Merge pull request #2813 from str4d/2074-uacomment

Implement uacomment config parameter

src/init.cpp

@@ -1250,6 +1250,20 @@ bool AppInit2(boost::thread_group& threadGroup, CScheduler& scheduler)
 
     RegisterNodeSignals(GetNodeSignals());
 
+    // sanitize comments per BIP-0014, format user agent and check total size
+    std::vector<string> uacomments;
+    BOOST_FOREACH(string cmt, mapMultiArgs["-uacomment"])
+    {
+        if (cmt != SanitizeString(cmt, SAFE_CHARS_UA_COMMENT))
+            return InitError(strprintf("User Agent comment (%s) contains unsafe characters.", cmt));
+        uacomments.push_back(SanitizeString(cmt, SAFE_CHARS_UA_COMMENT));
+    }
+    strSubVersion = FormatSubVersion(CLIENT_NAME, CLIENT_VERSION, uacomments);
+    if (strSubVersion.size() > MAX_SUBVERSION_LENGTH) {
+        return InitError(strprintf("Total length of network version string %i exceeds maximum of %i characters. Reduce the number and/or size of uacomments.",
+            strSubVersion.size(), MAX_SUBVERSION_LENGTH));
+    }
+
     if (mapArgs.count("-onlynet")) {
         std::set<enum Network> nets;
         BOOST_FOREACH(const std::string& snet, mapMultiArgs["-onlynet"]) {

src/main.cpp

@@ -4820,7 +4820,7 @@ bool static ProcessMessage(CNode* pfrom, string strCommand, CDataStream& vRecv,
         if (!vRecv.empty())
             vRecv >> addrFrom >> nNonce;
         if (!vRecv.empty()) {
-            vRecv >> LIMITED_STRING(pfrom->strSubVer, 256);
+            vRecv >> LIMITED_STRING(pfrom->strSubVer, MAX_SUBVERSION_LENGTH);
             pfrom->cleanSubVer = SanitizeString(pfrom->strSubVer);
         }
         if (!vRecv.empty())

src/net.cpp

@@ -73,6 +73,7 @@ static std::vector<ListenSocket> vhListenSocket;
 CAddrMan addrman;
 int nMaxConnections = DEFAULT_MAX_PEER_CONNECTIONS;
 bool fAddressesInitialized = false;
+std::string strSubVersion;
 
 vector<CNode*> vNodes;
 CCriticalSection cs_vNodes;
@@ -435,7 +436,7 @@ void CNode::PushVersion()
     else
         LogPrint("net", "send version message: version %d, blocks=%d, us=%s, peer=%d\n", PROTOCOL_VERSION, nBestHeight, addrMe.ToString(), id);
     PushMessage("version", PROTOCOL_VERSION, nLocalServices, nTime, addrYou, addrMe,
-                nLocalHostNonce, FormatSubVersion(CLIENT_NAME, CLIENT_VERSION, std::vector<string>()), nBestHeight, true);
+                nLocalHostNonce, strSubVersion, nBestHeight, true);
 }
 
 

src/net.h

@@ -49,6 +49,8 @@ static const unsigned int MAX_INV_SZ = 50000;
 static const unsigned int MAX_ADDR_TO_SEND = 1000;
 /** Maximum length of incoming protocol messages (no message over 2 MiB is currently acceptable). */
 static const unsigned int MAX_PROTOCOL_MESSAGE_LENGTH = 2 * 1024 * 1024;
+/** Maximum length of strSubVer in `version` message */
+static const unsigned int MAX_SUBVERSION_LENGTH = 256;
 /** -listen default */
 static const bool DEFAULT_LISTEN = true;
 /** The maximum number of entries in mapAskFor */
@@ -156,6 +158,9 @@ extern CCriticalSection cs_vAddedNodes;
 extern NodeId nLastNodeId;
 extern CCriticalSection cs_nLastNodeId;
 
+/** Subversion as sent to the P2P network in `version` messages */
+extern std::string strSubVersion;
+
 struct LocalServiceInfo {
     int nScore;
     int nPort;

src/rpcnet.cpp

@@ -469,8 +469,7 @@ UniValue getnetworkinfo(const UniValue& params, bool fHelp)
 
