Commit 1: Remove \p{...}/\P{...} named-character-property handling

std::regex ECMAScript mode does not support \p{Name}; \p tokenizes as an identity escape. Remove pStyleNamedCharacterProperty and its call sites; namedCharacterProperty now covers only POSIX brackets ([:name:], [.x.], [=x=]) and namedCharacterPropertyIsInverted keeps only the [[:^name:]] case (fixing the offset from start+3 to start+2 for POSIX). Update RegExpNamedCharacterProperty getName/isInverted docs to describe POSIX brackets. Remove the four \p{...}/\P{...} corpus lines from test.cpp; keep a plain [a-f\d]+ case for the class-with-escape shape. Bundle: github/codeql-action codeql-bundle-v2.26.1 (CodeQL CLI 2.26.1).
2026-07-31 15:33:00 +02:00 · 2026-07-21 12:16:37 +00:00
parent bfa65301f1
commit 448ae6af53
6 changed files with 196 additions and 266 deletions
--- a/cpp/ql/lib/semmle/code/cpp/regex/RegexTreeView.qll
+++ b/cpp/ql/lib/semmle/code/cpp/regex/RegexTreeView.qll
@@ -1192,13 +1192,13 @@ private module Impl implements RegexTreeViewSig {
    override string getAPrimaryQlClass() { result = "RegExpNamedCharacterProperty" }

    /**
-     * Gets the property name. For example, in `\p{Space}`, the result is
-     * `"Space"`.
+     * Gets the property name. For example, in `[[:digit:]]`, the result is
+     * `"digit"`.
     */
    string getName() { result = re.getCharacterPropertyName(start, end) }

    /**
-     * Holds if the property is inverted. For example, it holds for `\p{^Digit}`,
+     * Holds if the property is inverted. For example, it holds for `[[:^digit:]]`,
     * which matches non-digits.
     */
    predicate isInverted() { re.namedCharacterPropertyIsInverted(start, end) }
--- a/cpp/ql/lib/semmle/code/cpp/regex/internal/ParseRegExp.qll
+++ b/cpp/ql/lib/semmle/code/cpp/regex/internal/ParseRegExp.qll
@@ -282,9 +282,8 @@ class RegExp extends StringLiteral {
    )
  }

-  /** Matches named character properties such as `\p{Word}` and `[[:digit:]]` */
+  /** Matches named character properties such as `[[:digit:]]`, `[.a.]`, and `[=a=]`. */
  predicate namedCharacterProperty(int start, int end, string name) {
-    this.pStyleNamedCharacterProperty(start, end, name) or
    this.posixStyleNamedCharacterProperty(start, end, name) or
    this.posixCollatingSymbol(start, end, name) or
    this.posixEquivalenceClass(start, end, name)
@@ -367,52 +366,16 @@ class RegExp extends StringLiteral {
    name = this.getText().substring(start + 2, end - 2)
  }

-  /**
-   * Matches named character properties. For example:
-   * - `\p{Space}`
-   * - `\P{Digit}` upper-case P means inverted
-   * - `\p{^Word}` caret also means inverted
-   *
-   * These can occur both inside and outside of character classes.
-   */
-  private predicate pStyleNamedCharacterProperty(int start, int end, string name) {
-    this.escapingChar(start) and
-    this.getChar(start + 1) in ["p", "P"] and
-    this.getChar(start + 2) = "{" and
-    this.getChar(end - 1) = "}" and
-    end > start and
-    not exists(int i | start + 2 < i and i < end - 1 | this.getChar(i) = "}") and
-    exists(int nameStart |
-      this.getChar(start + 3) = "^" and nameStart = start + 4
-      or
-      not this.getChar(start + 3) = "^" and nameStart = start + 3
-    |
-      name = this.getText().substring(nameStart, end - 1)
-    )
-  }
-
  /**
   * Holds if the named character property is inverted. Examples for which it holds:
-   * - `\P{Digit}` upper-case P means inverted
-   * - `\p{^Word}` caret also means inverted
   * - `[[:^digit:]]`
   *
   * Examples for which it doesn't hold:
-   * - `\p{Word}`
-   * - `\P{^Space}` - upper-case P and caret cancel each other out
   * - `[[:alnum:]]`
   */
  predicate namedCharacterPropertyIsInverted(int start, int end) {
-    this.pStyleNamedCharacterProperty(start, end, _) and
-    exists(boolean upperP, boolean caret |
-      (if this.getChar(start + 1) = "P" then upperP = true else upperP = false) and
-      (if this.getChar(start + 3) = "^" then caret = true else caret = false)
-    |
-      upperP.booleanXor(caret) = true
-    )
-    or
    this.posixStyleNamedCharacterProperty(start, end, _) and
-    this.getChar(start + 3) = "^"
+    this.getChar(start + 2) = "^"
  }

  /**
@@ -424,7 +387,6 @@ class RegExp extends StringLiteral {
    this.escapingChar(start) and
    not this.numberedBackreference(start, _, _) and
    not this.namedBackreference(start, _, _) and
-    not this.pStyleNamedCharacterProperty(start, _, _) and
    (
      // hex char \xhh
      this.getChar(start + 1) = "x" and end = start + 4
@@ -502,9 +464,6 @@ class RegExp extends StringLiteral {
    ) and
    not exists(int x, int y | this.groupStart(x, y) and x <= start and y >= end) and
    not exists(int x, int y | this.backreference(x, y) and x <= start and y >= end) and
-    not exists(int x, int y |
-      this.pStyleNamedCharacterProperty(x, y, _) and x <= start and y >= end
-    ) and
    not exists(int x, int y | this.multiples(x, y, _, _) and x <= start and y >= end)
  }

