add collections; improve identifier

grammar.js

@@ -1,12 +1,15 @@
 const
-  space = ' ',
-  decimal = /[0-9]/
+  //TODO figure out how to subtact regex sets
+  unicodeLetter = /\p{L}/
+  unicodePunctuation = /\p{Pc}/
+  unicodeDigit = /[0-9]/
 
 module.exports = grammar({
   name: 'hcl',
 
   conflicts: $ => [
       [$.body],
+      [$.object_elem, $.variable_expr],
   ],
 
   extras: $ => [
@@ -46,7 +49,7 @@ module.exports = grammar({
 
     expr_term: $ => choice(
       $.literal_value,
-      // $.collection_value,
+      $.collection_value,
       // $.template_expr,
       $.variable_expr,
       // $.function_call,
@@ -68,6 +71,35 @@ module.exports = grammar({
 
     numeric_lit: $ => /[0-9]+(\.[0-9]+([eE][-+]?[0-9]+)?)?/,
 
+    collection_value: $ => choice(
+      $.tuple,
+      $.object,
+    ),
+
+    tuple: $ => seq(
+      '[',
+      optional(seq(
+        $.expression,
+        repeat(seq(',', $.expression)),
+      )),
+      ']',
+    ),
+
+    object: $ => seq(
+      '{',
+      optional(seq(
+        $.object_elem,
+        repeat(seq(',', $.object_elem)),
+      )),
+      '}',
+    ),
+
+    object_elem: $ => seq(
+      choice($.identifier, $.expression),
+      choice('=', ':'),
+      $.expression,
+    ),
+
     variable_expr: $ => $.identifier,
 
     // TODO: template expressions
@@ -81,8 +113,11 @@ module.exports = grammar({
     // TODO: string_literals are special template literals
     string_lit: $ => seq('"', /\w+/, '"'),
 
-    // TODO: unicode identifier
-    identifier: $ => /\w+/,
+    // TODO: not to spec, but maybe good enough
+    identifier: $ => token(seq(
+      unicodeLetter,
+      repeat(choice(unicodeLetter, unicodeDigit, unicodePunctuation))
+    )),
 
     // http://stackoverflow.com/questions/13014947/regex-to-match-a-c-style-multiline-comment/36328890#36328890
     comment: $ => token(choice(
@@ -93,6 +128,6 @@ module.exports = grammar({
         /[^*]*\*+([^/*][^*]*\*+)*/,
         '/'
       )
-    ))
+    )),
   }
 });

src/grammar.json

@@ -104,6 +104,10 @@
           "type": "SYMBOL",
           "name": "literal_value"
         },
+        {
+          "type": "SYMBOL",
+          "name": "collection_value"
+        },
         {
           "type": "SYMBOL",
           "name": "variable_expr"
@@ -156,6 +160,146 @@
       "type": "PATTERN",
       "value": "[0-9]+(\\.[0-9]+([eE][-+]?[0-9]+)?)?"
     },
+    "collection_value": {
+      "type": "CHOICE",
+      "members": [
+        {
+          "type": "SYMBOL",
+          "name": "tuple"
+        },
+        {
+          "type": "SYMBOL",
+          "name": "object"
+        }
+      ]
+    },
+    "tuple": {
+      "type": "SEQ",
+      "members": [
+        {
+          "type": "STRING",
+          "value": "["
+        },
+        {
+          "type": "CHOICE",
+          "members": [
+            {
+              "type": "SEQ",
+              "members": [
+                {
+                  "type": "SYMBOL",
+                  "name": "expression"
+                },
+                {
+                  "type": "REPEAT",
+                  "content": {
+                    "type": "SEQ",
+                    "members": [
+                      {
+                        "type": "STRING",
+                        "value": ","
+                      },
+                      {
+                        "type": "SYMBOL",
+                        "name": "expression"
+                      }
+                    ]
+                  }
+                }
+              ]
+            },
+            {
+              "type": "BLANK"
+            }
+          ]
+        },
+        {
+          "type": "STRING",
+          "value": "]"
+        }
+      ]
+    },
+    "object": {
+      "type": "SEQ",
+      "members": [
+        {
+          "type": "STRING",
+          "value": "{"
+        },
+        {
+          "type": "CHOICE",
+          "members": [
+            {
+              "type": "SEQ",
+              "members": [
+                {
+                  "type": "SYMBOL",
+                  "name": "object_elem"
+                },
+                {
+                  "type": "REPEAT",
+                  "content": {
+                    "type": "SEQ",
+                    "members": [
+                      {
+                        "type": "STRING",
+                        "value": ","
+                      },
+                      {
+                        "type": "SYMBOL",
+                        "name": "object_elem"
+                      }
+                    ]
+                  }
+                }
+              ]
+            },
+            {
+              "type": "BLANK"
+            }
+          ]
+        },
+        {
+          "type": "STRING",
+          "value": "}"
+        }
+      ]
+    },
+    "object_elem": {
+      "type": "SEQ",
+      "members": [
+        {
+          "type": "CHOICE",
+          "members": [
+            {
+              "type": "SYMBOL",
+              "name": "identifier"
+            },
+            {
+              "type": "SYMBOL",
+              "name": "expression"
+            }
+          ]
+        },
+        {
+          "type": "CHOICE",
+          "members": [
+            {
+              "type": "STRING",
+              "value": "="
+            },
+            {
+              "type": "STRING",
+              "value": ":"
+            }
+          ]
+        },
+        {
+          "type": "SYMBOL",
+          "name": "expression"
+        }
+      ]
+    },
     "variable_expr": {
       "type": "SYMBOL",
       "name": "identifier"
@@ -178,8 +322,36 @@
       ]
     },
     "identifier": {
-      "type": "PATTERN",
-      "value": "\\w+"
+      "type": "TOKEN",
+      "content": {
+        "type": "SEQ",
+        "members": [
+          {
+            "type": "PATTERN",
+            "value": "\\p{L}"
+          },
+          {
+            "type": "REPEAT",
+            "content": {
+              "type": "CHOICE",
+              "members": [
+                {
+                  "type": "PATTERN",
+                  "value": "\\p{L}"
+                },
+                {
+                  "type": "PATTERN",
+                  "value": "[0-9]"
+                },
+                {
+                  "type": "PATTERN",
+                  "value": "\\p{Pc}"
+                }
+              ]
+            }
+          }
+        ]
+      }
     },
     "comment": {
       "type": "TOKEN",
@@ -246,6 +418,10 @@
   "conflicts": [
     [
       "body"
+    ],
+    [
+      "object_elem",
+      "variable_expr"
     ]
   ],
   "precedences": [],

