes some outputs; the explicit suffix for variables with multiple states is ignored as well) * one benefit is that disambiguation works just like for structures when there are conflicts on field names * however, it's now a conflict if a scope and a structure have the same name (side-note: issues with conflicting enum / struct names or scope variables / subscope names were silent and are now properly reported) - you can consequently use scope names as types for variables as well. Writing literals is not allowed though, they can only be obtained by calling the scope. Remaining TODOs: - context variables are not handled properly at the moment - error handling on invalid calls - tests show a small error message regression; lots of examples will need tweaking to avoid scope/struct name or struct fields / output variable conflicts - add a `->` syntax to make struct field access distinct from scope output var access, enforced with typing. This is expected to reduce confusion of users and add a little typing precision. - document the new syntax & implications (tutorial, cheat-sheet) - a consequence of the changes is that subscope variables also can now be typed. A possible future evolution / simplification would be to rewrite subscopes as explicit scope calls early in the pipeline. That could also allow to manipulate them as expressions (bind them in let-ins, return them...)
2022-10-21 16:47:17 +03:00
(fun name ->
2022-12-13 18:06:36 +03:00
typecheck_expr_top_down ~leave_unresolved ctx env
(ast_to_typ (A.ScopeVar.Map.find name vars)))
args
Make scopes directly callable Quite a few changes are included here, some of which have some extra implications visible in the language: - adds the `Scope of { -- input_v: value; ... }` construct in the language - handle it down the pipeline: * `ScopeCall` in the surface AST * `EScopeCall` in desugared and scopelang * expressions are now traversed to detect dependencies between scopes * transformed into a normal function call in dcalc - defining a scope now implicitely defines a structure with the same name, with the output variables of the scope defined as fields. This allows us to type the return value from a scope call and access its fields easily. * the implications are mostly in surface/name_resolution.ml code-wise * the `Scope_out` struct that was defined in scope_to_dcalc is no longer needed/used and the fields are no longer renamed (changes some outputs; the explicit suffix for variables with multiple states is ignored as well) * one benefit is that disambiguation works just like for structures when there are conflicts on field names * however, it's now a conflict if a scope and a structure have the same name (side-note: issues with conflicting enum / struct names or scope variables / subscope names were silent and are now properly reported) - you can consequently use scope names as types for variables as well. Writing literals is not allowed though, they can only be obtained by calling the scope. Remaining TODOs: - context variables are not handled properly at the moment - error handling on invalid calls - tests show a small error message regression; lots of examples will need tweaking to avoid scope/struct name or struct fields / output variable conflicts - add a `->` syntax to make struct field access distinct from scope output var access, enforced with typing. This is expected to reduce confusion of users and add a little typing precision. - document the new syntax & implications (tutorial, cheat-sheet) - a consequence of the changes is that subscope variables also can now be typed. A possible future evolution / simplification would be to rewrite subscopes as explicit scope calls early in the pipeline. That could also allow to manipulate them as expressions (bind them in let-ins, return them...)
2022-10-21 16:47:17 +03:00
in
Expr.escopecall ~scope ~args:args' mark
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
| A.ERaise ex -> Expr.eraise ex context_mark
| A.ECatch { body; exn; handler } ->
let body' = typecheck_expr_top_down ~leave_unresolved ctx env tau body in
2022-12-13 18:06:36 +03:00
let handler' =
typecheck_expr_top_down ~leave_unresolved ctx env tau handler
in
Expr.ecatch body' exn handler' context_mark
| A.EVar v ->
let tau' =
match Env.get env v with
| Some t -> t
| None ->
Message.raise_spanned_error pos_e
"Variable %s not found in the current context" (Bindlib.name_of v)
in
Expr.evar (Var.translate v) (mark_with_tau_and_unify tau')
| A.EExternal { name } ->
let ty =
let not_found pr x =
Message.raise_spanned_error pos_e
"Could not resolve the reference to %a.@ Make sure the corresponding \
module was properly loaded?"
