Parse escape sequences

corpus/expressions.txt

@@ -81,7 +81,7 @@ int main() {
     (compound_statement
       (expression_statement (call_expression
         (identifier)
-        (argument_list (string_literal) (identifier)))))))
+        (argument_list (string_literal (escape_sequence)) (identifier)))))))
 
 ============================================
 String literals
@@ -102,7 +102,7 @@ int main() {
     (compound_statement
       (expression_statement (string_literal))
       (expression_statement (concatenated_string (string_literal) (string_literal) (string_literal)))
-      (expression_statement (string_literal)))))
+      (expression_statement (string_literal (escape_sequence) (escape_sequence))))))
 
 ============================================
 Character literals
@@ -123,9 +123,9 @@ int main() {
     (function_declarator (identifier) (parameter_list))
     (compound_statement
       (expression_statement (char_literal))
-      (expression_statement (char_literal))
-      (expression_statement (char_literal))
-      (expression_statement (char_literal)))))
+      (expression_statement (char_literal (escape_sequence)))
+      (expression_statement (char_literal (escape_sequence)))
+      (expression_statement (char_literal (escape_sequence))))))
 
 ============================================
 Field access

grammar.js

@@ -673,26 +673,40 @@ module.exports = grammar({
       repeat(choice('u', 'l', 'U', 'L'))
     )),
 
-    char_literal: $ => token(seq(
+    char_literal: $ => seq(
       '\'',
       choice(
-        seq(optional('\\'), /[^']/),
-        seq('\\', '\'')
+        $.escape_sequence,
+        /[^\n']/
       ),
       '\''
-    )),
+    ),
 
     concatenated_string: $ => seq(
       $.string_literal,
       repeat1($.string_literal)
     ),
 
-    string_literal: $ => token(seq(
+    string_literal: $ => seq(
       '"',
-      repeat(choice(/[^\\"\n]/, /\\./)),
-      '"')
+      repeat(choice(
+        token(prec(1, /[^\\"\n]/)),
+        $.escape_sequence
+      )),
+      '"'
     ),
 
+    escape_sequence: $ => token(seq(
+      '\\',
+      choice(
+        /[^xuU]/,
+        /\d{2,3}/,
+        /x[0-9a-fA-F]{2,}/,
+        /u[0-9a-fA-F]{4}/,
+        /U[0-9a-fA-F]{8}/
+      )
+    )),
+
     system_lib_string: $ => token(seq(
       '<',
       repeat(choice(/[^>\n]/, '\\>')),

src/grammar.json

@@ -4060,59 +4060,30 @@
       }
     },
     "char_literal": {
-      "type": "TOKEN",
-      "content": {
-        "type": "SEQ",
-        "members": [
-          {
-            "type": "STRING",
-            "value": "'"
-          },
-          {
-            "type": "CHOICE",
-            "members": [
-              {
-                "type": "SEQ",
-                "members": [
-                  {
-                    "type": "CHOICE",
-                    "members": [
-                      {
-                        "type": "STRING",
-                        "value": "\\"
-                      },
-                      {
-                        "type": "BLANK"
-                      }
-                    ]
-                  },
-                  {
-                    "type": "PATTERN",
-                    "value": "[^']"
-                  }
-                ]
-              },
-              {
-                "type": "SEQ",
-                "members": [
-                  {
-                    "type": "STRING",
-                    "value": "\\"
-                  },
-                  {
-                    "type": "STRING",
-                    "value": "'"
-                  }
-                ]
-              }
-            ]
-          },
-          {
-            "type": "STRING",
-            "value": "'"
-          }
-        ]
-      }
+      "type": "SEQ",
+      "members": [
+        {
+          "type": "STRING",
+          "value": "'"
+        },
+        {
+          "type": "CHOICE",
+          "members": [
+            {
+              "type": "SYMBOL",
+              "name": "escape_sequence"
+            },
+            {
+              "type": "PATTERN",
+              "value": "[^\\n']"
+            }
+          ]
+        },
+        {
+          "type": "STRING",
+          "value": "'"
+        }
+      ]
     },
     "concatenated_string": {
       "type": "SEQ",
@@ -4131,33 +4102,74 @@
       ]
     },
     "string_literal": {
+      "type": "SEQ",
+      "members": [
+        {
+          "type": "STRING",
+          "value": "\""
+        },
+        {
+          "type": "REPEAT",
+          "content": {
+            "type": "CHOICE",
+            "members": [
+              {
+                "type": "TOKEN",
+                "content": {
+                  "type": "PREC",
+                  "value": 1,
+                  "content": {
+                    "type": "PATTERN",
+                    "value": "[^\\\\\"\\n]"
+                  }
+                }
+              },
+              {
+                "type": "SYMBOL",
+                "name": "escape_sequence"
+              }
+            ]
+          }
+        },
+        {
+          "type": "STRING",
+          "value": "\""
+        }
+      ]
+    },
+    "escape_sequence": {
       "type": "TOKEN",
       "content": {
         "type": "SEQ",
         "members": [
           {
             "type": "STRING",
-            "value": "\""
-          },
-          {
-            "type": "REPEAT",
-            "content": {
-              "type": "CHOICE",
-              "members": [
-                {
-                  "type": "PATTERN",
-                  "value": "[^\\\\\"\\n]"
-                },
-                {
-                  "type": "PATTERN",
-                  "value": "\\\\."
-                }
-              ]
-            }
+            "value": "\\"
           },
           {
-            "type": "STRING",
-            "value": "\""
+            "type": "CHOICE",
+            "members": [
+              {
+                "type": "PATTERN",
+                "value": "[^xuU]"
+              },
+              {
+                "type": "PATTERN",
+                "value": "\\d{2,3}"
+              },
+              {
+                "type": "PATTERN",
+                "value": "x[0-9a-fA-F]{2,}"
+              },
+              {
+                "type": "PATTERN",
+                "value": "u[0-9a-fA-F]{4}"
+              },
+              {
+                "type": "PATTERN",
+                "value": "U[0-9a-fA-F]{8}"
+              }
+            ]
           }
         ]
       }