Add bitmaps support (OCaml backend).

skvadrik · Sep 20, 2024 · eac735f · eac735f
1 parent 5a4afb1
commit eac735f
Show file tree

Hide file tree

Showing 8 changed files with 140 additions and 23 deletions.
diff --git a/bootstrap/src/default_syntax_ocaml.cc b/bootstrap/src/default_syntax_ocaml.cc
@@ -6,7 +6,8 @@ const char* DEFAULT_SYNTAX_OCAML =
     "supported_api_styles = [\"free-form\"];\n"
     "supported_code_models = [\"recursive-functions\"];\n"
     "supported_targets = [\"code\", \"dot\"];\n"
-    "supported_features = [\"nested-ifs\", \"case-ranges\", \"tags\", \"captures\", \"captvars\"];\n"
+    "supported_features = [\"nested-ifs\", \"bitmaps\", \"case-ranges\",\n"
+    "    \"tags\", \"captures\", \"captvars\"];\n"
     "\n"
     "\n"
     "// language-specific options ---------------------------------------------------\n"
@@ -159,14 +160,19 @@ const char* DEFAULT_SYNTAX_OCAML =
     "code:const_local = topindent \"let \" name \" = \" init \" in\" nl;\n"
     "code:const_global = topindent \"let \" name \" = \" init nl;\n"
     "\n"
-    "code:array_local = <undefined>;\n"
-    "code:array_global = <undefined>;\n"
+    "code:array_local =\n"
+    "    topindent \"and \" name \" = [|\" nl indent\n"
+    "        [row: topindent [elem{0:-2}: elem \"; \"] [elem{-1}: elem \";\"] nl]\n"
+    "    dedent topindent \"|]\" nl;\n"
+    "\n"
+    "code:array_global = code:array_local;\n"
+    "\n"
     "code:array_elem = array \".(\" index \")\";\n"
     "\n"
     "code:type_int = \"int\";\n"
     "code:type_uint = \"uint\";\n"
     "code:type_cond_enum = (storable_state? \"int\" : \"uint\");\n"
-    "code:type_yybm = <undefined>;\n"
+    "code:type_yybm = \"int\";\n"
     "code:type_yytarget = <undefined>;\n"
     "\n"
     "code:assign = topindent lhs \" <- \" rhs \";\" nl;\n"
@@ -384,7 +390,7 @@ const char* DEFAULT_SYNTAX_OCAML =
     "            : YYLIMIT \" <= \" YYCURSOR)\n"
     "        : YYLESSTHAN);\n"
     "\n"
-    "code:yybm_filter = <undefined>;\n"
+    "code:yybm_filter = yych \" land ~0xFF\";\n"
     "\n"
-    "code:yybm_match = <undefined>;\n"
+    "code:yybm_match = \"(\" yybm \".(\" offset \" + Char.code \" yych \") land \" mask \") != 0\";\n"
     ;
diff --git a/include/syntax/ocaml b/include/syntax/ocaml
@@ -4,7 +4,8 @@ supported_apis = ["record", "generic"];
 supported_api_styles = ["free-form"];
 supported_code_models = ["recursive-functions"];
 supported_targets = ["code", "dot"];
-supported_features = ["nested-ifs", "case-ranges", "tags", "captures", "captvars"];
+supported_features = ["nested-ifs", "bitmaps", "case-ranges",
+    "tags", "captures", "captvars"];
 
 
 // language-specific options ---------------------------------------------------
@@ -157,14 +158,19 @@ code:var_global = topindent "let " name " = " init nl;
 code:const_local = topindent "let " name " = " init " in" nl;
 code:const_global = topindent "let " name " = " init nl;
 
-code:array_local = <undefined>;
-code:array_global = <undefined>;
+code:array_local =
+    topindent "and " name " = [|" nl indent
+        [row: topindent [elem{0:-2}: elem "; "] [elem{-1}: elem ";"] nl]
+    dedent topindent "|]" nl;
+
+code:array_global = code:array_local;
+
 code:array_elem = array ".(" index ")";
 
 code:type_int = "int";
 code:type_uint = "uint";
 code:type_cond_enum = (storable_state? "int" : "uint");
-code:type_yybm = <undefined>;
+code:type_yybm = "int";
 code:type_yytarget = <undefined>;
 
 code:assign = topindent lhs " <- " rhs ";" nl;
@@ -382,6 +388,6 @@ code:yylessthan =
             : YYLIMIT " <= " YYCURSOR)
         : YYLESSTHAN);
 
-code:yybm_filter = <undefined>;
+code:yybm_filter = yych " land ~0xFF";
 
-code:yybm_match = <undefined>;
+code:yybm_match = "(" yybm ".(" offset " + Char.code " yych ") land " mask ") != 0";
diff --git a/src/codegen/pass2_generate.cc b/src/codegen/pass2_generate.cc
@@ -1724,18 +1724,20 @@ CodeList* gen_bitmap(Output& output, const CodeBitmap* bitmap, const std::string
     return stmts;
 }
 
-static void gen_bitmaps(Output& output, CodeList* code, bool* local_decls) {
+static bool gen_bitmaps(Output& output, CodeList* code) {
     OutputBlock& b = output.block();
 
-    if (!b.opts->bitmaps) return;
+    if (!b.opts->bitmaps) return false;
 