pr x
in
match Mark.remove name with
| A.External_value name -> (
try ast_to_typ (A.TopdefName.Map.find name ctx.ctx_topdefs)
with A.TopdefName.Map.Not_found _ ->
not_found A.TopdefName.format name)
| A.External_scope name -> (
try
let scope_info = A.ScopeName.Map.find name ctx.ctx_scopes in
ast_to_typ
( TArrow
( [TStruct scope_info.in_struct_name, pos_e],
(TStruct scope_info.out_struct_name, pos_e) ),
pos_e )
with A.ScopeName.Map.Not_found _ -> not_found A.ScopeName.format name)
in
Expr.eexternal ~name (mark_with_tau_and_unify ty)
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
| A.ELit lit -> Expr.elit lit (ty_mark (lit_type lit))
| A.ETuple es ->
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
let tys = List.map (fun _ -> unionfind (TAny (Any.fresh ()))) es in
let mark = mark_with_tau_and_unify (unionfind (TTuple tys)) in
2022-12-13 18:06:36 +03:00
let es' =
List.map2 (typecheck_expr_top_down ~leave_unresolved ctx env) tys es
in
Expr.etuple es' mark
| A.ETupleAccess { e = e1; index; size } ->
if index >= size then
Message.raise_spanned_error (Expr.pos e)
"Tuple access out of bounds (%d/%d)" index size;
let tuple_ty =
TTuple
(List.init size (fun n ->
if n = index then tau else unionfind ~pos:e1 (TAny (Any.fresh ()))))
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
in
2022-12-13 18:06:36 +03:00
let e1' =
typecheck_expr_top_down ~leave_unresolved ctx env
(unionfind ~pos:e1 tuple_ty)
e1
in
Expr.etupleaccess ~e:e1' ~index ~size context_mark
| A.EAbs { binder; tys = t_args } ->
2022-06-23 15:04:51 +03:00
if Bindlib.mbinder_arity binder <> List.length t_args then
Message.raise_spanned_error (Expr.pos e)
"function has %d variables but was supplied %d types\n%a"
2022-06-23 15:04:51 +03:00
(Bindlib.mbinder_arity binder)
(List.length t_args) Expr.format e
else
let tau_args = List.map ast_to_typ t_args in
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
let t_ret = unionfind (TAny (Any.fresh ())) in
2023-01-11 13:23:08 +03:00
let t_func = unionfind (TArrow (tau_args, t_ret)) in
let mark = mark_with_tau_and_unify t_func in
2022-06-23 15:04:51 +03:00
let xs, body = Bindlib.unmbind binder in
let xs' = Array.map Var.translate xs in
2022-06-23 15:04:51 +03:00
let env =
List.fold_left2
(fun env x tau_arg -> Env.add x tau_arg env)
env (Array.to_list xs) tau_args
in
2022-12-13 18:06:36 +03:00
let body' =
typecheck_expr_top_down ~leave_unresolved ctx env t_ret body
in
let binder' = Bindlib.bind_mvar xs' (Expr.Box.lift body') in
2022-12-13 19:21:37 +03:00
Expr.eabs binder' (List.map (typ_to_ast ~leave_unresolved) tau_args) mark
| A.EApp { f = e1; args; tys } ->
(* Here we type the arguments first (in order), to ensure we know the types
of the arguments if [f] is [EAbs] before disambiguation. This is also the
right order for the [let-in] form. *)
let t_args =
match tys with
| [] -> List.map (fun _ -> unionfind (TAny (Any.fresh ()))) args
| tys -> List.map ast_to_typ tys
in
let args' =
List.map2 (typecheck_expr_top_down ~leave_unresolved ctx env) t_args args
in
let t_args =
2023-12-19 16:44:01 +03:00
match t_args, tys with
| [t], [] -> (
(* Handles typing before detuplification: if [tys] was not yet set, we
are allowed to destruct a tuple into multiple arguments (see
[Scopelang.from_desugared] for the corresponding code
transformation) *)
match UnionFind.get t with TTuple tys, _ -> tys | _ -> t_args)
2023-12-19 16:44:01 +03:00
| _ ->
if List.length t_args <> List.length args' then
Message.raise_spanned_error (Expr.pos e)
(match e1 with
| EAbs _, _ -> "This binds %d variables, but %d were provided."
| _ -> "This function application has %d arguments, but expects %d.")
