From 004f9f9a36765fdcb92d6b83d4d69dccba165251 Mon Sep 17 00:00:00 2001
From: Konstantin Isakov <dragonroot@gmail.com>
Date: Tue, 26 Jul 2011 21:26:59 -0700
Subject: [PATCH] Don't attempt adding very large words into index, they are
 usually a sign of malformed dictionary and can consume huge amount of
 resources.

---
 btreeidx.cc | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/btreeidx.cc b/btreeidx.cc
index d9f5d9c8..8daebf1a 100644
--- a/btreeidx.cc
+++ b/btreeidx.cc
@@ -899,6 +899,11 @@ void IndexedWords::addWord( wstring const & word, uint32_t articleOffset )
   wchar const * wordBegin = word.c_str();
   string::size_type wordSize = word.size();
 
+  // Safeguard us against various bugs here. Don't attempt adding words
+  // which are freakishly huge.
+  if ( wordSize > 256 )
+    return;
+
   // Skip any leading whitespace
   while( *wordBegin && Folding::isWhitespace( *wordBegin ) )
   {