Merge pull request #16481 from hvitved/treesitter/bump2

Tree-sitter: Bump to 0.22.6

Author: hvitved

python/extractor/tsg-python/Cargo.toml

@@ -7,7 +7,7 @@ authors = ["Taus Brock-Nannestad <tausbn@github.com>"]
 edition = "2018"
 
 # When changing/updating these, the `Cargo.Bazel.lock` file has to be regenerated.
-# Run `CARGO_BAZEL_REPIN=true CARGO_BAZEL_REPIN_ONLY=py_deps ./build --bazel sync --only=py_deps`
+# Run `CARGO_BAZEL_REPIN=true CARGO_BAZEL_REPIN_ONLY=py_deps ./tools/bazel sync --only=py_deps`
 # in the `semmle-code` repository to do so.
 # For more information, check out the documentation at
 # https://bazelbuild.github.io/rules_rust/crate_universe.html#repinning--updating-dependencies

ql/Cargo.lock

@@ -20,7 +20,10 @@ include = [
 path = "bindings/rust/lib.rs"
 
 [dependencies]
-tree-sitter = "~0.20.3"
+tree-sitter = ">= 0.22.6"
 
 [build-dependencies]
 cc = "1.0"
+
+[patch.crates-io]
+tree-sitter = {git = "https://github.com/redsun82/tree-sitter.git", rev = "1f5c1112ceaa8fc6aff61d1852690407670d2a96"}

ql/buramu/tree-sitter-blame/Cargo.toml

@@ -2,19 +2,22 @@
 name = "codeql-extractor-ql"
 version = "0.1.0"
 authors = ["GitHub"]
-edition = "2018"
+edition = "2021"
 
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [dependencies]
-tree-sitter = ">= 0.20, < 0.21"
-tree-sitter-ql = { git = "https://github.com/tree-sitter/tree-sitter-ql.git", rev = "d08db734f8dc52f6bc04db53a966603122bc6985"}
-tree-sitter-ql-dbscheme = { git = "https://github.com/erik-krogh/tree-sitter-ql-dbscheme.git", rev = "63e1344353f63931e88bfbc2faa2e78e1421b213"}
+tree-sitter = ">= 0.22.6"
+tree-sitter-ql = { git = "https://github.com/tree-sitter/tree-sitter-ql.git", rev = "fa5c3821dd2161f5c8528a8cbdb258daa6dc4de6"}
+tree-sitter-ql-dbscheme = { git = "https://github.com/tree-sitter/tree-sitter-ql-dbscheme.git", rev = "5f770f57fa415607ff50e3d237d47c8f11440eb3"}
 tree-sitter-blame = {path = "../buramu/tree-sitter-blame"}
-tree-sitter-json = {git = "https://github.com/tausbn/tree-sitter-json.git", rev = "745663ee997f1576fe1e7187e6347e0db36ec7a9"}
+tree-sitter-json = {git = "https://github.com/tree-sitter/tree-sitter-json.git", rev = "94f5c527b2965465956c2000ed6134dd24daf2a7"}
 clap = { version = "4.2", features = ["derive"] }
 tracing = "0.1"
 tracing-subscriber = { version = "0.3.18", features = ["env-filter"] }
 rayon = "1.9.0"
 regex = "1.10.4"
 codeql-extractor = { path = "../../shared/tree-sitter-extractor" }
+
+[patch.crates-io]
+tree-sitter = {git = "https://github.com/redsun82/tree-sitter.git", rev = "1f5c1112ceaa8fc6aff61d1852690407670d2a96"}

ql/extractor/Cargo.toml

@@ -31,15 +31,25 @@ private Predicate getPredicateFromPosition(string s) {
   )
 }
 
+pragma[nomagic]
+private string getJsonStringComponent(JSON::String s, int i) {
+  result = s.getChild(i).(JSON::Token).getValue()
+}
+
+pragma[nomagic]
+private string getJsonString(JSON::String s) {
+  result = concat(string c, int i | c = getJsonStringComponent(s, i) | c order by i)
+}
+
 class Object extends JSON::Object {
-  JSON::Value getValue(string key) {
+  JSON::UnderscoreValue getValue(string key) {
     exists(JSON::Pair p | p = this.getChild(_) |
-      key = p.getKey().(JSON::String).getChild().getValue() and
+      key = getJsonString(p.getKey()) and
       result = p.getValue()
     )
   }
 
-  string getString(string key) { result = this.getValue(key).(JSON::String).getChild().getValue() }
+  string getString(string key) { result = getJsonString(this.getValue(key)) }
 
   int getNumber(string key) { result = this.getValue(key).(JSON::Number).getValue().toInt() }
 
@@ -61,7 +71,7 @@ class Object extends JSON::Object {
 class Array extends JSON::Array {
   Object getObject(int i) { result = this.getChild(i) }
 
-  string getString(int i) { result = this.getChild(i).(JSON::String).getChild().getValue() }
+  string getString(int i) { result = getJsonString(this.getChild(i)) }
 
   int getNumber(int i) { result = this.getChild(i).(JSON::Number).getValue().toInt() }
 

ql/ql/src/codeql_ql/StructuredLogs.qll

@@ -1767,13 +1767,15 @@ module JSON {
     final override string getAPrimaryQlClass() { result = "ReservedWord" }
   }
 
+  class UnderscoreValue extends @json_underscore_value, AstNode { }
+
   /** A class representing `array` nodes. */
   class Array extends @json_array, AstNode {
     /** Gets the name of the primary QL class for this element. */
     final override string getAPrimaryQlClass() { result = "Array" }
 