(List.length t_args) (List.length args');
t_args
in
let t_func = unionfind ~pos:e1 (TArrow (t_args, tau)) in
let e1' = typecheck_expr_top_down ~leave_unresolved ctx env t_func e1 in
Expr.eapp ~f:e1' ~args:args'
~tys:(List.map (typ_to_ast ~leave_unresolved) t_args)
context_mark
| A.EAppOp { op; tys; args } ->
let t_args = List.map ast_to_typ tys in
2023-01-11 13:23:08 +03:00
let t_func = unionfind (TArrow (t_args, tau)) in
let args =
Operator.kind_dispatch op
~polymorphic:(fun op ->
(* Type the operator first, then right-to-left: polymorphic operators
are required to allow the resolution of all type variables this
way *)
unify ctx e (polymorphic_op_type (Mark.add pos_e op)) t_func;
List.rev_map2
(typecheck_expr_top_down ~leave_unresolved ctx env)
(List.rev t_args) (List.rev args))
~overloaded:(fun op ->
(* Typing the arguments first is required to resolve the operator *)
2022-12-13 18:06:36 +03:00
let args' =
List.map2
(typecheck_expr_top_down ~leave_unresolved ctx env)
t_args args
in
unify ctx e tau
(resolve_overload_ret_type ~leave_unresolved ctx e op t_args);
args')
~monomorphic:(fun op ->
(* Here it doesn't matter but may affect the error messages *)
unify ctx e
(ast_to_typ (Operator.monomorphic_type (Mark.add pos_e op)))
t_func;
List.map2
(typecheck_expr_top_down ~leave_unresolved ctx env)
t_args args)
~resolved:(fun op ->
(* This case should not fail *)
unify ctx e
(ast_to_typ (Operator.resolved_type (Mark.add pos_e op)))
t_func;
List.map2
(typecheck_expr_top_down ~leave_unresolved ctx env)
t_args args)
in
(* All operator applications are monomorphised at this point *)
let tys = List.map (typ_to_ast ~leave_unresolved) t_args in
Expr.eappop ~op ~args ~tys context_mark
| A.EDefault { excepts; just; cons } ->
let cons' = typecheck_expr_top_down ~leave_unresolved ctx env tau cons in
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
let just' =
2022-12-13 18:06:36 +03:00
typecheck_expr_top_down ~leave_unresolved ctx env
(unionfind ~pos:just (TLit TBool))
just
in
let excepts' =
List.map (typecheck_expr_top_down ~leave_unresolved ctx env tau) excepts
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
in
Expr.edefault ~excepts:excepts' ~just:just' ~cons:cons' context_mark
| A.EPureDefault e1 ->
let inner_ty = unionfind ~pos:e1 (TAny (Any.fresh ())) in
let mark =
mark_with_tau_and_unify (unionfind ~pos:e1 (TDefault inner_ty))
in
let e1' = typecheck_expr_top_down ~leave_unresolved ctx env inner_ty e1 in
Expr.epuredefault e1' mark
| A.EIfThenElse { cond; etrue = et; efalse = ef } ->
2022-12-13 18:06:36 +03:00
let et' = typecheck_expr_top_down ~leave_unresolved ctx env tau et in
let ef' = typecheck_expr_top_down ~leave_unresolved ctx env tau ef in
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
let cond' =
2022-12-13 18:06:36 +03:00
typecheck_expr_top_down ~leave_unresolved ctx env
(unionfind ~pos:cond (TLit TBool))
cond
in
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
Expr.eifthenelse cond' et' ef' context_mark
| A.EAssert e1 ->
let mark = mark_with_tau_and_unify (unionfind (TLit TUnit)) in
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
let e1' =
2022-12-13 18:06:36 +03:00