@@ -831,8 +790,6 @@ class RegExp extends StringLiteral {
    this.charSet(start, end)
    or
    this.backreference(start, end)
-    or
-    this.pStyleNamedCharacterProperty(start, end, _)
  }

  private predicate qualifier(int start, int end, boolean maybeEmpty, boolean mayRepeatForever) {
--- a/cpp/ql/test/library-tests/regex/locations.expected
+++ b/cpp/ql/test/library-tests/regex/locations.expected
@@ -1,40 +1,36 @@
-| test.cpp:144:23:144:23 | a | 22 | 0 | 1 | 23 | 23 |
-| test.cpp:144:23:144:24 | a+ | 22 | 0 | 2 | 23 | 24 |
-| test.cpp:144:23:144:25 | a+b | 22 | 0 | 3 | 23 | 25 |
-| test.cpp:144:25:144:25 | b | 22 | 2 | 3 | 25 | 25 |
-| test.cpp:147:23:147:23 | a | 20 | 0 | 1 | 23 | 23 |
-| test.cpp:147:23:147:24 | a+ | 20 | 0 | 2 | 23 | 24 |
-| test.cpp:147:23:147:25 | a+b | 20 | 0 | 3 | 23 | 25 |
-| test.cpp:147:25:147:25 | b | 20 | 2 | 3 | 25 | 25 |
-| test.cpp:150:24:150:25 | \\s | 21 | 0 | 2 | 24 | 25 |
-| test.cpp:150:24:150:26 | \\s+ | 21 | 0 | 3 | 24 | 26 |
-| test.cpp:150:24:150:27 | \\s+$ | 21 | 0 | 4 | 24 | 27 |
-| test.cpp:150:27:150:27 | $ | 21 | 3 | 4 | 27 | 27 |
-| test.cpp:153:24:153:25 | \\( | 21 | 0 | 2 | 24 | 25 |
-| test.cpp:153:24:153:37 | \\(([,\\w]+)+\\)$ | 21 | 0 | 14 | 24 | 37 |
-| test.cpp:153:26:153:33 | ([,\\w]+) | 21 | 2 | 10 | 26 | 33 |
-| test.cpp:153:26:153:34 | ([,\\w]+)+ | 21 | 2 | 11 | 26 | 34 |
-| test.cpp:153:27:153:31 | [,\\w] | 21 | 3 | 8 | 27 | 31 |
-| test.cpp:153:27:153:32 | [,\\w]+ | 21 | 3 | 9 | 27 | 32 |
-| test.cpp:153:28:153:28 | , | 21 | 4 | 5 | 28 | 28 |
-| test.cpp:153:29:153:30 | \\w | 21 | 5 | 7 | 29 | 30 |
-| test.cpp:153:35:153:36 | \\) | 21 | 11 | 13 | 35 | 36 |
-| test.cpp:153:37:153:37 | $ | 21 | 13 | 14 | 37 | 37 |
-| test.cpp:156:25:156:25 | a | 21 | 0 | 1 | 25 | 25 |
-| test.cpp:156:25:156:26 | a+ | 21 | 0 | 2 | 25 | 26 |
-| test.cpp:156:25:156:27 | a+b | 21 | 0 | 3 | 25 | 27 |
-| test.cpp:156:27:156:27 | b | 21 | 2 | 3 | 27 | 27 |
-| test.cpp:159:24:159:24 | a | 22 | 0 | 1 | 24 | 24 |
-| test.cpp:159:24:159:25 | a+ | 22 | 0 | 2 | 24 | 25 |
-| test.cpp:159:24:159:26 | a+b | 22 | 0 | 3 | 24 | 26 |
-| test.cpp:159:26:159:26 | b | 22 | 2 | 3 | 26 | 26 |
-| test.cpp:162:30:162:30 | a | 26 | 0 | 1 | 30 | 30 |
-| test.cpp:162:30:162:31 | a+ | 26 | 0 | 2 | 30 | 31 |
-| test.cpp:162:30:162:32 | a+b | 26 | 0 | 3 | 30 | 32 |
-| test.cpp:162:32:162:32 | b | 26 | 2 | 3 | 32 | 32 |
-| test.cpp:166:22:166:23 | \\s | 21 | 0 | 2 | 22 | 23 |
-| test.cpp:166:22:166:24 | \\s+ | 21 | 0 | 3 | 22 | 24 |
-| test.cpp:169:22:169:22 | a | 21 | 0 | 1 | 22 | 22 |
-| test.cpp:169:22:169:25 | a\\.b | 21 | 0 | 4 | 22 | 25 |
-| test.cpp:169:23:169:24 | \\. | 21 | 1 | 3 | 23 | 24 |
-| test.cpp:169:25:169:25 | b | 21 | 3 | 4 | 25 | 25 |
+| test.cpp:144:23:144:23 | a | 20 | 0 | 1 | 23 | 23 |
+| test.cpp:144:23:144:24 | a+ | 20 | 0 | 2 | 23 | 24 |
+| test.cpp:144:23:144:25 | a+b | 20 | 0 | 3 | 23 | 25 |
+| test.cpp:144:25:144:25 | b | 20 | 2 | 3 | 25 | 25 |
+| test.cpp:147:24:147:25 | \\s | 21 | 0 | 2 | 24 | 25 |
+| test.cpp:147:24:147:26 | \\s+ | 21 | 0 | 3 | 24 | 26 |
+| test.cpp:147:24:147:27 | \\s+$ | 21 | 0 | 4 | 24 | 27 |
+| test.cpp:147:27:147:27 | $ | 21 | 3 | 4 | 27 | 27 |
+| test.cpp:150:24:150:25 | \\( | 21 | 0 | 2 | 24 | 25 |
+| test.cpp:150:24:150:37 | \\(([,\\w]+)+\\)$ | 21 | 0 | 14 | 24 | 37 |
+| test.cpp:150:26:150:33 | ([,\\w]+) | 21 | 2 | 10 | 26 | 33 |
+| test.cpp:150:26:150:34 | ([,\\w]+)+ | 21 | 2 | 11 | 26 | 34 |
+| test.cpp:150:27:150:31 | [,\\w] | 21 | 3 | 8 | 27 | 31 |
+| test.cpp:150:27:150:32 | [,\\w]+ | 21 | 3 | 9 | 27 | 32 |
+| test.cpp:150:28:150:28 | , | 21 | 4 | 5 | 28 | 28 |
+| test.cpp:150:29:150:30 | \\w | 21 | 5 | 7 | 29 | 30 |
+| test.cpp:150:35:150:36 | \\) | 21 | 11 | 13 | 35 | 36 |
+| test.cpp:150:37:150:37 | $ | 21 | 13 | 14 | 37 | 37 |
+| test.cpp:153:25:153:25 | a | 21 | 0 | 1 | 25 | 25 |
+| test.cpp:153:25:153:26 | a+ | 21 | 0 | 2 | 25 | 26 |
+| test.cpp:153:25:153:27 | a+b | 21 | 0 | 3 | 25 | 27 |
+| test.cpp:153:27:153:27 | b | 21 | 2 | 3 | 27 | 27 |
+| test.cpp:156:24:156:24 | a | 22 | 0 | 1 | 24 | 24 |
+| test.cpp:156:24:156:25 | a+ | 22 | 0 | 2 | 24 | 25 |
+| test.cpp:156:24:156:26 | a+b | 22 | 0 | 3 | 24 | 26 |
+| test.cpp:156:26:156:26 | b | 22 | 2 | 3 | 26 | 26 |
+| test.cpp:159:30:159:30 | a | 26 | 0 | 1 | 30 | 30 |
+| test.cpp:159:30:159:31 | a+ | 26 | 0 | 2 | 30 | 31 |
+| test.cpp:159:30:159:32 | a+b | 26 | 0 | 3 | 30 | 32 |
+| test.cpp:159:32:159:32 | b | 26 | 2 | 3 | 32 | 32 |
+| test.cpp:163:22:163:23 | \\s | 21 | 0 | 2 | 22 | 23 |
+| test.cpp:163:22:163:24 | \\s+ | 21 | 0 | 3 | 22 | 24 |
+| test.cpp:166:22:166:22 | a | 21 | 0 | 1 | 22 | 22 |
+| test.cpp:166:22:166:25 | a\\.b | 21 | 0 | 4 | 22 | 25 |
+| test.cpp:166:23:166:24 | \\. | 21 | 1 | 3 | 23 | 24 |
+| test.cpp:166:25:166:25 | b | 21 | 3 | 4 | 25 | 25 |
--- a/cpp/ql/test/library-tests/regex/parse.expected
+++ b/cpp/ql/test/library-tests/regex/parse.expected
@@ -442,86 +442,82 @@ test.cpp:
 #  113| [RegExpPlus] \k<qux>+
 #-----| 0 -> [RegExpBackRef] \k<qux>

