Ruby: Move regex/non-regex split into TAstNode to convey disjointness

Author: hvitved

ruby/ql/lib/codeql/ruby/ast/Literal.qll

@@ -274,12 +274,10 @@ class StringComponent extends AstNode, TStringComponent {
  * "foo#{ bar() } baz"
  * ```
  */
-class StringTextComponent extends StringComponent, TStringTextComponent {
+class StringTextComponent extends StringComponent, TStringTextComponentNonRegexp {
   private Ruby::Token g;
 
-  StringTextComponent() {
-    this = TStringTextComponent(g) and not g.getParent() instanceof Ruby::Regex
-  }
+  StringTextComponent() { this = TStringTextComponentNonRegexp(g) }
 
   final override string toString() { result = g.getValue() }
 
@@ -291,12 +289,10 @@ class StringTextComponent extends StringComponent, TStringTextComponent {
 /**
  * An escape sequence component of a string or string-like literal.
  */
-class StringEscapeSequenceComponent extends StringComponent, TStringEscapeSequenceComponent {
+class StringEscapeSequenceComponent extends StringComponent, TStringEscapeSequenceComponentNonRegexp {
   private Ruby::EscapeSequence g;
 
-  StringEscapeSequenceComponent() {
-    this = TStringEscapeSequenceComponent(g) and not g.getParent() instanceof Ruby::Regex
-  }
+  StringEscapeSequenceComponent() { this = TStringEscapeSequenceComponentNonRegexp(g) }
 
   final override string toString() { result = g.getValue() }
 
@@ -309,12 +305,10 @@ class StringEscapeSequenceComponent extends StringComponent, TStringEscapeSequen
  * An interpolation expression component of a string or string-like literal.
  */
 class StringInterpolationComponent extends StringComponent, StmtSequence,
-  TStringInterpolationComponent {
+  TStringInterpolationComponentNonRegexp {
   private Ruby::Interpolation g;
 
-  StringInterpolationComponent() {
-    this = TStringInterpolationComponent(g) and not g.getParent() instanceof Ruby::Regex
-  }
+  StringInterpolationComponent() { this = TStringInterpolationComponentNonRegexp(g) }
 
   final override string toString() { result = "#{...}" }
 
@@ -325,14 +319,15 @@ class StringInterpolationComponent extends StringComponent, StmtSequence,
   final override string getAPrimaryQlClass() { result = "StringInterpolationComponent" }
 }
 
+private class TRegExpComponent =
+  TStringTextComponentRegexp or TStringEscapeSequenceComponentRegexp or
+      TStringInterpolationComponentRegexp;
+
 /**
  * The base class for a component of a regular expression literal.
  */
-class RegExpComponent extends AstNode, TStringComponent {
-  private RegExpLiteral parent;
-
-  RegExpComponent() { toGenerated(this).getParent() = toGenerated(parent) }
-
+class RegExpComponent extends AstNode, TRegExpComponent {
+  /** Gets the source text for this regex component, if any. */
   string getValueText() { none() }
 }
 
@@ -348,10 +343,10 @@ class RegExpComponent extends AstNode, TStringComponent {
  * "foo#{ bar() } baz"
  * ```
  */
-class RegExpTextComponent extends RegExpComponent, TStringTextComponent {
+class RegExpTextComponent extends RegExpComponent, TStringTextComponentRegexp {
   private Ruby::Token g;
 
-  RegExpTextComponent() { this = TStringTextComponent(g) and g.getParent() instanceof Ruby::Regex }
+  RegExpTextComponent() { this = TStringTextComponentRegexp(g) }
 
   final override string toString() { result = g.getValue() }
 
@@ -367,10 +362,10 @@ class RegExpTextComponent extends RegExpComponent, TStringTextComponent {
 /**
  * An escape sequence component of a regex literal.
  */
-class RegExpEscapeSequenceComponent extends RegExpComponent, TStringEscapeSequenceComponent {
+class RegExpEscapeSequenceComponent extends RegExpComponent, TStringEscapeSequenceComponentRegexp {
   private Ruby::EscapeSequence g;
 
