Precalculate all ascii values

SWAT-engineering · Jul 9, 2024 · aae40ab · aae40ab
1 parent d9a3f21
commit aae40ab
Show file tree

Hide file tree

Showing 2 changed files with 30 additions and 56 deletions.
diff --git a/rascal-textmate-core/src/main/rascal/lang/oniguruma/Conversion.rsc b/rascal-textmate-core/src/main/rascal/lang/oniguruma/Conversion.rsc
@@ -170,59 +170,14 @@ str encode(list[int] chars, bool withBounds = false)
     ? "\\b<encode(chars, withBounds = false)>\\b"
     : intercalate("", [encode(i) | i <- chars]);
 
-str encode(int char) {
-    if (char in alnum) {
-        return stringChar(char);
-    }
-    if (char in shorthands) {
-        return shorthands[char];
-    }
-    if (char < 256) {
-        return "\\x<right(toHex(char),2, "0")>";
-    }
-    return "\\x{<toHex(char)>}";
-
-}
+str encode(int char) 
+    = char in preEncoded
+    ? preEncoded[char]
+    : "\\x{<toHex(char)>}"
+    ;
 
-map[int, str] shorthands = (
-    0x09: "\\t",
-    0x0A: "\\n",
-    0x0D: "\\r",
-    0x20: " ",
-    0x21: "\\!",
-    0x22: "\"",
-    0x23: "\\#",
-    0x24: "\\$",
-    0x25: "\\%",
-    0x26: "\\&",
-    0x27: "\'",
-    0x28: "\\(",
-    0x29: "\\)",
-    0x2A: "\\*",
-    0x2B: "\\+",
-    0x2C: "\\,",
-    0x2D: "\\-",
-    0x2E: "\\.",
-    0x2F: "\\/",
-    0x3A: "\\:",
-    0x3B: "\\;",
-    0x3C: "\\\<",
-    0x3D: "\\=",
-    0x3E: "\\\>",
-    0x3F: "\\?",
-    0x40: "\\@",
-    0x5B: "\\[",
-    0x5C: "\\\\",
-    0x5D: "\\]",
-    0x5E: "\\^",
-    0x5F: "\\_",
-    0x60: "\\`",
-    0x7B: "\\{",
-    0x7C: "\\|",
-    0x7D: "\\{",
-    0x7E: "\\~"
-);
 
+private set[int] charRange(str from, str to) = {*[charAt(from, 0)..charAt(to, 0) + 1]};
 
 private str toHex(int i)
     = i < 16 
@@ -233,5 +188,24 @@ private list[str] hex
     = ["<i>" | i <- [0..10]]
     + ["A", "B", "C", "D", "E", "F"];
 