src/node-types.json

@@ -60,6 +60,25 @@
       ]
     }
   },
+  {
+    "type": "collection_value",
+    "named": true,
+    "fields": {},
+    "children": {
+      "multiple": false,
+      "required": true,
+      "types": [
+        {
+          "type": "object",
+          "named": true
+        },
+        {
+          "type": "tuple",
+          "named": true
+        }
+      ]
+    }
+  },
   {
     "type": "config_file",
     "named": true,
@@ -83,6 +102,10 @@
       "multiple": false,
       "required": true,
       "types": [
+        {
+          "type": "collection_value",
+          "named": true
+        },
         {
           "type": "expression",
           "named": true
@@ -113,11 +136,6 @@
       ]
     }
   },
-  {
-    "type": "identifier",
-    "named": true,
-    "fields": {}
-  },
   {
     "type": "literal_value",
     "named": true,
@@ -137,11 +155,60 @@
       ]
     }
   },
+  {
+    "type": "object",
+    "named": true,
+    "fields": {},
+    "children": {
+      "multiple": true,
+      "required": false,
+      "types": [
+        {
+          "type": "object_elem",
+          "named": true
+        }
+      ]
+    }
+  },
+  {
+    "type": "object_elem",
+    "named": true,
+    "fields": {},
+    "children": {
+      "multiple": true,
+      "required": true,
+      "types": [
+        {
+          "type": "expression",
+          "named": true
+        },
+        {
+          "type": "identifier",
+          "named": true
+        }
+      ]
+    }
+  },
   {
     "type": "string_lit",
     "named": true,
     "fields": {}
   },
+  {
+    "type": "tuple",
+    "named": true,
+    "fields": {},
+    "children": {
+      "multiple": true,
+      "required": false,
+      "types": [
+        {
+          "type": "expression",
+          "named": true
+        }
+      ]
+    }
+  },
   {
     "type": "variable_expr",
     "named": true,
@@ -169,10 +236,26 @@
     "type": ")",
     "named": false
   },
+  {
+    "type": ",",
+    "named": false
+  },
+  {
+    "type": ":",
+    "named": false
+  },
   {
     "type": "=",
     "named": false
   },
+  {
+    "type": "[",
+    "named": false
+  },
+  {
+    "type": "]",
+    "named": false
+  },
   {
     "type": "comment",
     "named": true
@@ -181,6 +264,10 @@
     "type": "false",
     "named": false
   },
+  {
+    "type": "identifier",
+    "named": true
+  },
   {
     "type": "null",
     "named": false