typecheck_expr_top_down ~leave_unresolved ctx env
(unionfind ~pos:e1 (TLit TBool))
e1
2022-07-11 12:32:23 +03:00
in
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
Expr.eassert e1' mark
2023-10-31 13:58:54 +03:00
| A.EEmptyError ->
Expr.eemptyerror (ty_mark (TDefault (unionfind (TAny (Any.fresh ())))))
| A.EErrorOnEmpty e1 ->
2023-10-31 13:58:54 +03:00
let tau' = unionfind (TDefault tau) in
let e1' = typecheck_expr_top_down ~leave_unresolved ctx env tau' e1 in
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
Expr.eerroronempty e1' context_mark
| A.EArray es ->
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
let cell_type = unionfind (TAny (Any.fresh ())) in
let mark = mark_with_tau_and_unify (unionfind (TArray cell_type)) in
2022-12-13 18:06:36 +03:00
let es' =
List.map (typecheck_expr_top_down ~leave_unresolved ctx env cell_type) es
in
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
Expr.earray es' mark
| A.ECustom { obj; targs; tret } ->
let mark =
mark_with_tau_and_unify (ast_to_typ (A.TArrow (targs, tret), Expr.pos e))
in
Expr.ecustom obj targs tret mark
2022-07-11 12:32:23 +03:00
let wrap ctx f e =
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
try f e
2022-07-11 12:32:23 +03:00
with Type_error (e, ty1, ty2) -> (
let bt = Printexc.get_raw_backtrace () in
try handle_type_error ctx e ty1 ty2
with e -> Printexc.raise_with_backtrace e bt)
2020-11-23 12:44:06 +03:00
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
let wrap_expr ctx f e =
(* We need to unbox here, because the typing may otherwise be stored in
Bindlib closures and not yet applied, and would escape the `try..with` *)
wrap ctx (fun e -> Expr.unbox (f e)) e
2020-12-14 20:09:38 +03:00
(** {1 API} *)
let get_ty_mark ~leave_unresolved (A.Custom { A.custom = uf; pos }) =
2022-12-13 18:06:36 +03:00
A.Typed { ty = typ_to_ast ~leave_unresolved uf; pos }
let expr_raw
(type a)
~(leave_unresolved : resolving_strategy)
(ctx : A.decl_ctx)
?(env = Env.empty ctx)
?(typ : A.typ option)
(e : (a, 'm) A.gexpr) : (a, unionfind_typ A.custom) A.gexpr =
let fty =
match typ with
2022-12-13 18:06:36 +03:00
| None -> typecheck_expr_bottom_up ~leave_unresolved ctx env
| Some typ ->
typecheck_expr_top_down ~leave_unresolved ctx env (ast_to_typ typ)
in
wrap_expr ctx fty e
2022-12-13 18:06:36 +03:00
let check_expr ~leave_unresolved ctx ?env ?typ e =
Expr.map_marks
~f:(fun (Custom { pos; _ }) -> A.Untyped { pos })
2022-12-13 18:06:36 +03:00
(expr_raw ctx ~leave_unresolved ?env ?typ e)
(* Infer the type of an expression *)
2022-12-13 18:06:36 +03:00
let expr ~leave_unresolved ctx ?env ?typ e =
Expr.map_marks
~f:(get_ty_mark ~leave_unresolved)
(expr_raw ~leave_unresolved ctx ?env ?typ e)
2022-12-13 18:06:36 +03:00
let rec scope_body_expr ~leave_unresolved ctx env ty_out body_expr =
match body_expr with
| A.Result e ->
2022-12-13 18:06:36 +03:00
let e' =
wrap_expr ctx (typecheck_expr_top_down ~leave_unresolved ctx env ty_out) e
in
let e' = Expr.map_marks ~f:(get_ty_mark ~leave_unresolved) e' in
Bindlib.box_apply (fun e -> A.Result e) (Expr.Box.lift e')
| A.ScopeLet
{
scope_let_kind;
scope_let_typ;
