Ruby: Remove RegExpLiteral.getAMatch

This predicate is a duplicate of getAMatchedString, which matches the
naming in the JS version.

Author: hmac

ruby/ql/lib/codeql/ruby/regexp/RegExpTreeView.qll

@@ -271,9 +271,6 @@ class RegExpTerm extends RegExpParent {
 
   /** Holds if this regular expression term can match the empty string. */
   predicate isNullable() { none() }
-
-  /** Gets a string matched by this regular expression. */
-  string getAMatch() { none() }
 }
 
 /**
@@ -458,20 +455,6 @@ class RegExpSequence extends RegExpTerm, TRegExpSequence {
   override predicate isNullable() {
     forall(RegExpTerm child | child = this.getAChild() | child.isNullable())
   }
-
-  // Why can't we use concat(...) with language[monotonicAggregates] here instead?
-  override string getAMatch() { result = this.getAMatchFromChildAtIndex(0) }
-
-  private string getAMatchFromChildAtIndex(int i) {
-    i = this.getNumChild() and result = ""
-    or
-    exists(string substring, string rest |
-      substring = this.getChild(i).getAMatch() and
-      rest = this.getAMatchFromChildAtIndex(i + 1)
-    |
-      result = substring + rest
-    )
-  }
 }
 
 pragma[nomagic]
@@ -703,8 +686,6 @@ class RegExpCharacterClass extends RegExpTerm, TRegExpCharacterClass {
   override string getAPrimaryQlClass() { result = "RegExpCharacterClass" }
 
   override predicate isNullable() { none() }
-
-  override string getAMatch() { not this.isInverted() and result = this.getAChild().getAMatch() }
 }
 
 /**
@@ -819,8 +800,6 @@ class RegExpConstant extends RegExpTerm {
   override string getAPrimaryQlClass() { result = "RegExpConstant" }
 
   override predicate isNullable() { none() }
-
-  override string getAMatch() { result = this.getValue() }
 }
 
 /**
@@ -870,8 +849,6 @@ class RegExpGroup extends RegExpTerm, TRegExpGroup {
   override string getAPrimaryQlClass() { result = "RegExpGroup" }
 
   override predicate isNullable() { this.getAChild().isNullable() }
-
-  override string getAMatch() { result = this.getAChild().getAMatch() }
 }
 
 /**

ruby/ql/lib/codeql/ruby/security/IncompleteMultiCharacterSanitizationQuery.qll

@@ -60,7 +60,7 @@ pragma[noinline]
 private DangerousPrefixSubstring getADangerousMatchedChar(EmptyReplaceRegExpTerm t) {
   t.isNullable() and result = ""
   or
-  result = t.getAMatch()
+  result = t.getAMatchedString()
   or
   // A substring matched by some character class. This is only used to match the "word" part of a HTML tag (e.g. "iframe" in "<iframe").
   exists(ReDoSUtil::CharacterClass cc |
@@ -123,14 +123,14 @@ private predicate matchesDangerousPrefix(EmptyReplaceRegExpTerm t, string prefix
     kind = "path injection" and
     prefix = ["/..", "../"] and
     // If the regex is matching explicit path components, it is unlikely that it's being used as a sanitizer.
-    not t.getSuccessor*().getAMatch().regexpMatch("(?is).*[a-z0-9_-].*")
+    not t.getSuccessor*().getAMatchedString().regexpMatch("(?is).*[a-z0-9_-].*")
     or
     kind = "HTML element injection" and
     (
       // comments
       prefix = "<!--" and
       // If the regex is matching explicit textual content of an HTML comment, it is unlikely that it's being used as a sanitizer.
-      not t.getSuccessor*().getAMatch().regexpMatch("(?is).*[a-z0-9_].*")
+      not t.getSuccessor*().getAMatchedString().regexpMatch("(?is).*[a-z0-9_].*")
       or
       // specific tags
       // the `cript|scrip` case has been observed in the wild several times
@@ -148,11 +148,11 @@ private predicate matchesDangerousPrefix(EmptyReplaceRegExpTerm t, string prefix
     ] and
   (
     // explicit matching: `onclick` and `ng-bind`
-    t.getAMatch().regexpMatch("(?i)" + prefix + "[a-z]+")
+    t.getAMatchedString().regexpMatch("(?i)" + prefix + "[a-z]+")
     or
     // regexp-based matching: `on[a-z]+`
     exists(EmptyReplaceRegExpTerm start | start = t.getAChild() |
-      start.getAMatch().regexpMatch("(?i)[^a-z]*" + prefix) and
+      start.getAMatchedString().regexpMatch("(?i)[^a-z]*" + prefix) and
       isCommonWordMatcher(start.getSuccessor())
     )
   )