Formatting and other nitpicks

.editorconfig

@@ -7,6 +7,10 @@ indent_size = 2
 indent_style = space
 insert_final_newline = true
 trim_trailing_whitespace = true
+max_line_length = 120
 
 [*.js]
 quote_type = single
+
+[*.{c,h,cc,rs}]
+indent_size = 4

.gitattributes

@@ -1,4 +1,4 @@
 * text=auto
 
-**/src/** linguist-generated
-**/bindings/node/binding.cc linguist-generated
+**/src/*.json linguist-generated
+**/src/parser.c linguist-generated

.github/workflows/ci.yml

@@ -18,7 +18,5 @@ jobs:
           node-version: 16
       - run: npm install
         name: Install dependencies
-      - run: >-
-          npm run test-dtd &&
-          npm run test-xml
+      - run: npm test
         name: Run tests

Cargo.toml

@@ -1,7 +1,7 @@
 [package]
 name = "tree-sitter-xml"
 description = "XML grammar for tree-sitter"
-version = "0.0.1"
+version = "0.2.0"
 license = "MIT"
 readme = "README.md"
 keywords = ["incremental", "parsing", "dtd", "xml"]
@@ -11,20 +11,20 @@ edition = "2021"
 
 build = "bindings/rust/build.rs"
 include = [
-  "bindings/rust",
-  "dtd/grammar.js",
-  "dtd/src",
-  "dtd/queries",
-  "xml/grammar.js",
-  "xml/src",
-  "xml/queries",
+  "bindings/rust/*",
+  "tree-sitter-dtd/grammar.js",
+  "tree-sitter-dtd/src/*",
+  "tree-sitter-dtd/queries/*",
+  "tree-sitter-xml/grammar.js",
+  "tree-sitter-xml/src/*",
+  "tree-sitter-xml/queries/*",
 ]
 
 [lib]
 path = "bindings/rust/lib.rs"
 
 [dependencies]
-tree-sitter = ">= 0.19, < 0.21"
+tree-sitter = "~0.20.10"
 
 [build-dependencies]
 cc = "1.0"

binding.gyp

@@ -4,14 +4,15 @@
       "target_name": "tree_sitter_xml_binding",
       "include_dirs": [
         "<!(node -e \"require('nan')\")",
-        "tree-sitter-xml/src"
+        "tree-sitter-dtd/src",
+        "tree-sitter-xml/src",
       ],
       "sources": [
         "tree-sitter-dtd/src/parser.c",
         "tree-sitter-dtd/src/scanner.c",
         "tree-sitter-xml/src/parser.c",
         "tree-sitter-xml/src/scanner.c",
-        "bindings/node/binding.cc"
+        "bindings/node/binding.cc",
       ],
       "cflags_c": [
         "-std=c99",

bindings/node/index.js

@@ -1,11 +1,11 @@
 try {
-		module.exports = require("../../build/Release/tree_sitter_xml_binding");
+  module.exports = require('../../build/Release/tree_sitter_xml_binding');
 } catch (error1) {
   if (error1.code !== 'MODULE_NOT_FOUND') {
     throw error1;
   }
   try {
-				module.exports = require("../../build/Debug/tree_sitter_xml_binding");
+    module.exports = require('../../build/Debug/tree_sitter_xml_binding');
   } catch (error2) {
     if (error2.code !== 'MODULE_NOT_FOUND') {
       throw error2;
@@ -15,10 +15,6 @@ try {
 }
 
 try {
-		module.exports.xml.nodeTypeInfo = require("../../tree-sitter-xml/src/node-types.json");
-		module.exports.dtd.nodeTypeInfo = require("../../tree-sitter-dtd/src/node-types.json");
+  module.exports.xml.nodeTypeInfo = require('../../tree-sitter-xml/src/node-types.json');
+  module.exports.dtd.nodeTypeInfo = require('../../tree-sitter-dtd/src/node-types.json');
 } catch (_) { }
-
-
-
-

bindings/node/xml.js

@@ -1,2 +1 @@
 module.exports = require('./index').xml;
-

bindings/rust/build.rs

@@ -2,6 +2,7 @@ fn main() {
     let root_dir = std::path::Path::new(".");
     let xml_dir = root_dir.join("tree-sitter-xml").join("src");
     let dtd_dir = root_dir.join("tree-sitter-dtd").join("src");
+    let common_dir = root_dir.join("common");
 
     let mut config = cc::Build::new();
     config.include(&xml_dir);
@@ -19,10 +20,7 @@ fn main() {
         println!("cargo:rerun-if-changed={}", path.to_str().unwrap());
     }
 