scope_let_expr = e0;
scope_let_next;
scope_let_pos;
} ->
let ty_e = ast_to_typ scope_let_typ in
2022-12-13 18:06:36 +03:00
let e =
wrap_expr ctx (typecheck_expr_bottom_up ~leave_unresolved ctx env) e0
in
Swap boxing and annotations in expressions This was the only reasonable solution I found to the issue raised [here](https://github.com/CatalaLang/catala/pull/334#discussion_r987175884). This was a pretty tedious rewrite, but it should now ensure we are doing things correctly. As a bonus, the "smart" expression constructors are now used everywhere to build expressions (so another refactoring like this one should be much easier) and this makes the code overall feel more straightforward (`Bindlib.box_apply` or `let+` no longer need to be visible!) --- Basically, we were using values of type `gexpr box = naked_gexpr marked box` throughout when (re-)building expressions. This was done 99% of the time by using `Bindlib.box_apply add_mark naked_e` right after building `naked_e`. In lots of places, we needed to recover the annotation of this expression later on, typically to build its parent term (to inherit the position, or build the type). Since it wasn't always possible to wrap these uses within `box_apply` (esp. as bindlib boxes aren't a monad), here and there we had to call `Bindlib.unbox`, just to recover the position or type. This had the very unpleasant effect of forcing the resolution of the whole box (including applying any stored closures) to reach the top-level annotation which isn't even dependant on specific variable bindings. Then, generally, throwing away the result. Therefore, the change proposed here transforms - `naked_gexpr marked Bindlib.box` into - `naked_gexpr Bindlib.box marked` (aliased to `boxed_gexpr` or `gexpr boxed` for convenience) This means only 1. not fitting the mark into the box right away when building, and 2. accessing the top-level mark directly without unboxing The functions for building terms from module `Shared_ast.Expr` could be changed easily. But then they needed to be consistently used throughout, without manually building terms through `Bindlib.apply_box` -- which covers most of the changes in this patch. `Expr.Box.inj` is provided to swap back to a box, before binding for example. Additionally, this gives a 40% speedup on `make -C examples pass_all_tests`, which hints at the amount of unnecessary work we were doing --'
2022-10-06 20:13:45 +03:00
wrap ctx (fun t -> unify ctx e0 (ty e) t) ty_e;
(* We could use [typecheck_expr_top_down] rather than this manual
unification, but we get better messages with this order of the [unify]
parameters, which keeps location of the type as defined instead of as
inferred. *)
let var, next = Bindlib.unbind scope_let_next in
let env = Env.add var ty_e env in
2022-12-13 18:06:36 +03:00
let next = scope_body_expr ~leave_unresolved ctx env ty_out next in
let scope_let_next = Bindlib.bind_var (Var.translate var) next in
Bindlib.box_apply2
(fun scope_let_expr scope_let_next ->
A.ScopeLet
{
scope_let_kind;
2022-12-13 16:12:43 +03:00
scope_let_typ =
(match Mark.remove scope_let_typ with
2022-12-13 18:06:36 +03:00
