2010-03-03 Tatsuhiro Tsujikawa <t-tujikawa@users.sourceforge.net>

	In Metalink4, if size element contains invalid size, discard whole
	document. Added strict hash value check for metalink3/4.
	* src/MetalinkParserController.cc
	* src/MetalinkParserStateMachine.cc
	* src/MetalinkParserStateMachine.h
	* src/MetalinkParserStateV3Impl.cc
	* src/MetalinkParserStateV4Impl.cc
	* src/util.cc
	* src/util.h
	* test/MetalinkParserControllerTest.cc
	* test/MetalinkProcessorTest.cc
	* test/metalink4.xml

7 files changed, 53 insertions, 8 deletions

diff --git a/src/MetalinkParserController.cc b/src/MetalinkParserController.cc
index d3045b97..b340af49 100644
--- a/src/MetalinkParserController.cc
+++ b/src/MetalinkParserController.cc
@@ -53,6 +53,13 @@
 
 namespace aria2 {
 
+static bool isValidHash(const std::string& algo, const std::string& hash)
+{
+  return util::isHexDigit(hash) &&
+    MessageDigestContext::supports(algo) &&
+    MessageDigestContext::digestLength(algo)*2 == hash.size();
+}
+
 MetalinkParserController::MetalinkParserController():
   _metalinker(new Metalinker())
 {}
@@ -281,7 +288,11 @@ void MetalinkParserController::setHashOfChecksum(const std::string& md)
   if(_tChecksum.isNull()) {
     return;
   }
-  _tChecksum->setMessageDigest(md);
+  if(isValidHash(_tChecksum->getAlgo(), md)) {
+    _tChecksum->setMessageDigest(md);
+  } else {
+    cancelChecksumTransaction();
+  }
 #endif // ENABLE_MESSAGE_DIGEST
 }
 
@@ -352,7 +363,11 @@ void MetalinkParserController::addHashOfChunkChecksumV4(const std::string& md)
   if(_tChunkChecksumV4.isNull()) {
     return;
   }
-  _tempChunkChecksumsV4.push_back(md);
+  if(isValidHash(_tChunkChecksumV4->getAlgo(), md)) {
+    _tempChunkChecksumsV4.push_back(md);
+  } else {
+    cancelChunkChecksumTransactionV4();
+  }
 #endif // ENABLE_MESSAGE_DIGEST  
 }
 
@@ -426,7 +441,11 @@ void MetalinkParserController::addHashOfChunkChecksum(size_t order, const std::s
   if(_tChunkChecksum.isNull()) {
     return;
   }
-  _tempChunkChecksums.push_back(std::pair<size_t, std::string>(order, md));
+  if(isValidHash(_tChunkChecksum->getAlgo(), md)) {
+    _tempChunkChecksums.push_back(std::make_pair(order, md));
+  } else {
+    cancelChunkChecksumTransaction();
+  }
 #endif // ENABLE_MESSAGE_DIGEST
 }
 
@@ -446,7 +465,11 @@ void MetalinkParserController::setMessageDigestOfChunkChecksum(const std::string
   if(_tChunkChecksum.isNull()) {
     return;
   }
-  _tempHashPair.second = md;
+  if(isValidHash(_tChunkChecksum->getAlgo(), md)) {
+    _tempHashPair.second = md;
+  } else {
+    cancelChunkChecksumTransaction();
+  }
 #endif // ENABLE_MESSAGE_DIGEST
 }
 
diff --git a/src/MetalinkParserStateMachine.cc b/src/MetalinkParserStateMachine.cc
index 2f114d4a..5d28ef3c 100644
--- a/src/MetalinkParserStateMachine.cc
+++ b/src/MetalinkParserStateMachine.cc
@@ -290,6 +290,11 @@ void MetalinkParserStateMachine::commitEntryTransaction()
   _ctrl->commitEntryTransaction();
 }
 
+void MetalinkParserStateMachine::cancelEntryTransaction()
+{
+  _ctrl->cancelEntryTransaction();
+}
+
 void MetalinkParserStateMachine::newResourceTransaction()
 {
   _ctrl->newResourceTransaction();
diff --git a/src/MetalinkParserStateMachine.h b/src/MetalinkParserStateMachine.h
index 472b05da..1eea88b5 100644
--- a/src/MetalinkParserStateMachine.h
+++ b/src/MetalinkParserStateMachine.h
@@ -166,6 +166,8 @@ public:
 
   void commitEntryTransaction();
 
+  void cancelEntryTransaction();
+
   void newResourceTransaction();
 
   void setURLOfResource(const std::string& url);
diff --git a/src/MetalinkParserStateV3Impl.cc b/src/MetalinkParserStateV3Impl.cc
index c4068836..98a931dd 100644
--- a/src/MetalinkParserStateV3Impl.cc
+++ b/src/MetalinkParserStateV3Impl.cc
@@ -183,7 +183,7 @@ void SizeMetalinkParserState::endElement
  const std::string& characters)
 {
   try {
-    stm->setFileLengthOfEntry(util::parseLLInt(characters));
+    stm->setFileLengthOfEntry(util::parseULLInt(characters));
   } catch(RecoverableException& e) {
     // current metalink specification doesn't require size element.
   }
diff --git a/src/MetalinkParserStateV4Impl.cc b/src/MetalinkParserStateV4Impl.cc
index d95b2772..fc81e89a 100644
--- a/src/MetalinkParserStateV4Impl.cc
+++ b/src/MetalinkParserStateV4Impl.cc
@@ -281,9 +281,10 @@ void SizeMetalinkParserStateV4::endElement
  const std::string& characters)
 {
   try {
-    stm->setFileLengthOfEntry(util::parseLLInt(characters));
+    stm->setFileLengthOfEntry(util::parseULLInt(characters));
   } catch(RecoverableException& e) {
-    // current metalink specification doesn't require size element.
+    stm->cancelEntryTransaction();
+    stm->logError("Bad size");
   }
 }
 
diff --git a/src/util.cc b/src/util.cc
index 5b6b6832..01e4bb96 100644
--- a/src/util.cc
+++ b/src/util.cc
@@ -216,6 +216,16 @@ bool isHexDigit(const char c)
   return isDigit(c) || ('A' <= c && c <= 'F') || ('a' <= c && c <= 'f');
 }
 
+bool isHexDigit(const std::string& s)
+{
+  for(std::string::const_iterator i = s.begin(), eoi = s.end(); i != eoi; ++i) {
+    if(!isHexDigit(*i)) {
+      return false;
+    }
+  }
+  return true;
+}
+
 bool inRFC3986ReservedChars(const char c)
 {
   static const char reserved[] = {
diff --git a/src/util.h b/src/util.h
index d1985209..4aefe6b0 100644
--- a/src/util.h
+++ b/src/util.h
@@ -254,11 +254,15 @@ void sleep(long seconds);
 void usleep(long microseconds);
   
 bool isNumber(const std::string& what);
+
+bool isHexDigit(const char c);
+
+bool isHexDigit(const std::string& s);
   
 bool isLowercase(const std::string& what);
   
 bool isUppercase(const std::string& what);
-  
+
 /**
  * Converts alphabets to unsigned int, assuming alphabets as a base 26
  * integer and 'a' or 'A' is 0.