     /** Gets the `i`th child of this node. */
-    final Value getChild(int i) { json_array_child(this, i, result) }
+    final UnderscoreValue getChild(int i) { json_array_child(this, i, result) }
 
     /** Gets a field or child node of this node. */
     final override AstNode getAFieldOrChild() { json_array_child(this, _, result) }
@@ -1791,12 +1793,18 @@ module JSON {
     final override string getAPrimaryQlClass() { result = "Document" }
 
     /** Gets the `i`th child of this node. */
-    final Value getChild(int i) { json_document_child(this, i, result) }
+    final UnderscoreValue getChild(int i) { json_document_child(this, i, result) }
 
     /** Gets a field or child node of this node. */
     final override AstNode getAFieldOrChild() { json_document_child(this, _, result) }
   }
 
+  /** A class representing `escape_sequence` tokens. */
+  class EscapeSequence extends @json_token_escape_sequence, Token {
+    /** Gets the name of the primary QL class for this element. */
+    final override string getAPrimaryQlClass() { result = "EscapeSequence" }
+  }
+
   /** A class representing `false` tokens. */
   class False extends @json_token_false, Token {
     /** Gets the name of the primary QL class for this element. */
@@ -1833,10 +1841,10 @@ module JSON {
     final override string getAPrimaryQlClass() { result = "Pair" }
 
     /** Gets the node corresponding to the field `key`. */
-    final AstNode getKey() { json_pair_def(this, result, _) }
+    final String getKey() { json_pair_def(this, result, _) }
 
     /** Gets the node corresponding to the field `value`. */
-    final Value getValue() { json_pair_def(this, _, result) }
+    final UnderscoreValue getValue() { json_pair_def(this, _, result) }
 
     /** Gets a field or child node of this node. */
     final override AstNode getAFieldOrChild() {
@@ -1849,11 +1857,11 @@ module JSON {
     /** Gets the name of the primary QL class for this element. */
     final override string getAPrimaryQlClass() { result = "String" }
 
-    /** Gets the child of this node. */
-    final StringContent getChild() { json_string_child(this, result) }
+    /** Gets the `i`th child of this node. */
+    final AstNode getChild(int i) { json_string_child(this, i, result) }
 
     /** Gets a field or child node of this node. */
-    final override AstNode getAFieldOrChild() { json_string_child(this, result) }
+    final override AstNode getAFieldOrChild() { json_string_child(this, _, result) }
   }
 
   /** A class representing `string_content` tokens. */
@@ -1867,6 +1875,4 @@ module JSON {
     /** Gets the name of the primary QL class for this element. */
     final override string getAPrimaryQlClass() { result = "True" }
   }
-
-  class Value extends @json_value, AstNode { }
 }

ql/ql/src/codeql_ql/ast/internal/TreeSitter.qll

@@ -1239,11 +1239,13 @@ blame_ast_node_parent(
 );
 
 /*- JSON dbscheme -*/
+@json_underscore_value = @json_array | @json_object | @json_string__ | @json_token_false | @json_token_null | @json_token_number | @json_token_true
+
 #keyset[json_array, index]
 json_array_child(
   int json_array: @json_array ref,
   int index: int ref,
-  unique int child: @json_value ref
+  unique int child: @json_underscore_value ref
 );
 
 json_array_def(
@@ -1254,7 +1256,7 @@ json_array_def(
 json_document_child(
   int json_document: @json_document ref,
   int index: int ref,
-  unique int child: @json_value ref
+  unique int child: @json_underscore_value ref
 );
 
 json_document_def(
@@ -1272,25 +1274,25 @@ json_object_def(
   unique int id: @json_object
 );
 
-@json_pair_key_type = @json_string__ | @json_token_number
-
 json_pair_def(
   unique int id: @json_pair,
-  int key__: @json_pair_key_type ref,
-  int value: @json_value ref
+  int key__: @json_string__ ref,
+  int value: @json_underscore_value ref
 );
 
+@json_string_child_type = @json_token_escape_sequence | @json_token_string_content
+
+#keyset[json_string__, index]
 json_string_child(
-  unique int json_string__: @json_string__ ref,
-  unique int child: @json_token_string_content ref
+  int json_string__: @json_string__ ref,
+  int index: int ref,
+  unique int child: @json_string_child_type ref
 );
 
 json_string_def(
   unique int id: @json_string__
 );
 
-@json_value = @json_array | @json_object | @json_string__ | @json_token_false | @json_token_null | @json_token_number | @json_token_true
-
 json_tokeninfo(
   unique int id: @json_token,
   int kind: int ref,
@@ -1300,11 +1302,12 @@ json_tokeninfo(
 case @json_token.kind of
   0 = @json_reserved_word
 | 1 = @json_token_comment
-| 2 = @json_token_false
-| 3 = @json_token_null
-| 4 = @json_token_number
-| 5 = @json_token_string_content
-| 6 = @json_token_true
+| 2 = @json_token_escape_sequence
+| 3 = @json_token_false
+| 4 = @json_token_null
+| 5 = @json_token_number
+| 6 = @json_token_string_content
+| 7 = @json_token_true
 ;
 
 

ql/ql/src/ql.dbscheme

@@ -2,6 +2,6 @@
 # extractor. It is set to the lowest version of Rust we want to support.
 
 [toolchain]
-channel = "1.68"
+channel = "1.70"
 profile = "minimal"
-components = [ "rustfmt" ]
+components = [ "rustfmt" ]