add a test case;some clean up before unicode (rescript-lang#5043)

* clean up string API to prepare unicode support

* add a complex test case

* fix CI

Author: bobzhang

jscomp/build_tests/ocamlgraph/libx/bitv.ml

@@ -451,7 +451,7 @@ let to_string v =
   let n = v.length in
   let s = Bytes.make n '0' in
   for i = 0 to n - 1 do
-    if unsafe_get v i then s.[i] <- '1'
+    if unsafe_get v i then Bytes.set s i  '1'
   done;
   Bytes.to_string s
 

jscomp/main/builtin_cmi_datasets.ml

@@ -23,14 +23,10 @@
 
 external length : string -> int = "%string_length"
 external get : string -> int -> char = "%string_safe_get"
-external set : bytes -> int -> char -> unit = "%bytes_safe_set"
-external create : int -> bytes = "caml_create_bytes"
 external unsafe_get : string -> int -> char = "%string_unsafe_get"
-external unsafe_set : bytes -> int -> char -> unit = "%bytes_unsafe_set"
+
 external unsafe_blit : string -> int ->  bytes -> int -> int -> unit
                      = "caml_blit_string" [@@noalloc]
-external unsafe_fill : bytes -> int -> int -> char -> unit
-                     = "caml_fill_bytes" [@@noalloc]
 
 module B = Bytes
 
@@ -41,12 +37,8 @@ external make : int -> char -> string = "caml_string_repeat"
 
 let init n f =
   B.init n f |> bts
-let copy s =
-  B.copy (bos s) |> bts
 let sub s ofs len =
   B.sub (bos s) ofs len |> bts
-let fill =
-  B.fill
 let blit =
   B.blit_string
 

jscomp/main/builtin_cmj_datasets.ml

@@ -58,24 +58,7 @@ external get : string -> int -> char = "%string_safe_get"
    Raise [Invalid_argument] if [n] not a valid index in [s]. *)
 
 
-external set : bytes -> int -> char -> unit = "%bytes_safe_set"
-  [@@ocaml.deprecated "Use Bytes.set instead."]
-(** [String.set s n c] modifies byte sequence [s] in place,
-   replacing the byte at index [n] with [c].
-   You can also write [s.[n] <- c] instead of [String.set s n c].
 
-   Raise [Invalid_argument] if [n] is not a valid index in [s].
-
-   @deprecated This is a deprecated alias of {!Bytes.set}.[ ] *)
-
-external create : int -> bytes = "caml_create_bytes"
-  [@@ocaml.deprecated "Use Bytes.create instead."]
-(** [String.create n] returns a fresh byte sequence of length [n].
-   The sequence is uninitialized and contains arbitrary bytes.
-
-   Raise [Invalid_argument] if [n < 0] or [n > ]{!Sys.max_string_length}.
-
-   @deprecated This is a deprecated alias of {!Bytes.create}.[ ] *)
 
 val make : int -> char -> string
 (** [String.make n c] returns a fresh string of length [n],
@@ -93,11 +76,6 @@ val init : int -> (int -> char) -> string
     @since 4.02.0
 *)
 
-val copy : string -> string [@@ocaml.deprecated]
-(** Return a copy of the given string.
-
-    @deprecated Because strings are immutable, it doesn't make much
-    sense to make identical copies of them. *)
 
 val sub : string -> int -> int -> string
 (** [String.sub s start len] returns a fresh string of length [len],
@@ -107,15 +85,7 @@ val sub : string -> int -> int -> string
    Raise [Invalid_argument] if [start] and [len] do not
    designate a valid substring of [s]. *)
 
-val fill : bytes -> int -> int -> char -> unit
-  [@@ocaml.deprecated "Use Bytes.fill instead."]
-(** [String.fill s start len c] modifies byte sequence [s] in place,
-   replacing [len] bytes with [c], starting at [start].
-
-   Raise [Invalid_argument] if [start] and [len] do not
-   designate a valid range of [s].
 
