Detect invalid unicode class specification (#74)

Previously, the parser would unconditionally attempt to read the unicode class spec after a \p or \P, even if the pattern is truncated and invalid. Fixes issue #72.
google · Oct 23, 2018 · f6c10c7 · f6c10c7
1 parent 06d4a72
commit f6c10c7
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 0 deletions.
diff --git a/java/com/google/re2j/Parser.java b/java/com/google/re2j/Parser.java
@@ -1539,6 +1539,10 @@ private boolean parseUnicodeClass(StringIterator t, CharClass cc) throws Pattern
     if (c == 'P') {
       sign = -1;
     }
+    if (!t.more()) {
+      t.rewindTo(startPos);
+      throw new PatternSyntaxException(ERR_INVALID_CHAR_RANGE, t.rest());
+    }
     c = t.pop();
     String name;
     if (c != '{') {

diff --git a/javatests/com/google/re2j/RE2CompileTest.java b/javatests/com/google/re2j/RE2CompileTest.java
@@ -44,6 +44,8 @@ public static String[][] testData() {
       {"a**", "invalid nested repetition operator: `**`"},
       {"a*+", "invalid nested repetition operator: `*+`"},
       {"\\x", "invalid escape sequence: `\\x`"},
+      {"\\p", "invalid character class range: `\\p`"},
+      {"\\p{", "invalid character class range: `\\p{`"}
     };
   }