| TAny -> typ_to_ast ~leave_unresolved (ty e)
2022-12-13 16:12:43 +03:00
| _ -> scope_let_typ);
scope_let_expr;
scope_let_next;
scope_let_pos;
})
2022-12-13 18:06:36 +03:00
(Expr.Box.lift (Expr.map_marks ~f:(get_ty_mark ~leave_unresolved) e))
scope_let_next
2022-12-13 18:06:36 +03:00
let scope_body ~leave_unresolved ctx env body =
let get_pos struct_name = Mark.get (A.StructName.get_info struct_name) in
let struct_ty struct_name =
UnionFind.make (Mark.add (get_pos struct_name) (TStruct struct_name))
in
let ty_in = struct_ty body.A.scope_body_input_struct in
let ty_out = struct_ty body.A.scope_body_output_struct in
let var, e = Bindlib.unbind body.A.scope_body_expr in
let env = Env.add var ty_in env in
2022-12-13 18:06:36 +03:00
let e' = scope_body_expr ~leave_unresolved ctx env ty_out e in
( Bindlib.box_apply
(fun scope_body_expr -> { body with scope_body_expr })
(Bindlib.bind_var (Var.translate var) e'),
UnionFind.make
(Mark.add
(get_pos body.A.scope_body_output_struct)
2023-01-11 13:23:08 +03:00
(TArrow ([ty_in], ty_out))) )
2022-12-13 18:06:36 +03:00
let rec scopes ~leave_unresolved ctx env = function
2023-06-13 11:48:21 +03:00
| A.Nil -> Bindlib.box A.Nil, env
| A.Cons (item, next_bind) ->
let var, next = Bindlib.unbind next_bind in
let env, def =
match item with
| A.ScopeDef (name, body) ->
2023-03-21 14:14:10 +03:00
let body_e, ty_scope = scope_body ~leave_unresolved ctx env body in
( Env.add var ty_scope env,
Bindlib.box_apply (fun body -> A.ScopeDef (name, body)) body_e )
| A.Topdef (name, typ, e) ->
2023-03-21 14:14:10 +03:00
let e' = expr_raw ~leave_unresolved ctx ~env ~typ e in
let (A.Custom { custom = uf; _ }) = Mark.get e' in
2023-03-21 14:14:10 +03:00
let e' = Expr.map_marks ~f:(get_ty_mark ~leave_unresolved) e' in
( Env.add var uf env,
Bindlib.box_apply
(fun e -> A.Topdef (name, Expr.ty e', e))
(Expr.Box.lift e') )
in
2023-06-13 11:48:21 +03:00
let next', env = scopes ~leave_unresolved ctx env next in
let next_bind' = Bindlib.bind_var (Var.translate var) next' in
2023-06-13 11:48:21 +03:00
( Bindlib.box_apply2 (fun item next -> A.Cons (item, next)) def next_bind',
env )
2022-12-13 18:06:36 +03:00
let program ~leave_unresolved prg =
2023-06-13 11:48:21 +03:00
let code_items, new_env =
scopes ~leave_unresolved prg.A.decl_ctx (Env.empty prg.A.decl_ctx)
prg.A.code_items
in
2023-06-13 11:48:21 +03:00
{
A.lang = prg.lang;
A.module_name = prg.A.module_name;
2023-06-13 11:48:21 +03:00
A.code_items = Bindlib.unbox code_items;
decl_ctx =
{
prg.decl_ctx with
ctx_structs =
A.StructName.Map.mapi
(fun s_name fields ->
A.StructField.Map.mapi
2023-06-13 11:48:21 +03:00
(fun f_name (t : A.typ) ->
match Mark.remove t with
| TAny ->
typ_to_ast ~leave_unresolved
(A.StructField.Map.find f_name
(A.StructName.Map.find s_name new_env.structs))
| _ -> t)
fields)
prg.decl_ctx.ctx_structs;
ctx_enums =
A.EnumName.Map.mapi
(fun e_name cons ->
A.EnumConstructor.Map.mapi
2023-06-13 11:48:21 +03:00
(fun cons_name (t : A.typ) ->
match Mark.remove t with
| TAny ->
typ_to_ast ~leave_unresolved
(A.EnumConstructor.Map.find cons_name
(A.EnumName.Map.find e_name new_env.enums))
| _ -> t)
cons)
prg.decl_ctx.ctx_enums;
};
}