-#  116| [RegExpNamedCharacterProperty] \p{Word}
-
-#  116| [RegExpStar] \p{Word}*
-#-----| 0 -> [RegExpNamedCharacterProperty] \p{Word}
-
-#  117| [RegExpNamedCharacterProperty] \P{Digit}
-
-#  117| [RegExpPlus] \P{Digit}+
-#-----| 0 -> [RegExpNamedCharacterProperty] \P{Digit}
-
-#  118| [RegExpNamedCharacterProperty] \p{^Alnum}
-
-#  118| [RegExpRange] \p{^Alnum}{2,3}
-#-----| 0 -> [RegExpNamedCharacterProperty] \p{^Alnum}
-
-#  119| [RegExpCharacterClass] [a-f\p{Digit}]
+#  116| [RegExpCharacterClass] [a-f\d]
 #-----| 0 -> [RegExpCharacterRange] a-f
-#-----| 1 -> [RegExpNamedCharacterProperty] \p{Digit}
+#-----| 1 -> [RegExpCharacterClassEscape] \d

-#  119| [RegExpPlus] [a-f\p{Digit}]+
-#-----| 0 -> [RegExpCharacterClass] [a-f\p{Digit}]
+#  116| [RegExpPlus] [a-f\d]+
+#-----| 0 -> [RegExpCharacterClass] [a-f\d]

-#  119| [RegExpConstant, RegExpNormalChar] a
+#  116| [RegExpConstant, RegExpNormalChar] a