-private set[int] alnum
-    = {*[48..58], *[65..91], *[97..123]};
+private set[int] printable
+    = charRange("0", "9")
+    + charRange("a", "z")
+    + charRange("A", "Z")
+    ;
+
+private map[int, str] escapes = (
+    0x09: "\\t",
+    0x0A: "\\n",
+    0x0D: "\\r",
+    0x20: "\\x20" // spaces look a bit strange in a regex, although they are valid, people tend to read over them as layout
+) + ( c : "\\<stringChar(c)>" | c <- [0x21..0x7F], c notin printable); // regular ascii characters that might have special meaning in a regex
+
+
+private map[int, str] addFallBack(map[int, str] defined)
+    = ( char : "\\x<right(toHex(char),2, "0")>" | char <- [0..256], char notin defined)
+    + defined
+    ;
+
+private map[int, str] preEncoded
+    = addFallBack(escapes + ( c : stringChar(c) | c <- printable));
diff --git a/vscode-extension/syntaxes/rascal.tmLanguage.json b/vscode-extension/syntaxes/rascal.tmLanguage.json
@@ -1,7 +1,7 @@
 {
   "repository": {
     "prod(lex(\"Comment\"),[lit(\"//\"),conditional(\\iter-star(\\char-class([range(1,9),range(11,1114111)])),{\\not-follow(\\char-class([range(9,9),range(13,13),range(32,32),range(160,160),range(5760,5760),range(8192,8202),range(8239,8239),range(8287,8287),range(12288,12288)])),\\end-of-line()})],{tag(\"category\"(\"comment\"))})": {
-      "match": "((?:\\/\\/)(?:(?:[\\x01-\\t]|[\\x0B-\\x{10FFFF}])*?(?!(?:[\\t-\\t]|[\\r-\\r]|[ - ]|[\\xA0-\\xA0]|[\\x{1680}-\\x{1680}]|[\\x{2000}-\\x{200A}]|[\\x{202F}-\\x{202F}]|[\\x{205F}-\\x{205F}]|[\\x{3000}-\\x{3000}]))(?:$)))",
+      "match": "((?:\\/\\/)(?:(?:[\\x01-\\t]|[\\x0B-\\x{10FFFF}])*?(?!(?:[\\t-\\t]|[\\r-\\r]|[\\x20-\\x20]|[\\xA0-\\xA0]|[\\x{1680}-\\x{1680}]|[\\x{2000}-\\x{200A}]|[\\x{202F}-\\x{202F}]|[\\x{205F}-\\x{205F}]|[\\x{3000}-\\x{3000}]))(?:$)))",
       "name": "prod(lex(\"Comment\"),[lit(\"//\"),conditional(\\iter-star(\\char-class([range(1,9),range(11,1114111)])),{\\not-follow(\\char-class([range(9,9),range(13,13),range(32,32),range(160,160),range(5760,5760),range(8192,8202),range(8239,8239),range(8287,8287),range(12288,12288)])),\\end-of-line()})],{tag(\"category\"(\"comment\"))})",
       "captures": {
         "1": {
@@ -19,7 +19,7 @@
       }
     },
     "prod(lex(\"delimiters\"),[alt({lit(\"bottom-up-break\"),lit(\")\"),lit(\"≫\"),lit(\"\\n\"),lit(\"!:=\"),lit(\"\\'\"),lit(\"!=\"),lit(\"\\>=\"),lit(\"://\"),lit(\"non-assoc\"),lit(\"&=\"),lit(\"\\<-\"),lit(\"*=\"),lit(\"+=\"),lit(\"top-down-break\"),lit(\",\"),lit(\"...\"),lit(\"/=\"),lit(\"⇨\"),lit(\"(\"),lit(\"*/\"),lit(\"%\"),lit(\"!\\<\\<\"),lit(\"=\\>\"),lit(\"!\\>\\>\"),lit(\"||\"),lit(\"\\>\\>\"),lit(\"::\"),lit(\"⚠\"),lit(\"&&\"),lit(\"@\"),lit(\":=\"),lit(\"#\"),lit(\"\\<==\\>\"),lit(\"\\\"\"),lit(\"\\<\\<=\"),lit(\"}\"),lit(\"?=\"),lit(\"\\<:\"),lit(\"==\\>\"),lit(\"^\"),lit(\"/*\"),lit(\";\"),lit(\"{\"),lit(\"-=\"),lit(\"$T\")})],{})": {