-   @deprecated This is a deprecated alias of {!Bytes.fill}.[ ] *)
 
 val blit : string -> int -> bytes -> int -> int -> unit
 (** Same as {!Bytes.blit_string}. *)
@@ -338,11 +308,6 @@ val split_on_char: char -> string -> string list
 (* The following is for system use only. Do not call directly. *)
 
 external unsafe_get : string -> int -> char = "%string_unsafe_get"
-external unsafe_set : bytes -> int -> char -> unit = "%bytes_unsafe_set"
-  [@@ocaml.deprecated]
 external unsafe_blit :
   string -> int -> bytes -> int -> int -> unit
   = "caml_blit_string" [@@noalloc]
-external unsafe_fill :
-  bytes -> int -> int -> char -> unit = "caml_fill_bytes" [@@noalloc]
-  [@@ocaml.deprecated]

jscomp/stdlib-406/string.ml

@@ -24,24 +24,6 @@ external get : string -> int -> char = "%string_safe_get"
 
    Raise [Invalid_argument] if [n] not a valid index in [s]. *)
 
-external set : bytes -> int -> char -> unit = "%bytes_safe_set"
-  [@@ocaml.deprecated "Use BytesLabels.set instead."]
-(** [String.set s n c] modifies byte sequence [s] in place,
-   replacing the byte at index [n] with [c].
-   You can also write [s.[n] <- c] instead of [String.set s n c].
-
-   Raise [Invalid_argument] if [n] is not a valid index in [s].
-
-   @deprecated This is a deprecated alias of {!BytesLabels.set}. *)
-
-external create : int -> bytes = "caml_create_bytes"
-  [@@ocaml.deprecated "Use BytesLabels.create instead."]
-(** [String.create n] returns a fresh byte sequence of length [n].
-   The sequence is uninitialized and contains arbitrary bytes.
-
-   Raise [Invalid_argument] if [n < 0] or [n > ]{!Sys.max_string_length}.
-
-   @deprecated This is a deprecated alias of {!BytesLabels.create}. *)
 
 val make : int -> char -> string
 (** [String.make n c] returns a fresh string of length [n],
@@ -56,8 +38,6 @@ val init : int -> f:(int -> char) -> string
    Raise [Invalid_argument] if [n < 0] or [n > ]{!Sys.max_string_length}.
    @since 4.02.0 *)
 
-val copy : string -> string  [@@ocaml.deprecated]
-(** Return a copy of the given string. *)
 
 val sub : string -> pos:int -> len:int -> string
 (** [String.sub s start len] returns a fresh string of length [len],
@@ -67,15 +47,6 @@ val sub : string -> pos:int -> len:int -> string
    Raise [Invalid_argument] if [start] and [len] do not
    designate a valid substring of [s]. *)
 
-val fill : bytes -> pos:int -> len:int -> char -> unit
-  [@@ocaml.deprecated "Use BytesLabels.fill instead."]
-(** [String.fill s start len c] modifies byte sequence [s] in place,
-   replacing [len] bytes by [c], starting at [start].
-
-   Raise [Invalid_argument] if [start] and [len] do not
-   designate a valid substring of [s].
-
-   @deprecated This is a deprecated alias of {!BytesLabels.fill}. *)
 
 val blit :
   src:string -> src_pos:int -> dst:bytes -> dst_pos:int -> len:int
@@ -293,11 +264,3 @@ val split_on_char: sep:char -> string -> string list
 (* The following is for system use only. Do not call directly. *)
 
 external unsafe_get : string -> int -> char = "%string_unsafe_get"
-external unsafe_set : bytes -> int -> char -> unit = "%bytes_unsafe_set"
-  [@@ocaml.deprecated]
-external unsafe_blit :
-  src:string -> src_pos:int -> dst:bytes -> dst_pos:int -> len:int ->
-    unit = "caml_blit_string" [@@noalloc]
-external unsafe_fill :
-  bytes -> pos:int -> len:int -> char -> unit = "caml_fill_bytes" [@@noalloc]
-  [@@ocaml.deprecated]

jscomp/stdlib-406/string.mli

@@ -409,6 +409,7 @@ o test/key_word_property.cmi test/key_word_property.cmj : cc test/key_word_prope
 o test/key_word_property2.cmi test/key_word_property2.cmj : cc test/key_word_property2.ml | test/export_keyword.cmj $stdlib
 o test/key_word_property_plus_test.cmi test/key_word_property_plus_test.cmj : cc test/key_word_property_plus_test.ml | test/global_mangles.cmj test/mt.cmj $stdlib
 o test/label_uncurry.cmi test/label_uncurry.cmj : cc test/label_uncurry.ml | $stdlib
+o test/large_integer_pat.cmi test/large_integer_pat.cmj : cc test/large_integer_pat.ml | $stdlib
 o test/large_record_duplication_test.cmi test/large_record_duplication_test.cmj : cc test/large_record_duplication_test.ml | test/mt.cmj $stdlib
 o test/largest_int_flow.cmi test/largest_int_flow.cmj : cc test/largest_int_flow.ml | $stdlib
 o test/lazy_demo.cmi test/lazy_demo.cmj : cc test/lazy_demo.re | $stdlib