-    println!(
-        "cargo:rerun-if-changed={}",
-        root_dir.join("common").join("scanner.h").to_str().unwrap()
-    );
+    println!("cargo:rerun-if-changed={}", common_dir.join("scanner.h").to_str().unwrap());
 
     config.compile("parser-scanner");
 }

bindings/rust/lib.rs

@@ -50,18 +50,18 @@ pub fn language_xml() -> Language {
     unsafe { tree_sitter_xml() }
 }
 
-/// The syntax highlighting query for this language.
+/// The syntax highlighting queries for XML.
 pub const XML_HIGHLIGHT_QUERY: &str = include_str!("../../tree-sitter-xml/queries/highlights.scm");
-pub const DTD_HIGHLIGHT_QUERY: &str = include_str!("../../tree-sitter-dtd/queries/highlights.scm");
-
-/// The local-variable syntax highlighting query for this language.
-// pub const LOCALS_QUERY: &str = include_str!("../../tree-sitter-xml/queries/locals.scm");
 
-/// The symbol tagging query for this language.
-// pub const TAGGING_QUERY: &str = include_str!("../../tree-sitter-xml/queries/tags.scm");
+/// The syntax highlighting queries for DTD.
+pub const DTD_HIGHLIGHT_QUERY: &str = include_str!("../../tree-sitter-dtd/queries/highlights.scm");
 
-/// The content of the [`node-types.json`][] file for this grammar.
+/// The content of the [`node-types.json`][] file for XML.
 ///
 /// [`node-types.json`]: https://tree-sitter.github.io/tree-sitter/using-parsers#static-node-types
 pub const XML_NODE_TYPES: &str = include_str!("../../tree-sitter-xml/src/node-types.json");
+
+/// The content of the [`node-types.json`][] file for DTD.
+///
+/// [`node-types.json`]: https://tree-sitter.github.io/tree-sitter/using-parsers#static-node-types
 pub const DTD_NODE_TYPES: &str = include_str!("../../tree-sitter-dtd/src/node-types.json");

common/scanner.h

@@ -1,3 +1,5 @@
+#pragma once
+
 #include <ctype.h>
 #include <tree_sitter/parser.h>
 
@@ -7,20 +9,13 @@ enum TokenType {
     CharData,
 };
 
-static bool in_dtd_error_recovery(const bool *valid_symbols) {
-    return valid_symbols[PITarget] && valid_symbols[PIContent];
-}
-
-static bool in_xml_error_recovery(const bool *valid_symbols) {
-    return valid_symbols[PITarget] && valid_symbols[PIContent] &&
-           valid_symbols[CharData];
-}
-
-static void advance(TSLexer *lexer) { lexer->advance(lexer, false); }
+/// Advance the lexer to the next token
+static inline void advance(TSLexer *lexer) { lexer->advance(lexer, false); }
 
+/// Scan for the target of a PI node
 static bool scan_pi_target(TSLexer *lexer) {
-    bool advanced_once = false;
-    bool found_x_first = false;
+    bool advanced_once = false, found_x_first = false;
+
     if (isalpha(lexer->lookahead) || lexer->lookahead == '_') {
         if (lexer->lookahead == 'x' || lexer->lookahead == 'X') {
             found_x_first = true;
@@ -67,10 +62,13 @@ static bool scan_pi_target(TSLexer *lexer) {
     return false;
 }
 
+/// Scan for the content of a PI node
 static bool scan_pi_content(TSLexer *lexer) {
     bool advanced_once = false;
-    while (lexer->lookahead != '\n' && lexer->lookahead != '?' &&
-           !lexer->eof(lexer)) {
+
+    while (!lexer->eof(lexer) &&
+           lexer->lookahead != '\n' &&
+           lexer->lookahead != '?') {
         advanced_once = true;
         advance(lexer);
     }
@@ -93,10 +91,13 @@ static bool scan_pi_content(TSLexer *lexer) {
     return false;
 }
 