-#  119| [RegExpCharacterRange] a-f
+#  116| [RegExpCharacterRange] a-f
 #-----| 0 -> [RegExpConstant, RegExpNormalChar] a
 #-----| 1 -> [RegExpConstant, RegExpNormalChar] f

-#  119| [RegExpConstant, RegExpNormalChar] f
+#  116| [RegExpConstant, RegExpNormalChar] f

-#  119| [RegExpNamedCharacterProperty] \p{Digit}
+#  116| [RegExpCharacterClassEscape] \d

-#  122| [RegExpCharacterClass] [[:alpha:]]
+#  119| [RegExpCharacterClass] [[:alpha:]]
 #-----| 0 -> [RegExpNamedCharacterProperty] [:alpha:]

-#  122| [RegExpSequence] [[:alpha:]][[:digit:]]
+#  119| [RegExpSequence] [[:alpha:]][[:digit:]]
 #-----| 0 -> [RegExpCharacterClass] [[:alpha:]]
 #-----| 1 -> [RegExpCharacterClass] [[:digit:]]

-#  122| [RegExpNamedCharacterProperty] [:alpha:]
+#  119| [RegExpNamedCharacterProperty] [:alpha:]

-#  122| [RegExpCharacterClass] [[:digit:]]
+#  119| [RegExpCharacterClass] [[:digit:]]
 #-----| 0 -> [RegExpNamedCharacterProperty] [:digit:]

-#  122| [RegExpNamedCharacterProperty] [:digit:]
+#  119| [RegExpNamedCharacterProperty] [:digit:]

-#  125| [RegExpCharacterClass] [[:alpha:][:digit:]]
+#  122| [RegExpCharacterClass] [[:alpha:][:digit:]]
 #-----| 0 -> [RegExpNamedCharacterProperty] [:alpha:]
 #-----| 1 -> [RegExpNamedCharacterProperty] [:digit:]

-#  125| [RegExpNamedCharacterProperty] [:alpha:]
+#  122| [RegExpNamedCharacterProperty] [:alpha:]

-#  125| [RegExpNamedCharacterProperty] [:digit:]
+#  122| [RegExpNamedCharacterProperty] [:digit:]

-#  128| [RegExpCharacterClass] [A-F[:digit:]a-f]
+#  125| [RegExpCharacterClass] [A-F[:digit:]a-f]
 #-----| 0 -> [RegExpCharacterRange] A-F
 #-----| 1 -> [RegExpNamedCharacterProperty] [:digit:]
 #-----| 2 -> [RegExpCharacterRange] a-f

-#  128| [RegExpConstant, RegExpNormalChar] A
+#  125| [RegExpConstant, RegExpNormalChar] A

-#  128| [RegExpCharacterRange] A-F
+#  125| [RegExpCharacterRange] A-F
 #-----| 0 -> [RegExpConstant, RegExpNormalChar] A
 #-----| 1 -> [RegExpConstant, RegExpNormalChar] F

-#  128| [RegExpConstant, RegExpNormalChar] F
+#  125| [RegExpConstant, RegExpNormalChar] F

-#  128| [RegExpNamedCharacterProperty] [:digit:]
+#  125| [RegExpNamedCharacterProperty] [:digit:]

-#  128| [RegExpConstant, RegExpNormalChar] a
+#  125| [RegExpConstant, RegExpNormalChar] a

-#  128| [RegExpCharacterRange] a-f
+#  125| [RegExpCharacterRange] a-f
 #-----| 0 -> [RegExpConstant, RegExpNormalChar] a
 #-----| 1 -> [RegExpConstant, RegExpNormalChar] f

-#  128| [RegExpConstant, RegExpNormalChar] f
+#  125| [RegExpConstant, RegExpNormalChar] f

-#  131| [RegExpNamedCharacterProperty] [:digit:]
+#  128| [RegExpNamedCharacterProperty] [:digit:]

-#  136| [RegExpConstant, RegExpEscape] \u9879
+#  133| [RegExpConstant, RegExpEscape] \u9879
+
+#  141| [RegExpConstant, RegExpNormalChar] a
+
+#  141| [RegExpPlus] a+
+#-----| 0 -> [RegExpConstant, RegExpNormalChar] a
+
+#  141| [RegExpSequence] a+b
+#-----| 0 -> [RegExpPlus] a+
+#-----| 1 -> [RegExpConstant, RegExpNormalChar] b
+
+#  141| [RegExpConstant, RegExpNormalChar] b

 #  144| [RegExpConstant, RegExpNormalChar] a

@@ -534,56 +530,56 @@ test.cpp:

 #  144| [RegExpConstant, RegExpNormalChar] b

-#  147| [RegExpConstant, RegExpNormalChar] a
+#  147| [RegExpCharacterClassEscape] \s

-#  147| [RegExpPlus] a+
-#-----| 0 -> [RegExpConstant, RegExpNormalChar] a
-
-#  147| [RegExpSequence] a+b
-#-----| 0 -> [RegExpPlus] a+
-#-----| 1 -> [RegExpConstant, RegExpNormalChar] b
-
-#  147| [RegExpConstant, RegExpNormalChar] b
-
-#  150| [RegExpCharacterClassEscape] \s
-
-#  150| [RegExpPlus] \s+
+#  147| [RegExpPlus] \s+
 #-----| 0 -> [RegExpCharacterClassEscape] \s

