From f6c10c72ee9cc8ecbdc69c671e1a673dc686bb49 Mon Sep 17 00:00:00 2001
From: James Ring <sjr@jdns.org>
Date: Tue, 23 Oct 2018 08:17:12 -0700
Subject: [PATCH] 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.
---
 java/com/google/re2j/Parser.java              | 4 ++++
 javatests/com/google/re2j/RE2CompileTest.java | 2 ++
 2 files changed, 6 insertions(+)

diff --git a/java/com/google/re2j/Parser.java b/java/com/google/re2j/Parser.java
index d1870c69..2b965052 100644
--- 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
index 3fc7f0dd..fb31f982 100644
--- 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{`"}
     };
   }