-  RegExpEscapeSequenceComponent() { this = TStringEscapeSequenceComponent(g) }
+  RegExpEscapeSequenceComponent() { this = TStringEscapeSequenceComponentRegexp(g) }
 
   final override string toString() { result = g.getValue() }
 
@@ -387,10 +382,10 @@ class RegExpEscapeSequenceComponent extends RegExpComponent, TStringEscapeSequen
  * An interpolation expression component of a regex literal.
  */
 class RegExpInterpolationComponent extends RegExpComponent, StmtSequence,
-  TStringInterpolationComponent {
+  TStringInterpolationComponentRegexp {
   private Ruby::Interpolation g;
 
-  RegExpInterpolationComponent() { this = TStringInterpolationComponent(g) }
+  RegExpInterpolationComponent() { this = TStringInterpolationComponentRegexp(g) }
 
   final override string toString() { result = "#{...}" }
 

ruby/ql/lib/codeql/ruby/ast/internal/AST.qll

@@ -272,10 +272,25 @@ private module Cached {
     TStmtSequenceSynth(AST::AstNode parent, int i) { mkSynthChild(StmtSequenceKind(), parent, i) } or
     TStringArrayLiteral(Ruby::StringArray g) or
     TStringConcatenation(Ruby::ChainedString g) or
-    TStringEscapeSequenceComponent(Ruby::EscapeSequence g) or
-    TStringInterpolationComponent(Ruby::Interpolation g) or
-    TStringTextComponent(Ruby::Token g) {
-      g instanceof Ruby::StringContent or g instanceof Ruby::HeredocContent
+    TStringEscapeSequenceComponentNonRegexp(Ruby::EscapeSequence g) {
+      not g.getParent() instanceof Ruby::Regex
+    } or
+    TStringEscapeSequenceComponentRegexp(Ruby::EscapeSequence g) {
+      g.getParent() instanceof Ruby::Regex
+    } or
+    TStringInterpolationComponentNonRegexp(Ruby::Interpolation g) {
+      not g.getParent() instanceof Ruby::Regex
+    } or
+    TStringInterpolationComponentRegexp(Ruby::Interpolation g) {
+      g.getParent() instanceof Ruby::Regex
+    } or
+    TStringTextComponentNonRegexp(Ruby::Token g) {
+      (g instanceof Ruby::StringContent or g instanceof Ruby::HeredocContent) and
+      not g.getParent() instanceof Ruby::Regex
+    } or
+    TStringTextComponentRegexp(Ruby::Token g) {
+      (g instanceof Ruby::StringContent or g instanceof Ruby::HeredocContent) and
+      g.getParent() instanceof Ruby::Regex
     } or
     TSubExprReal(Ruby::Binary g) { g instanceof @ruby_binary_minus } or
     TSubExprSynth(AST::AstNode parent, int i) { mkSynthChild(SubExprKind(), parent, i) } or
@@ -489,9 +504,12 @@ private module Cached {
     n = TSplatParameter(result) or
     n = TStringArrayLiteral(result) or
     n = TStringConcatenation(result) or
-    n = TStringEscapeSequenceComponent(result) or
-    n = TStringInterpolationComponent(result) or
-    n = TStringTextComponent(result) or
+    n = TStringEscapeSequenceComponentNonRegexp(result) or
+    n = TStringEscapeSequenceComponentRegexp(result) or
+    n = TStringInterpolationComponentNonRegexp(result) or
+    n = TStringInterpolationComponentRegexp(result) or
+    n = TStringTextComponentNonRegexp(result) or
+    n = TStringTextComponentRegexp(result) or
     n = TSubExprReal(result) or
     n = TSubshellLiteral(result) or
     n = TSymbolArrayLiteral(result) or
@@ -680,6 +698,14 @@ class TIntegerLiteral = TIntegerLiteralReal or TIntegerLiteralSynth;
 
 class TBooleanLiteral = TTrueLiteral or TFalseLiteral;
 
+class TStringTextComponent = TStringTextComponentNonRegexp or TStringTextComponentRegexp;
+
+class TStringEscapeSequenceComponent =
+  TStringEscapeSequenceComponentNonRegexp or TStringEscapeSequenceComponentRegexp;
+
+class TStringInterpolationComponent =
+  TStringInterpolationComponentNonRegexp or TStringInterpolationComponentRegexp;
+
 class TStringComponent =
   TStringTextComponent or TStringEscapeSequenceComponent or TStringInterpolationComponent;