-#  150| [RegExpSequence] \s+$
+#  147| [RegExpSequence] \s+$
 #-----| 0 -> [RegExpPlus] \s+
 #-----| 1 -> [RegExpDollar] $

-#  150| [RegExpDollar] $
+#  147| [RegExpDollar] $

-#  153| [RegExpConstant, RegExpEscape] \(
+#  150| [RegExpConstant, RegExpEscape] \(

-#  153| [RegExpSequence] \(([,\w]+)+\)$
+#  150| [RegExpSequence] \(([,\w]+)+\)$
 #-----| 0 -> [RegExpConstant, RegExpEscape] \(
 #-----| 1 -> [RegExpPlus] ([,\w]+)+
 #-----| 2 -> [RegExpConstant, RegExpEscape] \)
 #-----| 3 -> [RegExpDollar] $

-#  153| [RegExpGroup] ([,\w]+)
+#  150| [RegExpGroup] ([,\w]+)
 #-----| 0 -> [RegExpPlus] [,\w]+

-#  153| [RegExpPlus] ([,\w]+)+
+#  150| [RegExpPlus] ([,\w]+)+
 #-----| 0 -> [RegExpGroup] ([,\w]+)

-#  153| [RegExpCharacterClass] [,\w]
+#  150| [RegExpCharacterClass] [,\w]
 #-----| 0 -> [RegExpConstant, RegExpNormalChar] ,
 #-----| 1 -> [RegExpCharacterClassEscape] \w

-#  153| [RegExpPlus] [,\w]+
+#  150| [RegExpPlus] [,\w]+
 #-----| 0 -> [RegExpCharacterClass] [,\w]

-#  153| [RegExpConstant, RegExpNormalChar] ,
+#  150| [RegExpConstant, RegExpNormalChar] ,

-#  153| [RegExpCharacterClassEscape] \w
+#  150| [RegExpCharacterClassEscape] \w

-#  153| [RegExpConstant, RegExpEscape] \)
+#  150| [RegExpConstant, RegExpEscape] \)

-#  153| [RegExpDollar] $
+#  150| [RegExpDollar] $
+
+#  153| [RegExpConstant, RegExpNormalChar] a
+
+#  153| [RegExpPlus] a+
+#-----| 0 -> [RegExpConstant, RegExpNormalChar] a
+
+#  153| [RegExpSequence] a+b
+#-----| 0 -> [RegExpPlus] a+
+#-----| 1 -> [RegExpConstant, RegExpNormalChar] b
+
+#  153| [RegExpConstant, RegExpNormalChar] b

 #  156| [RegExpConstant, RegExpNormalChar] a

@@ -607,29 +603,18 @@ test.cpp:

 #  159| [RegExpConstant, RegExpNormalChar] b

-#  162| [RegExpConstant, RegExpNormalChar] a
+#  163| [RegExpCharacterClassEscape] \s

-#  162| [RegExpPlus] a+
-#-----| 0 -> [RegExpConstant, RegExpNormalChar] a
-
-#  162| [RegExpSequence] a+b
-#-----| 0 -> [RegExpPlus] a+
-#-----| 1 -> [RegExpConstant, RegExpNormalChar] b
-
-#  162| [RegExpConstant, RegExpNormalChar] b
-
-#  166| [RegExpCharacterClassEscape] \s
-
-#  166| [RegExpPlus] \s+
+#  163| [RegExpPlus] \s+
 #-----| 0 -> [RegExpCharacterClassEscape] \s

-#  169| [RegExpConstant, RegExpNormalChar] a
+#  166| [RegExpConstant, RegExpNormalChar] a

-#  169| [RegExpSequence] a\.b
+#  166| [RegExpSequence] a\.b
 #-----| 0 -> [RegExpConstant, RegExpNormalChar] a
 #-----| 1 -> [RegExpConstant, RegExpEscape] \.
 #-----| 2 -> [RegExpConstant, RegExpNormalChar] b

-#  169| [RegExpConstant, RegExpEscape] \.
+#  166| [RegExpConstant, RegExpEscape] \.

