OCaml: use unsafe_get instead of get as the default YYPEEK implem…

…entation. The generated code already handles the end of input, so it should use `unsafe_get` to avoid redundant bounds checks. If the user wants to use `get`, they can always do so by overriding YYPEEK. Update all examples and remove '\x00' character at the end of string literals, as they are already NULL-terminated.
skvadrik · Dec 25, 2024 · e2d249f · e2d249f
1 parent 54d0eaa
commit e2d249f
Show file tree

Hide file tree

Showing 33 changed files with 604 additions and 604 deletions.
diff --git a/bootstrap/src/default_syntax_ocaml.h b/bootstrap/src/default_syntax_ocaml.h
@@ -133,7 +133,7 @@ static constexpr const char* DEFAULT_SYNTAX_OCAML =
     "re2c:YYMAXNMATCH = \"yymaxnmatch\";\n"
     "re2c:YYMTAGN = \"yymatgn\";\n"
     "re2c:YYMTAGP = \"yymtagp\";\n"
-    "re2c:YYPEEK = (.api.record ? \"get\" : \"yypeek\");\n"
+    "re2c:YYPEEK = (.api.record ? \"unsafe_get\" : \"yypeek\");\n"
     "re2c:YYRESTORE = \"yyrestore\";\n"
     "re2c:YYRESTORECTX = \"yyrestorectx\";\n"
     "re2c:YYRESTORETAG = \"yyrestoretag\";\n"

diff --git a/examples/ocaml/01_basic.ml b/examples/ocaml/01_basic.ml
@@ -10,7 +10,7 @@ type state = {
 
 
 let rec yy0 (yyrecord : state) : bool =
-    let yych = get yyrecord.yyinput yyrecord.yycursor in
+    let yych = unsafe_get yyrecord.yyinput yyrecord.yycursor in
     yyrecord.yycursor <- yyrecord.yycursor + 1;
     match yych with
         | '1'..'9' -> (yy2 [@tailcall]) yyrecord
@@ -20,7 +20,7 @@ and yy1 (yyrecord : state) : bool =
     false
 
 and yy2 (yyrecord : state) : bool =
-    let yych = get yyrecord.yyinput yyrecord.yycursor in
+    let yych = unsafe_get yyrecord.yyinput yyrecord.yycursor in
     match yych with
         | '0'..'9' ->
             yyrecord.yycursor <- yyrecord.yycursor + 1;
@@ -36,7 +36,7 @@ and lex (yyrecord : state) : bool =
 
 
 let main () =
-    let st = {yyinput = "1234\x00"; yycursor = 0}
+    let st = {yyinput = "1234"; yycursor = 0}
     in if not (lex st) then raise (Failure "error")
 
 let _ = main ()
diff --git a/examples/ocaml/01_basic.re b/examples/ocaml/01_basic.re
@@ -16,7 +16,7 @@ type state = {
 %}
 
 let main () =
-    let st = {yyinput = "1234\x00"; yycursor = 0}
+    let st = {yyinput = "1234"; yycursor = 0}
     in if not (lex st) then raise (Failure "error")
 
 let _ = main ()
diff --git a/examples/ocaml/conditions/parse_u32_blocks.ml b/examples/ocaml/conditions/parse_u32_blocks.ml
@@ -22,7 +22,7 @@ let add (num: int option) (dgt: int) (base: int) : int option =
 
 
 let rec yy0 (st : state) (num : int option) : int option =
-    let yych = get st.yyinput st.yycursor in
+    let yych = unsafe_get st.yyinput st.yycursor in
     st.yycursor <- st.yycursor + 1;
     match yych with
         | '0'..'1' -> (yy2 [@tailcall]) st num
@@ -41,7 +41,7 @@ and parse_bin (st : state) (num : int option) : int option =
 
 
 let rec yy3 (st : state) (num : int option) : int option =
-    let yych = get st.yyinput st.yycursor in
+    let yych = unsafe_get st.yyinput st.yycursor in
     st.yycursor <- st.yycursor + 1;
     match yych with
         | '0'..'7' -> (yy5 [@tailcall]) st num
@@ -60,7 +60,7 @@ and parse_oct (st : state) (num : int option) : int option =
 
 
 let rec yy6 (st : state) (num : int option) : int option =
-    let yych = get st.yyinput st.yycursor in
+    let yych = unsafe_get st.yyinput st.yycursor in
     st.yycursor <- st.yycursor + 1;
     match yych with
         | '0'..'9' -> (yy8 [@tailcall]) st num
@@ -79,7 +79,7 @@ and parse_dec (st : state) (num : int option) : int option =
 
 
 let rec yy9 (st : state) (num : int option) : int option =
-    let yych = get st.yyinput st.yycursor in
+    let yych = unsafe_get st.yyinput st.yycursor in
     st.yycursor <- st.yycursor + 1;
     match yych with
         | '0'..'9' -> (yy11 [@tailcall]) st num
@@ -106,7 +106,7 @@ and parse_hex (st : state) (num : int option) : int option =
 