+/// Scan for a CharData node
 static bool scan_char_data(TSLexer *lexer) {
     bool advanced_once = false;
-    while (lexer->lookahead != '<' && lexer->lookahead != '&' &&
-           !lexer->eof(lexer)) {
+
+    while (!lexer->eof(lexer) &&
+           lexer->lookahead != '<' &&
+           lexer->lookahead != '&') {
         if (lexer->lookahead == ']') {
             lexer->mark_end(lexer);
             advance(lexer);
@@ -122,3 +123,24 @@ static bool scan_char_data(TSLexer *lexer) {
     }
     return false;
 }
+
+/// Scan for the common symbols
+#define SCAN_COMMON(lexer, valid_symbols) \
+    if (in_error_recovery(valid_symbols)) return false; \
+    \
+    if (valid_symbols[PITarget]) return scan_pi_target(lexer); \
+    \
+    if (valid_symbols[PIContent]) return scan_pi_content(lexer);
+
+/// Define the boilerplate functions of the scanner
+/// @param name the name of the language
+#define SCANNER_BOILERPLATE(name) \
+    void *tree_sitter_##name##_external_scanner_create() { return NULL; } \
+    \
+    void tree_sitter_##name##_external_scanner_destroy(void *payload) {} \
+    \
+    void tree_sitter_##name##_external_scanner_reset(void *payload) {} \
+    \
+    unsigned tree_sitter_##name##_external_scanner_serialize(void *payload, char *buffer) { return 0; } \
+    \
+    void tree_sitter_##name##_external_scanner_deserialize(void *payload, const char *buffer, unsigned length) {}

package.json

@@ -1,6 +1,6 @@
 {
   "name": "tree-sitter-xml",
-  "version": "0.1.0",
+  "version": "0.2.0",
   "license": "MIT",
   "description": "XML & DTD grammars for tree-sitter",
   "repository": "ObserverOfTime/tree-sitter-xml",
@@ -53,7 +53,7 @@
       "file-types": [
         "dtd"
       ],
-      "injection-regex": "dtd",
+      "injection-regex": "^dtd$",
       "highlights": "tree-sitter-dtd/queries/highlights.scm",
       "path": "tree-sitter-dtd"
     }

tree-sitter-dtd/grammar.js

@@ -28,6 +28,7 @@ module.exports = grammar({
     $._EntityDecl,
     $._Reference
   ],
+
   rules: {
     // AKA: extSubset
     document: $ => seq(

tree-sitter-dtd/src/scanner.c

@@ -1,37 +1,13 @@
-#include <ctype.h>
-#include <tree_sitter/parser.h>
-#include <wctype.h>
-
 #include "../../common/scanner.h"
 
-void *tree_sitter_dtd_external_scanner_create() { return NULL; }
-
-void tree_sitter_dtd_external_scanner_destroy(void *payload) {}
-
-void tree_sitter_dtd_external_scanner_reset(void *payload) {}
-
-unsigned tree_sitter_dtd_external_scanner_serialize(void *payload,
-                                                    char *buffer) {
-    return 0;
+static inline bool in_error_recovery(const bool *valid_symbols) {
+    return valid_symbols[PITarget] && valid_symbols[PIContent] && valid_symbols[CharData];
 }
 
-void tree_sitter_dtd_external_scanner_deserialize(void *payload,
-                                                  const char *buffer,
-                                                  unsigned length) {}
+bool tree_sitter_dtd_external_scanner_scan(void *payload, TSLexer *lexer, const bool *valid_symbols) {
+    SCAN_COMMON(lexer, valid_symbols)
 
-bool tree_sitter_dtd_external_scanner_scan(void *payload, TSLexer *lexer,
-                                           const bool *valid_symbols) {
-    if (in_dtd_error_recovery(valid_symbols)) {
     return false;
 }
 
-    if (valid_symbols[PITarget]) {
-        return scan_pi_target(lexer);
-    }
-
-    if (valid_symbols[PIContent]) {
-        return scan_pi_content(lexer);
-    }
-
-    return false;
-}
+SCANNER_BOILERPLATE(dtd)

tree-sitter-dtd/src/tree_sitter/parser.h

@@ -13,8 +13,9 @@ extern "C" {
 #define ts_builtin_sym_end 0
 #define TREE_SITTER_SERIALIZATION_BUFFER_SIZE 1024
 
-#ifndef TREE_SITTER_API_H_
 typedef uint16_t TSStateId;
+
+#ifndef TREE_SITTER_API_H_
 typedef uint16_t TSSymbol;
 typedef uint16_t TSFieldId;
 typedef struct TSLanguage TSLanguage;
@@ -129,16 +130,9 @@ struct TSLanguage {
  *  Lexer Macros
  */
 
-#ifdef _MSC_VER
-#define UNUSED __pragma(warning(suppress : 4101))
-#else
-#define UNUSED __attribute__((unused))
-#endif
-
 #define START_LEXER()           \
   bool result = false;          \
   bool skip = false;            \
-  UNUSED                        \
   bool eof = false;             \
   int32_t lookahead;            \
   goto start;                   \
@@ -172,7 +166,7 @@ struct TSLanguage {
  *  Parse Table Macros
  */
 
-#define SMALL_STATE(id) ((id) - LARGE_STATE_COUNT)
+#define SMALL_STATE(id) id - LARGE_STATE_COUNT
 