-#  169| [RegExpConstant, RegExpNormalChar] b
+#  166| [RegExpConstant, RegExpNormalChar] b
--- a/cpp/ql/test/library-tests/regex/regexp.expected
+++ b/cpp/ql/test/library-tests/regex/regexp.expected
@@ -8,7 +8,7 @@ groupNumber
 | test.cpp:108:21:108:30 | (?<id>\\w+) | 1 |
 | test.cpp:112:23:112:26 | (a+) | 1 |
 | test.cpp:113:23:113:32 | (?<qux>q+) | 1 |
-| test.cpp:153:26:153:33 | ([,\\w]+) | 1 |
+| test.cpp:150:26:150:33 | ([,\\w]+) | 1 |
 term
 | test.cpp:33:21:33:23 | abc | RegExpConstant,RegExpNormalChar |
 | test.cpp:36:22:36:22 | a | RegExpConstant,RegExpNormalChar |
@@ -172,76 +172,70 @@ term
 | test.cpp:113:33:113:35 | \\s+ | RegExpPlus |
 | test.cpp:113:36:113:42 | \\k<qux> | RegExpBackRef |
 | test.cpp:113:36:113:43 | \\k<qux>+ | RegExpPlus |
-| test.cpp:116:23:116:30 | \\p{Word} | RegExpNamedCharacterProperty |
-| test.cpp:116:23:116:31 | \\p{Word}* | RegExpStar |
-| test.cpp:117:23:117:31 | \\P{Digit} | RegExpNamedCharacterProperty |
-| test.cpp:117:23:117:32 | \\P{Digit}+ | RegExpPlus |
-| test.cpp:118:23:118:32 | \\p{^Alnum} | RegExpNamedCharacterProperty |
-| test.cpp:118:23:118:37 | \\p{^Alnum}{2,3} | RegExpRange |
-| test.cpp:119:23:119:36 | [a-f\\p{Digit}] | RegExpCharacterClass |
-| test.cpp:119:23:119:37 | [a-f\\p{Digit}]+ | RegExpPlus |
-| test.cpp:119:24:119:24 | a | RegExpConstant,RegExpNormalChar |
-| test.cpp:119:24:119:26 | a-f | RegExpCharacterRange |
-| test.cpp:119:26:119:26 | f | RegExpConstant,RegExpNormalChar |
-| test.cpp:119:27:119:35 | \\p{Digit} | RegExpNamedCharacterProperty |
-| test.cpp:122:24:122:34 | [[:alpha:]] | RegExpCharacterClass |
-| test.cpp:122:24:122:45 | [[:alpha:]][[:digit:]] | RegExpSequence |
+| test.cpp:116:23:116:29 | [a-f\\d] | RegExpCharacterClass |
+| test.cpp:116:23:116:30 | [a-f\\d]+ | RegExpPlus |
+| test.cpp:116:24:116:24 | a | RegExpConstant,RegExpNormalChar |
+| test.cpp:116:24:116:26 | a-f | RegExpCharacterRange |
+| test.cpp:116:26:116:26 | f | RegExpConstant,RegExpNormalChar |
+| test.cpp:116:27:116:28 | \\d | RegExpCharacterClassEscape |
+| test.cpp:119:24:119:34 | [[:alpha:]] | RegExpCharacterClass |
+| test.cpp:119:24:119:45 | [[:alpha:]][[:digit:]] | RegExpSequence |
+| test.cpp:119:25:119:33 | [:alpha:] | RegExpNamedCharacterProperty |
+| test.cpp:119:35:119:45 | [[:digit:]] | RegExpCharacterClass |
+| test.cpp:119:36:119:44 | [:digit:] | RegExpNamedCharacterProperty |
+| test.cpp:122:24:122:43 | [[:alpha:][:digit:]] | RegExpCharacterClass |
 | test.cpp:122:25:122:33 | [:alpha:] | RegExpNamedCharacterProperty |
-| test.cpp:122:35:122:45 | [[:digit:]] | RegExpCharacterClass |
-| test.cpp:122:36:122:44 | [:digit:] | RegExpNamedCharacterProperty |
-| test.cpp:125:24:125:43 | [[:alpha:][:digit:]] | RegExpCharacterClass |
-| test.cpp:125:25:125:33 | [:alpha:] | RegExpNamedCharacterProperty |
-| test.cpp:125:34:125:42 | [:digit:] | RegExpNamedCharacterProperty |
-| test.cpp:128:24:128:40 | [A-F[:digit:]a-f] | RegExpCharacterClass |
-| test.cpp:128:25:128:25 | A | RegExpConstant,RegExpNormalChar |
-| test.cpp:128:25:128:27 | A-F | RegExpCharacterRange |
-| test.cpp:128:27:128:27 | F | RegExpConstant,RegExpNormalChar |
-| test.cpp:128:28:128:36 | [:digit:] | RegExpNamedCharacterProperty |
-| test.cpp:128:37:128:37 | a | RegExpConstant,RegExpNormalChar |
-| test.cpp:128:37:128:39 | a-f | RegExpCharacterRange |
-| test.cpp:128:39:128:39 | f | RegExpConstant,RegExpNormalChar |
-| test.cpp:131:24:131:32 | [:digit:] | RegExpNamedCharacterProperty |
-| test.cpp:136:21:136:26 | \\u9879 | RegExpConstant,RegExpEscape |
+| test.cpp:122:34:122:42 | [:digit:] | RegExpNamedCharacterProperty |
+| test.cpp:125:24:125:40 | [A-F[:digit:]a-f] | RegExpCharacterClass |
+| test.cpp:125:25:125:25 | A | RegExpConstant,RegExpNormalChar |
+| test.cpp:125:25:125:27 | A-F | RegExpCharacterRange |