 
 let rec yy14 (st : state) : int option =
-    let yych = get st.yyinput st.yycursor in
+    let yych = unsafe_get st.yyinput st.yycursor in
     st.yycursor <- st.yycursor + 1;
     match yych with
         | '0' -> (yy16 [@tailcall]) st
@@ -118,7 +118,7 @@ and yy15 (st : state) : int option =
 
 and yy16 (st : state) : int option =
     st.yymarker <- st.yycursor;
-    let yych = get st.yyinput st.yycursor in
+    let yych = unsafe_get st.yyinput st.yycursor in
     match yych with
         | 'B'
         | 'b' ->
@@ -138,7 +138,7 @@ and yy18 (st : state) : int option =
     parse_dec st (Some 0)
 
 and yy19 (st : state) : int option =
-    let yych = get st.yyinput st.yycursor in
+    let yych = unsafe_get st.yyinput st.yycursor in
     match yych with
         | '0'..'1' ->
             st.yycursor <- st.yycursor + 1;
@@ -150,7 +150,7 @@ and yy20 (st : state) : int option =
     (yy17 [@tailcall]) st
 
 and yy21 (st : state) : int option =
-    let yych = get st.yyinput st.yycursor in
+    let yych = unsafe_get st.yyinput st.yycursor in
     match yych with
         | '0'..'9'
         | 'A'..'F'

diff --git a/examples/ocaml/conditions/parse_u32_conditions.ml b/examples/ocaml/conditions/parse_u32_conditions.ml
@@ -24,7 +24,7 @@ let add (num: int option) (dgt: int) (base: int) : int option =
 
 
 let rec yy1 (st : state) (num : int option) : int option =
-    let yych = get st.yyinput st.yycursor in
+    let yych = unsafe_get st.yyinput st.yycursor in
     st.yycursor <- st.yycursor + 1;
     match yych with
         | '0' -> (yy3 [@tailcall]) st num
@@ -36,7 +36,7 @@ and yy2 (st : state) (num : int option) : int option =
 
 and yy3 (st : state) (num : int option) : int option =
     st.yymarker <- st.yycursor;
-    let yych = get st.yyinput st.yycursor in
+    let yych = unsafe_get st.yyinput st.yycursor in
     match yych with
         | 'B'
         | 'b' ->
@@ -58,7 +58,7 @@ and yy5 (st : state) (num : int option) : int option =
     (yyfndec [@tailcall]) st num
 
 and yy6 (st : state) (num : int option) : int option =
-    let yych = get st.yyinput st.yycursor in
+    let yych = unsafe_get st.yyinput st.yycursor in
     match yych with
         | '0'..'1' ->
             st.yycursor <- st.yycursor + 1;
@@ -70,7 +70,7 @@ and yy7 (st : state) (num : int option) : int option =
     (yy4 [@tailcall]) st num
 
 and yy8 (st : state) (num : int option) : int option =
-    let yych = get st.yyinput st.yycursor in
+    let yych = unsafe_get st.yyinput st.yycursor in
     match yych with
         | '0'..'9'
         | 'A'..'F'
@@ -93,7 +93,7 @@ and yyfninit (st : state) (num : int option) : int option =
     (yy1 [@tailcall]) st num
 
 and yy11 (st : state) (num : int option) : int option =
-    let yych = get st.yyinput st.yycursor in
+    let yych = unsafe_get st.yyinput st.yycursor in
     st.yycursor <- st.yycursor + 1;
     match yych with
         | '0'..'1' -> (yy13 [@tailcall]) st num
@@ -109,7 +109,7 @@ and yyfnbin (st : state) (num : int option) : int option =
     (yy11 [@tailcall]) st num
 
 and yy14 (st : state) (num : int option) : int option =
-    let yych = get st.yyinput st.yycursor in
+    let yych = unsafe_get st.yyinput st.yycursor in
     st.yycursor <- st.yycursor + 1;
     match yych with
         | '0'..'7' -> (yy16 [@tailcall]) st num
@@ -125,7 +125,7 @@ and yyfnoct (st : state) (num : int option) : int option =
     (yy14 [@tailcall]) st num
 
 and yy17 (st : state) (num : int option) : int option =
-    let yych = get st.yyinput st.yycursor in
+    let yych = unsafe_get st.yyinput st.yycursor in
     st.yycursor <- st.yycursor + 1;
     match yych with
         | '0'..'9' -> (yy19 [@tailcall]) st num
@@ -141,7 +141,7 @@ and yyfndec (st : state) (num : int option) : int option =
     (yy17 [@tailcall]) st num
 
 and yy20 (st : state) (num : int option) : int option =
-    let yych = get st.yyinput st.yycursor in
+    let yych = unsafe_get st.yyinput st.yycursor in
     st.yycursor <- st.yycursor + 1;
     match yych with
         | '0'..'9' -> (yy22 [@tailcall]) st num