 #define STATE(id) id
 
@@ -182,7 +176,7 @@ struct TSLanguage {
   {{                                  \
     .shift = {                        \
       .type = TSParseActionTypeShift, \
-      .state = (state_value)          \
+      .state = state_value            \
     }                                 \
   }}
 
@@ -190,7 +184,7 @@ struct TSLanguage {
   {{                                  \
     .shift = {                        \
       .type = TSParseActionTypeShift, \
-      .state = (state_value),         \
+      .state = state_value,           \
       .repetition = true              \
     }                                 \
   }}

tree-sitter-xml/grammar.js

@@ -29,8 +29,8 @@ module.exports = grammar(DTD, {
 
   rules: {
     document: $ => prec(2, seq(
-      optional($._S),
-      optional($.prolog),
+      O($._S),
+      O($.prolog),
       field('root', $.element),
       repeat($._Misc),
     )),
@@ -138,7 +138,7 @@ module.exports = grammar(DTD, {
 
     CDStart: _ => seq('<![', 'CDATA', '['),
 
-    CData: _ => /[^\]]*|][^\][\]>]*|]][^>]*/,
+    CData: _ => /([^\]]|][^\][\]]|]][^>])*/,
 
     StyleSheetPI: $ => seq(
       '<?',

tree-sitter-xml/src/scanner.c

@@ -1,41 +1,16 @@
-#include <ctype.h>
-#include <tree_sitter/parser.h>
-#include <wctype.h>
-
 #include "../../common/scanner.h"
 
-void *tree_sitter_xml_external_scanner_create() { return NULL; }
-
-void tree_sitter_xml_external_scanner_destroy(void *payload) {}
-
-void tree_sitter_xml_external_scanner_reset(void *payload) {}
-
-unsigned tree_sitter_xml_external_scanner_serialize(void *payload,
-                                                    char *buffer) {
-    return 0;
+static inline bool in_error_recovery(const bool *valid_symbols) {
+    return valid_symbols[PITarget] && valid_symbols[PIContent];
 }
 
-void tree_sitter_xml_external_scanner_deserialize(void *payload,
-                                                  const char *buffer,
-                                                  unsigned length) {}
-
 bool tree_sitter_xml_external_scanner_scan(void *payload, TSLexer *lexer,
                                            const bool *valid_symbols) {
-    if (in_xml_error_recovery(valid_symbols)) {
-        return false;
-    }
+    SCAN_COMMON(lexer, valid_symbols)
 
-    if (valid_symbols[PITarget]) {
-        return scan_pi_target(lexer);
-    }
-
-    if (valid_symbols[PIContent]) {
-        return scan_pi_content(lexer);
-    }
-
-    if (valid_symbols[CharData]) {
-        return scan_char_data(lexer);
-    }
+    if (valid_symbols[CharData]) return scan_char_data(lexer);
 
     return false;
 }
+
+SCANNER_BOILERPLATE(xml)

tree-sitter-xml/src/tree_sitter/parser.h

@@ -13,8 +13,9 @@ extern "C" {
 #define ts_builtin_sym_end 0
 #define TREE_SITTER_SERIALIZATION_BUFFER_SIZE 1024
 
-#ifndef TREE_SITTER_API_H_
 typedef uint16_t TSStateId;
+
+#ifndef TREE_SITTER_API_H_
 typedef uint16_t TSSymbol;
 typedef uint16_t TSFieldId;
 typedef struct TSLanguage TSLanguage;
@@ -129,16 +130,9 @@ struct TSLanguage {
  *  Lexer Macros
  */
 
-#ifdef _MSC_VER
-#define UNUSED __pragma(warning(suppress : 4101))
-#else
-#define UNUSED __attribute__((unused))
-#endif
-
 #define START_LEXER()           \
   bool result = false;          \
   bool skip = false;            \
-  UNUSED                        \
   bool eof = false;             \
   int32_t lookahead;            \
   goto start;                   \
@@ -172,7 +166,7 @@ struct TSLanguage {
  *  Parse Table Macros
  */
 
-#define SMALL_STATE(id) ((id) - LARGE_STATE_COUNT)
+#define SMALL_STATE(id) id - LARGE_STATE_COUNT
 
 #define STATE(id) id
 
@@ -182,7 +176,7 @@ struct TSLanguage {
   {{                                  \
     .shift = {                        \
       .type = TSParseActionTypeShift, \
-      .state = (state_value)          \
+      .state = state_value            \
     }                                 \
   }}
 
@@ -190,7 +184,7 @@ struct TSLanguage {
   {{                                  \
     .shift = {                        \
       .type = TSParseActionTypeShift, \
-      .state = (state_value),         \
+      .state = state_value,           \
       .repetition = true              \
     }                                 \
   }}