+| test.cpp:125:27:125:27 | F | RegExpConstant,RegExpNormalChar |
+| test.cpp:125:28:125:36 | [:digit:] | RegExpNamedCharacterProperty |
+| test.cpp:125:37:125:37 | a | RegExpConstant,RegExpNormalChar |
+| test.cpp:125:37:125:39 | a-f | RegExpCharacterRange |
+| test.cpp:125:39:125:39 | f | RegExpConstant,RegExpNormalChar |
+| test.cpp:128:24:128:32 | [:digit:] | RegExpNamedCharacterProperty |
+| test.cpp:133:21:133:26 | \\u9879 | RegExpConstant,RegExpEscape |
+| test.cpp:141:23:141:23 | a | RegExpConstant,RegExpNormalChar |
+| test.cpp:141:23:141:24 | a+ | RegExpPlus |
+| test.cpp:141:23:141:25 | a+b | RegExpSequence |
+| test.cpp:141:25:141:25 | b | RegExpConstant,RegExpNormalChar |
 | test.cpp:144:23:144:23 | a | RegExpConstant,RegExpNormalChar |
 | test.cpp:144:23:144:24 | a+ | RegExpPlus |
 | test.cpp:144:23:144:25 | a+b | RegExpSequence |
 | test.cpp:144:25:144:25 | b | RegExpConstant,RegExpNormalChar |
-| test.cpp:147:23:147:23 | a | RegExpConstant,RegExpNormalChar |
-| test.cpp:147:23:147:24 | a+ | RegExpPlus |
-| test.cpp:147:23:147:25 | a+b | RegExpSequence |
-| test.cpp:147:25:147:25 | b | RegExpConstant,RegExpNormalChar |
-| test.cpp:150:24:150:25 | \\s | RegExpCharacterClassEscape |
-| test.cpp:150:24:150:26 | \\s+ | RegExpPlus |
-| test.cpp:150:24:150:27 | \\s+$ | RegExpSequence |
-| test.cpp:150:27:150:27 | $ | RegExpDollar |
-| test.cpp:153:24:153:25 | \\( | RegExpConstant,RegExpEscape |
-| test.cpp:153:24:153:37 | \\(([,\\w]+)+\\)$ | RegExpSequence |
-| test.cpp:153:26:153:33 | ([,\\w]+) | RegExpGroup |
-| test.cpp:153:26:153:34 | ([,\\w]+)+ | RegExpPlus |
-| test.cpp:153:27:153:31 | [,\\w] | RegExpCharacterClass |
-| test.cpp:153:27:153:32 | [,\\w]+ | RegExpPlus |
-| test.cpp:153:28:153:28 | , | RegExpConstant,RegExpNormalChar |
-| test.cpp:153:29:153:30 | \\w | RegExpCharacterClassEscape |
-| test.cpp:153:35:153:36 | \\) | RegExpConstant,RegExpEscape |
-| test.cpp:153:37:153:37 | $ | RegExpDollar |
-| test.cpp:156:25:156:25 | a | RegExpConstant,RegExpNormalChar |
-| test.cpp:156:25:156:26 | a+ | RegExpPlus |
-| test.cpp:156:25:156:27 | a+b | RegExpSequence |
-| test.cpp:156:27:156:27 | b | RegExpConstant,RegExpNormalChar |
-| test.cpp:159:24:159:24 | a | RegExpConstant,RegExpNormalChar |
-| test.cpp:159:24:159:25 | a+ | RegExpPlus |
-| test.cpp:159:24:159:26 | a+b | RegExpSequence |
-| test.cpp:159:26:159:26 | b | RegExpConstant,RegExpNormalChar |
-| test.cpp:162:30:162:30 | a | RegExpConstant,RegExpNormalChar |
-| test.cpp:162:30:162:31 | a+ | RegExpPlus |
-| test.cpp:162:30:162:32 | a+b | RegExpSequence |
-| test.cpp:162:32:162:32 | b | RegExpConstant,RegExpNormalChar |
-| test.cpp:166:22:166:23 | \\s | RegExpCharacterClassEscape |
-| test.cpp:166:22:166:24 | \\s+ | RegExpPlus |
-| test.cpp:169:22:169:22 | a | RegExpConstant,RegExpNormalChar |
-| test.cpp:169:22:169:25 | a\\.b | RegExpSequence |
-| test.cpp:169:23:169:24 | \\. | RegExpConstant,RegExpEscape |
-| test.cpp:169:25:169:25 | b | RegExpConstant,RegExpNormalChar |
+| test.cpp:147:24:147:25 | \\s | RegExpCharacterClassEscape |
+| test.cpp:147:24:147:26 | \\s+ | RegExpPlus |
+| test.cpp:147:24:147:27 | \\s+$ | RegExpSequence |
+| test.cpp:147:27:147:27 | $ | RegExpDollar |
+| test.cpp:150:24:150:25 | \\( | RegExpConstant,RegExpEscape |
+| test.cpp:150:24:150:37 | \\(([,\\w]+)+\\)$ | RegExpSequence |
+| test.cpp:150:26:150:33 | ([,\\w]+) | RegExpGroup |
+| test.cpp:150:26:150:34 | ([,\\w]+)+ | RegExpPlus |
+| test.cpp:150:27:150:31 | [,\\w] | RegExpCharacterClass |
+| test.cpp:150:27:150:32 | [,\\w]+ | RegExpPlus |
+| test.cpp:150:28:150:28 | , | RegExpConstant,RegExpNormalChar |
+| test.cpp:150:29:150:30 | \\w | RegExpCharacterClassEscape |
+| test.cpp:150:35:150:36 | \\) | RegExpConstant,RegExpEscape |
+| test.cpp:150:37:150:37 | $ | RegExpDollar |
+| test.cpp:153:25:153:25 | a | RegExpConstant,RegExpNormalChar |
