Rename to ARM to not conflict with upstream ASM parser

- Also add include statement support
SethBarberee · Feb 16, 2024 · eee056f · eee056f
1 parent 0ac5741
commit eee056f
Show file tree

Hide file tree

Showing 9 changed files with 2,195 additions and 2,009 deletions.
diff --git a/README.md b/README.md
@@ -7,17 +7,17 @@ The start of a ARM/Thumb ASM tree-sitter parser. This isn't fully complete yet s
 1) Add the following snippet to your init.lua:
 ```
 local parser_config = require "nvim-treesitter.parsers".get_parser_configs()
-parser_config.asm = {
+parser_config.arm = {
   install_info = {
     url = <PATH to REPO>,
     files = {"src/parser.c"},
     -- optional entries:
     generate_requires_npm = false, -- if stand-alone parser without npm dependencies
     requires_generate_from_grammar = false, -- if folder contains pre-generated src/parser.c
   },
-  filetype = "asm", -- if filetype does not match the parser name
+  filetype = "arm", -- if filetype does not match the parser name
 }
 ```
-2) Run `TSInstallFromGrammar asm`
+2) Run `TSInstallFromGrammar arm`
 
-3) Copy `queries/highlights.scm` to `<path to nvim>/queries/after/asm/highlights.scm` (Optional, if you want the highlighting)
+3) Copy `queries/highlights.scm` to `<path to nvim>/queries/after/arm/highlights.scm` (Optional, if you want the highlighting)
diff --git a/grammar.js b/grammar.js
@@ -1,5 +1,5 @@
 module.exports = grammar({
-  name: "asm",
+  name: "arm", // have to use this since upstream has asm now
 
   word: ($) => $.identifier,
 
@@ -9,7 +9,8 @@ module.exports = grammar({
   rules: {
     source_file: ($) => optional($._statement),
 
-    _statement: ($) => repeat1(choice($.function_definition, $.directive)),
+    _statement: ($) =>
+      repeat1(choice($.function_definition, $.directive, $.include_statement)),
 
     function_definition: ($) =>
       seq(
@@ -23,10 +24,10 @@ module.exports = grammar({
             $.ldm_statement,
             $.push_statement,
             $.pool_statement,
-            $.label
-          )
+            $.label,
+          ),
         ),
-        $.return_statement
+        $.return_statement,
       ),
 
     math_statement: ($) =>
@@ -36,7 +37,7 @@ module.exports = grammar({
         ",",
         field("Rn", $.register),
         ",",
-        field("operand2", choice($.register, $.constant))
+        field("operand2", choice($.register, $.constant)),
       ),
 
     // We handle when [] or just a label and offsets
@@ -60,12 +61,12 @@ module.exports = grammar({
               choice(
                 field("offset", $.constant),
                 field("regoffset", $.register),
-                $.offset_statement
-              )
+                $.offset_statement,
+              ),
             ),
-            "]"
-          )
-        )
+            "]",
+          ),
+        ),
       ),
 
     ldm_statement: ($) =>
@@ -76,7 +77,7 @@ module.exports = grammar({
         ",",
         "{",
         field("registers", commaSep($.reg_list)),
-        "}"
+        "}",
       ),
 
     ldm_opcode: ($) => choice(/ldm([a-z]+)?/, /stm([a-z]+)?/),
@@ -102,7 +103,7 @@ module.exports = grammar({
         $.opcode,
         field("Rd", $.register),
         optional(","),
-        field("operand2", optional(choice($.register, $.constant)))
+        field("operand2", optional(choice($.register, $.constant))),
       ),
 
     // TODO: fairly limited and only to ARM/THUMB for now since that's all I use
@@ -133,7 +134,7 @@ module.exports = grammar({
         /umlal([a-z]+)?/,
         /smull([a-z]+)?/,
         /smlal([a-z]+)?/,
-        /nop/
+        /nop/,
       ),
 
     return_statement: ($) => seq($.branch_opcode, field("Rm", $.register)),
@@ -152,7 +153,7 @@ module.exports = grammar({
         /(bpl)\s+/,
         /(bx)\s+/,
         /(bl([a-z]+)?)\s+/,
-        /(bg([a-z]+)?)\s+/
+        /(bg([a-z]+)?)\s+/,
       ),
 
     //label: ($) => /(.*?):/,
@@ -163,12 +164,17 @@ module.exports = grammar({
     directive_statement: ($) => seq($.directive, $.constant),
     directive: ($) => token(/[.][0-9a-zA-Z]+/),
 
+    string: ($) => seq('"', $.identifier, '"'),
+
+    // TODO parse the file as a string
+    include_statement: ($) => seq("#include", $.string),
+
     comment: ($) =>
       token(
         choice(
           seq("@", /(\\(.|\r?\n)|[^\\\n])*/),
-          seq(";", /(\\(.|\r?\n)|[^\\\n])*/)
-        )
+          seq(";", /(\\(.|\r?\n)|[^\\\n])*/),
+        ),
       ),
 
     // Used in ldr/str and directives

diff --git a/queries/highlights.scm b/queries/highlights.scm
@@ -22,6 +22,8 @@
 (func_name (identifier)) @function
 (label (identifier)) @label
 (directive) @preproc
+(include_statement) @preproc
+(string) @string
 
 [
     (load_opcode)

diff --git a/src/grammar.json b/src/grammar.json
@@ -1,5 +1,5 @@
 {
-  "name": "asm",
+  "name": "arm",
   "word": "identifier",
   "rules": {
     "source_file": {
@@ -26,6 +26,10 @@
           {
             "type": "SYMBOL",
             "name": "directive"
+          },
+          {
+            "type": "SYMBOL",
+            "name": "include_statement"
           }
         ]
       }
@@ -762,6 +766,36 @@
         "value": "[.][0-9a-zA-Z]+"
       }
     },
+    "string": {
+      "type": "SEQ",
+      "members": [
+        {
+          "type": "STRING",
+          "value": "\""
+        },
+        {
+          "type": "SYMBOL",
+          "name": "identifier"
+        },
+        {
+          "type": "STRING",
+          "value": "\""
+        }
+      ]
+    },
+    "include_statement": {
+      "type": "SEQ",
+      "members": [
+        {
+          "type": "STRING",
+          "value": "#include"
+        },
+        {
+          "type": "SYMBOL",
+          "name": "string"
+        }
+      ]
+    },
     "comment": {
       "type": "TOKEN",
       "content": {

diff --git a/src/node-types.json b/src/node-types.json
@@ -108,6 +108,21 @@
       ]
     }
   },
+  {
+    "type": "include_statement",
+    "named": true,
+    "fields": {},
+    "children": {
+      "multiple": false,
+      "required": true,
+      "types": [
+        {
+          "type": "string",
+          "named": true
+        }
+      ]
+    }
+  },
   {
     "type": "label",
     "named": true,
@@ -460,6 +475,25 @@
         {
           "type": "function_definition",
           "named": true
+        },
+        {
+          "type": "include_statement",
+          "named": true
+        }
+      ]
+    }
+  },
+  {
+    "type": "string",
+    "named": true,
+    "fields": {},
+    "children": {
+      "multiple": false,
+      "required": true,
+      "types": [
+        {
+          "type": "identifier",
+          "named": true
         }
       ]
     }
@@ -468,6 +502,14 @@
     "type": "!",
     "named": false
   },
+  {
+    "type": "\"",
+    "named": false
+  },
+  {
+    "type": "#include",
+    "named": false
+  },
   {
     "type": ",",
     "named": false