+    bool have_bitmaps = false;
     for (const std::unique_ptr<Adfa>& dfa : b.dfas) {
         CodeList* bitmap = gen_bitmap(output, dfa->bitmap, dfa->cond);
         if (bitmap) {
-            *local_decls = true;
+            have_bitmaps = true;
             append(code, bitmap);
         }
     }
+    return have_bitmaps;
 }
 
 void gen_dfa_as_blocks_with_labels(Output& output, const Adfa& dfa, CodeList* stmts) {
@@ -1930,7 +1932,7 @@ LOCAL_NODISCARD(Ret gen_block_code(Output& output, const Adfas& dfas, CodeList*
             append(code, gen_cond_table(output));
         }
 
-        gen_bitmaps(output, code, &local_decls);
+        local_decls |= gen_bitmaps(output, code);
 
         if (opts->storable_state) {
             CHECK_RET(gen_state_goto_implicit(output, code));
@@ -1968,7 +1970,7 @@ LOCAL_NODISCARD(Ret gen_block_code(Output& output, const Adfas& dfas, CodeList*
         local_decls = true;
         append(code, gen_yystate_def(output));
 
-        gen_bitmaps(output, code, &local_decls);
+        local_decls |= gen_bitmaps(output, code);
 
         CodeCases* cases = code_cases(alc);
         for (const std::unique_ptr<Adfa>& dfa : dfas) {
@@ -1989,6 +1991,8 @@ LOCAL_NODISCARD(Ret gen_block_code(Output& output, const Adfas& dfas, CodeList*
         CHECK_RET(gen_start_function(output, *dfas[0], funcs));
 
         append(code, code_recursive_functions(alc, funcs));
+
+        gen_bitmaps(output, code); // no local declarations in rec/func mode => ignore return value
     }
 
     // If needed, wrap the block in braces, so that variable declarations have local scope.

diff --git a/src/options/opt.cc b/src/options/opt.cc
@@ -332,10 +332,6 @@ LOCAL_NODISCARD(Ret fix_mutopt(
             RET_FAIL(error("computed gotos are not supported in this code model"));
         }
     }
-    if (real.bitmaps && glob.code_model == CodeModel::REC_FUNC) {
-        // TODO: allow bitmaps before in rec/func mode
-        RET_FAIL(error("bitmaps are not supported in this code model"));
-    }
 
     return Ret::OK;
 }

diff --git a/test/codegen/ocaml/01_basic_b.ml b/test/codegen/ocaml/01_basic_b.ml
@@ -0,0 +1,82 @@
+(* Generated by re2ocaml *)
+#1 "codegen/ocaml/01_basic_b.re"
+(* re2ocaml $INPUT -o $OUTPUT -b *)
+
+type state = {
+    str: string;
+    mutable cur: int;
+}
+
+
+#12 "codegen/ocaml/01_basic_b.ml"
+let rec yy0 (st : state) : bool =
+	let yych = st.str.[st.cur] in
+	st.cur <- st.cur + 1;
+	if (yych <= '0') then (yy1 [@tailcall]) st
+	else if (yych <= '9') then (yy2 [@tailcall]) st
+	else (yy1 [@tailcall]) st
+
+and yy1 (st : state) : bool =
+#18 "codegen/ocaml/01_basic_b.re"
+	false
+#23 "codegen/ocaml/01_basic_b.ml"
+
+and yy2 (st : state) : bool =
+	let yych = st.str.[st.cur] in
+	if ((yybm.(0 + Char.code yych) land 128) != 0) then (
+		st.cur <- st.cur + 1;
+		(yy2 [@tailcall]) st
+	) else (
+		(yy3 [@tailcall]) st
+	)
+
+and yy3 (st : state) : bool =
+#17 "codegen/ocaml/01_basic_b.re"
+	true
+#37 "codegen/ocaml/01_basic_b.ml"
+
+and lex (st : state) : bool =
+	(yy0 [@tailcall]) st
+
+and yybm = [|
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	128; 128; 128; 128; 128; 128; 128; 128;
+	128; 128;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+	  0;   0;   0;   0;   0;   0;   0;   0;
+|]
+#19 "codegen/ocaml/01_basic_b.re"
+
+
+let main () =
+    let st = {str = "1234\x00"; cur = 0}
+    in if not (lex st) then raise (Failure "error")
+
+let _ = main ()
diff --git a/test/codegen/ocaml/01_basic_b.re b/test/codegen/ocaml/01_basic_b.re
@@ -0,0 +1,25 @@
+(* re2ocaml $INPUT -o $OUTPUT -b *)
+
+type state = {
+    str: string;
+    mutable cur: int;
+}
+
+/*!re2c
+    re2c:api = custom;
+    re2c:define:YYFN = ["lex;bool", "st;state"];
+    re2c:define:YYPEEK = "st.str.[st.cur]";
+    re2c:define:YYSKIP = "st.cur <- st.cur + 1;";
+    re2c:yyfill:enable = 0;
+
+    number = [1-9][0-9]*;
+
+    number { true }
+    *      { false }
+*/
+
+let main () =
+    let st = {str = "1234\x00"; cur = 0}
+    in if not (lex st) then raise (Failure "error")
+
+let _ = main ()
diff --git a/test/codegen/ocaml/unsupported_bitmaps.ml b/test/codegen/ocaml/unsupported_bitmaps.ml
diff --git a/test/codegen/ocaml/unsupported_bitmaps.re b/test/codegen/ocaml/unsupported_bitmaps.re