+| test.cpp:153:25:153:26 | a+ | RegExpPlus |
+| test.cpp:153:25:153:27 | a+b | RegExpSequence |
+| test.cpp:153:27:153:27 | b | RegExpConstant,RegExpNormalChar |
+| test.cpp:156:24:156:24 | a | RegExpConstant,RegExpNormalChar |
+| test.cpp:156:24:156:25 | a+ | RegExpPlus |
+| test.cpp:156:24:156:26 | a+b | RegExpSequence |
+| test.cpp:156:26:156:26 | b | RegExpConstant,RegExpNormalChar |
+| test.cpp:159:30:159:30 | a | RegExpConstant,RegExpNormalChar |
+| test.cpp:159:30:159:31 | a+ | RegExpPlus |
+| test.cpp:159:30:159:32 | a+b | RegExpSequence |
+| test.cpp:159:32:159:32 | b | RegExpConstant,RegExpNormalChar |
+| test.cpp:163:22:163:23 | \\s | RegExpCharacterClassEscape |
+| test.cpp:163:22:163:24 | \\s+ | RegExpPlus |
+| test.cpp:166:22:166:22 | a | RegExpConstant,RegExpNormalChar |
+| test.cpp:166:22:166:25 | a\\.b | RegExpSequence |
+| test.cpp:166:23:166:24 | \\. | RegExpConstant,RegExpEscape |
+| test.cpp:166:25:166:25 | b | RegExpConstant,RegExpNormalChar |
 regExpNormalCharValue
 | test.cpp:33:21:33:23 | abc | abc |
 | test.cpp:36:22:36:22 | a | a |
@@ -312,29 +306,30 @@ regExpNormalCharValue
 | test.cpp:112:27:112:27 | b | b |
 | test.cpp:113:30:113:30 | q | q |
 | test.cpp:113:33:113:34 | \\s | s |
-| test.cpp:119:24:119:24 | a | a |
-| test.cpp:119:26:119:26 | f | f |
-| test.cpp:128:25:128:25 | A | A |
-| test.cpp:128:27:128:27 | F | F |
-| test.cpp:128:37:128:37 | a | a |
-| test.cpp:128:39:128:39 | f | f |
-| test.cpp:136:21:136:26 | \\u9879 | \u9879 |
+| test.cpp:116:24:116:24 | a | a |
+| test.cpp:116:26:116:26 | f | f |
+| test.cpp:116:27:116:28 | \\d | d |
+| test.cpp:125:25:125:25 | A | A |
+| test.cpp:125:27:125:27 | F | F |
+| test.cpp:125:37:125:37 | a | a |
+| test.cpp:125:39:125:39 | f | f |
+| test.cpp:133:21:133:26 | \\u9879 | \u9879 |
+| test.cpp:141:23:141:23 | a | a |
+| test.cpp:141:25:141:25 | b | b |
 | test.cpp:144:23:144:23 | a | a |
 | test.cpp:144:25:144:25 | b | b |
-| test.cpp:147:23:147:23 | a | a |
-| test.cpp:147:25:147:25 | b | b |
-| test.cpp:150:24:150:25 | \\s | s |
-| test.cpp:153:24:153:25 | \\( | ( |
-| test.cpp:153:28:153:28 | , | , |
-| test.cpp:153:29:153:30 | \\w | w |
-| test.cpp:153:35:153:36 | \\) | ) |
-| test.cpp:156:25:156:25 | a | a |
-| test.cpp:156:27:156:27 | b | b |
-| test.cpp:159:24:159:24 | a | a |
-| test.cpp:159:26:159:26 | b | b |
-| test.cpp:162:30:162:30 | a | a |
-| test.cpp:162:32:162:32 | b | b |
-| test.cpp:166:22:166:23 | \\s | s |
-| test.cpp:169:22:169:22 | a | a |
-| test.cpp:169:23:169:24 | \\. | . |
-| test.cpp:169:25:169:25 | b | b |
+| test.cpp:147:24:147:25 | \\s | s |
+| test.cpp:150:24:150:25 | \\( | ( |
+| test.cpp:150:28:150:28 | , | , |
+| test.cpp:150:29:150:30 | \\w | w |
+| test.cpp:150:35:150:36 | \\) | ) |
+| test.cpp:153:25:153:25 | a | a |
+| test.cpp:153:27:153:27 | b | b |
+| test.cpp:156:24:156:24 | a | a |
+| test.cpp:156:26:156:26 | b | b |
+| test.cpp:159:30:159:30 | a | a |
+| test.cpp:159:32:159:32 | b | b |
+| test.cpp:163:22:163:23 | \\s | s |
+| test.cpp:166:22:166:22 | a | a |
+| test.cpp:166:23:166:24 | \\. | . |
+| test.cpp:166:25:166:25 | b | b |
--- a/cpp/ql/test/library-tests/regex/test.cpp
+++ b/cpp/ql/test/library-tests/regex/test.cpp
@@ -112,11 +112,8 @@ void test() {
  std::regex r_bref1("(a+)b+\\1");
  std::regex r_bref2("(?<qux>q+)\\s+\\k<qux>+");

-  // Named character properties using the p-style syntax
-  std::regex r_prop1("\\p{Word}*");
-  std::regex r_prop2("\\P{Digit}+");
-  std::regex r_prop3("\\p{^Alnum}{2,3}");
-  std::regex r_prop4("[a-f\\p{Digit}]+"); // Also valid inside character classes
+  // A character class combining a range with an escape class
+  std::regex r_prop1("[a-f\\d]+");

  // Two separate character classes, each containing a single POSIX bracket expression
  std::regex r_posix1("[[:alpha:]][[:digit:]]");