     UniValue obj(UniValue::VOBJ);
     obj.push_back(Pair("version",       CLIENT_VERSION));
-    obj.push_back(Pair("subversion",
-        FormatSubVersion(CLIENT_NAME, CLIENT_VERSION, std::vector<string>())));
+    obj.push_back(Pair("subversion",    strSubVersion));
     obj.push_back(Pair("protocolversion",PROTOCOL_VERSION));
     obj.push_back(Pair("localservices",       strprintf("%016x", nLocalServices)));
     obj.push_back(Pair("timeoffset",    GetTimeOffset()));

src/test/util_tests.cpp

@@ -413,7 +413,7 @@ BOOST_AUTO_TEST_CASE(test_FormatSubVersion)
     comments.push_back(std::string("comment1"));
     std::vector<std::string> comments2;
     comments2.push_back(std::string("comment1"));
-    comments2.push_back(std::string("comment2"));
+    comments2.push_back(SanitizeString(std::string("Comment2; .,_?@; !\"#$%&'()*+-/<=>[]\\^`{|}~"), SAFE_CHARS_UA_COMMENT)); // Semicolon is discouraged but not forbidden by BIP-0014
     BOOST_CHECK_EQUAL(FormatSubVersion("Test", 99900, std::vector<std::string>()), std::string("/Test:0.9.99-beta1/"));
     BOOST_CHECK_EQUAL(FormatSubVersion("Test", 99924, std::vector<std::string>()), std::string("/Test:0.9.99-beta25/"));
     BOOST_CHECK_EQUAL(FormatSubVersion("Test", 99925, std::vector<std::string>()), std::string("/Test:0.9.99-rc1/"));
@@ -423,8 +423,8 @@ BOOST_AUTO_TEST_CASE(test_FormatSubVersion)
     BOOST_CHECK_EQUAL(FormatSubVersion("Test", 99999, std::vector<std::string>()), std::string("/Test:0.9.99-49/"));
     BOOST_CHECK_EQUAL(FormatSubVersion("Test", 99900, comments),  std::string("/Test:0.9.99-beta1(comment1)/"));
     BOOST_CHECK_EQUAL(FormatSubVersion("Test", 99950, comments),  std::string("/Test:0.9.99(comment1)/"));
-    BOOST_CHECK_EQUAL(FormatSubVersion("Test", 99900, comments2), std::string("/Test:0.9.99-beta1(comment1; comment2)/"));
-    BOOST_CHECK_EQUAL(FormatSubVersion("Test", 99950, comments2), std::string("/Test:0.9.99(comment1; comment2)/"));
+    BOOST_CHECK_EQUAL(FormatSubVersion("Test", 99900, comments2), std::string("/Test:0.9.99-beta1(comment1; Comment2; .,_?@; )/"));
+    BOOST_CHECK_EQUAL(FormatSubVersion("Test", 99950, comments2), std::string("/Test:0.9.99(comment1; Comment2; .,_?@; )/"));
 }
 
 BOOST_AUTO_TEST_CASE(test_ParseFixedPoint)

src/utilstrencodings.cpp

@@ -15,17 +15,20 @@
 
 using namespace std;
 
-string SanitizeString(const string& str)
+static const string CHARS_ALPHA_NUM = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789";
+
+static const string SAFE_CHARS[] =
+{
+    CHARS_ALPHA_NUM + " .,;_/:?@()", // SAFE_CHARS_DEFAULT
+    CHARS_ALPHA_NUM + " .,;_?@" // SAFE_CHARS_UA_COMMENT
+};
+
+string SanitizeString(const string& str, int rule)
 {
-    /**
-     * safeChars chosen to allow simple messages/URLs/email addresses, but avoid anything
-     * even possibly remotely dangerous like & or >
-     */
-    static string safeChars("abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ01234567890 .,;_/:?@()");
     string strResult;
     for (std::string::size_type i = 0; i < str.size(); i++)
     {
-        if (safeChars.find(str[i]) != std::string::npos)
+        if (SAFE_CHARS[rule].find(str[i]) != std::string::npos)
             strResult.push_back(str[i]);
     }
     return strResult;

src/utilstrencodings.h

@@ -22,8 +22,22 @@
 /** This is needed because the foreach macro can't get over the comma in pair<t1, t2> */
 #define PAIRTYPE(t1, t2)    std::pair<t1, t2>
 
+/** Used by SanitizeString() */
+enum SafeChars
+{
+    SAFE_CHARS_DEFAULT, //!< The full set of allowed chars
+    SAFE_CHARS_UA_COMMENT //!< BIP-0014 subset
+};
+
 std::string SanitizeFilename(const std::string& str);
-std::string SanitizeString(const std::string& str);
+/**
+* Remove unsafe chars. Safe chars chosen to allow simple messages/URLs/email
+* addresses, but avoid anything even possibly remotely dangerous like & or >
+* @param[in] str    The string to sanitize
+* @param[in] rule   The set of safe chars to choose (default: least restrictive)
+* @return           A new string without unsafe chars
+*/
+std::string SanitizeString(const std::string& str, int rule = SAFE_CHARS_DEFAULT);
 std::string HexInt(uint32_t val);
 uint32_t ParseHexToUInt32(const std::string& str);
 std::vector<unsigned char> ParseHex(const char* psz);