-      "match": "(?:(?:bottom\\-up\\-break)|(?:\\))|(?:\\x{226B})|(?:\\n)|(?:\\!\\:\\=)|(?:')|(?:\\!\\=)|(?:\\>\\=)|(?:\\:\\/\\/)|(?:non\\-assoc)|(?:\\&\\=)|(?:\\<\\-)|(?:\\*\\=)|(?:\\+\\=)|(?:top\\-down\\-break)|(?:\\,)|(?:\\.\\.\\.)|(?:\\/\\=)|(?:\\x{21E8})|(?:\\()|(?:\\*\\/)|(?:\\%)|(?:\\!\\<\\<)|(?:\\=\\>)|(?:\\!\\>\\>)|(?:\\|\\|)|(?:\\>\\>)|(?:\\:\\:)|(?:\\x{26A0})|(?:\\&\\&)|(?:\\@)|(?:\\:\\=)|(?:\\#)|(?:\\<\\=\\=\\>)|(?:\")|(?:\\<\\<\\=)|(?:\\{)|(?:\\?\\=)|(?:\\<\\:)|(?:\\=\\=\\>)|(?:\\^)|(?:\\/\\*)|(?:\\;)|(?:\\{)|(?:\\-\\=)|(?:\\$T))",
+      "match": "(?:(?:bottom\\-up\\-break)|(?:\\))|(?:\\x{226B})|(?:\\n)|(?:\\!\\:\\=)|(?:\\')|(?:\\!\\=)|(?:\\>\\=)|(?:\\:\\/\\/)|(?:non\\-assoc)|(?:\\&\\=)|(?:\\<\\-)|(?:\\*\\=)|(?:\\+\\=)|(?:top\\-down\\-break)|(?:\\,)|(?:\\.\\.\\.)|(?:\\/\\=)|(?:\\x{21E8})|(?:\\()|(?:\\*\\/)|(?:\\%)|(?:\\!\\<\\<)|(?:\\=\\>)|(?:\\!\\>\\>)|(?:\\|\\|)|(?:\\>\\>)|(?:\\:\\:)|(?:\\x{26A0})|(?:\\&\\&)|(?:\\@)|(?:\\:\\=)|(?:\\#)|(?:\\<\\=\\=\\>)|(?:\\\")|(?:\\<\\<\\=)|(?:\\})|(?:\\?\\=)|(?:\\<\\:)|(?:\\=\\=\\>)|(?:\\^)|(?:\\/\\*)|(?:\\;)|(?:\\{)|(?:\\-\\=)|(?:\\$T))",
       "name": "prod(lex(\"delimiters\"),[alt({lit(\"bottom-up-break\"),lit(\")\"),lit(\"≫\"),lit(\"\\n\"),lit(\"!:=\"),lit(\"\\'\"),lit(\"!=\"),lit(\"\\>=\"),lit(\"://\"),lit(\"non-assoc\"),lit(\"&=\"),lit(\"\\<-\"),lit(\"*=\"),lit(\"+=\"),lit(\"top-down-break\"),lit(\",\"),lit(\"...\"),lit(\"/=\"),lit(\"⇨\"),lit(\"(\"),lit(\"*/\"),lit(\"%\"),lit(\"!\\<\\<\"),lit(\"=\\>\"),lit(\"!\\>\\>\"),lit(\"||\"),lit(\"\\>\\>\"),lit(\"::\"),lit(\"⚠\"),lit(\"&&\"),lit(\"@\"),lit(\":=\"),lit(\"#\"),lit(\"\\<==\\>\"),lit(\"\\\"\"),lit(\"\\<\\<=\"),lit(\"}\"),lit(\"?=\"),lit(\"\\<:\"),lit(\"==\\>\"),lit(\"^\"),lit(\"/*\"),lit(\";\"),lit(\"{\"),lit(\"-=\"),lit(\"$T\")})],{})",
       "captures": {}
     },
@@ -80,7 +80,7 @@
       "end": "(?:\\])",
       "patterns": [
         {
-          "match": "((?:\\\\)(?:[ - ]|[\"-\"]|['-']|[\\--\\-]|[\\<-\\<]|[\\>-\\>]|[\\[-\\]]|[b-b]|[f-f]|[n-n]|[r-r]|[t-t]))",
+          "match": "((?:\\\\)(?:[\\x20-\\x20]|[\\\"-\\\"]|[\\'-\\']|[\\--\\-]|[\\<-\\<]|[\\>-\\>]|[\\[-\\]]|[b-b]|[f-f]|[n-n]|[r-r]|[t-t]))",
           "name": "prod(lex(\"Char\"),[lit(\"\\\\\"),\\char-class([range(32,32),range(34,34),range(39,39),range(45,45),range(60,60),range(62,62),range(91,93),range(98,98),range(102,102),range(110,110),range(114,114),range(116,116)])],{tag(\"category\"(\"constant\"))})",
           "captures": {
             "1": {