Handle bad host names

* Detect bad characters in host names.
* Incorporate splitting host and port into the state
  machine that is parsing/decoding the host.

NOTE:
    IPv6address is not checked for bad characters yet.
    More research is needed to learn exactly what are
    the various ways to write an IPv6 address.

1 files changed, 141 insertions, 5 deletions

diff --git a/src/Uri.cpp b/src/Uri.cpp
index e218ecd..ba04298 100644
--- a/src/Uri.cpp
+++ b/src/Uri.cpp
@@ -305,13 +305,149 @@ namespace Uri {
             }
 
             // Next, parsing host and port from authority and path.
-            const auto portDelimiter = hostPortString.find(':');
-            if (portDelimiter == std::string::npos) {
-                host = hostPortString;
+            std::string portString;
+            size_t decoderState = 0;
+            int decodedCharacter = 0;
+            host.clear();
+            for (const auto c: hostPortString) {
+                switch(decoderState) {
+                    case 0: { // first character
+                        if (c == '[') {
+                            host.push_back(c);
+                            decoderState = 4;
+                            break;
+                        } else {
+                            decoderState = 1;
+                        }
+                    }
+
+                    case 1: { // reg-name or IPv4Address
+                        if (c == '%') {
+                            decoderState = 2;
+                        } else if (c == ':') {
+                            decoderState = 9;
+                        } else {
+                            if (
+                                IsCharacterInSet(
+                                    c,
+                                    {
+                                        // unreserved
+                                        'a','z', 'A','Z', // ALPHA
+                                        '0','9', // DIGIT
+                                        '-','-', '.','.', '_','_', '~','~',
+
+                                        // sub-delims
+                                        '!','!', '$','$', '&','&', '\'','\'', '(','(', ')',')',
+                                        '*','*', '+','+', ',',',', ';',';', '=','=',
+
+                                        // (also allowed in reg-name)
+                                        ':',':',
+                                    }
+                                )
+                            ) {
+                                host.push_back(c);
+                            } else {
+                                return false;
+                            }
+                        }
+                    } break;
+
+                    case 2: { // % ...
+                        decoderState = 3;
+                        decodedCharacter <<= 4;
+                        if (IsCharacterInSet(c, {'0','9'})) {
+                            decodedCharacter += (int)(c - '0');
+                        } else if (IsCharacterInSet(c, {'A','F'})) {
+                            decodedCharacter += (int)(c - 'A') + 10;
+                        } else {
+                            return false;
+                        }
+                    } break;
+
+                    case 3: { // %[0-9A-F] ...
+                        decoderState = 1;
+                        decodedCharacter <<= 4;
+                        if (IsCharacterInSet(c, {'0','9'})) {
+                            decodedCharacter += (int)(c - '0');
+                        } else if (IsCharacterInSet(c, {'A','F'})) {
+                            decodedCharacter += (int)(c - 'A') + 10;
+                        } else {
+                            return false;
+                        }
+                        host.push_back((char)decodedCharacter);
+                    } break;
+
+                    case 4: { // IP-literal
+                        if (c == 'v') {
+                            host.push_back(c);
+                            decoderState = 6;
+                            break;
+                        } else {
+                            decoderState = 5;
+                        }
+                    }
+
+                    case 5: { // IPv6Address
+                        // TODO: research this offline first
+                        // before attempting to code it
+                        host.push_back(c);
+                        if (c == ']') {
+                            decoderState = 8;
+                        }
+                    } break;
+
+                    case 6: { // IPvFuture: v ...
+                        if (c == '.') {
+                            decoderState = 7;
+                        } else if (!IsCharacterInSet(c, {'0','9', 'A','F'})) {
+                            return false;
+                        }
+                        host.push_back(c);
+                    } break;
+
+                    case 7: { // IPvFuture v 1*HEXDIG . ...
+                        host.push_back(c);
+                        if (c == ']') {
+                            decoderState = 8;
+                        } else if (
+                            !IsCharacterInSet(
+                                c,
+                                {
+                                    // unreserved
+                                    'a','z', 'A','Z', // ALPHA
+                                    '0','9', // DIGIT
+                                    '-','-', '.','.', '_','_', '~','~',
+
+                                    // sub-delims
+                                    '!','!', '$','$', '&','&', '\'','\'', '(','(', ')',')',
+                                    '*','*', '+','+', ',',',', ';',';', '=','=',
+
+                                    // (also allowed in IPvFuture)
+                                    ':',':',
+                                }
+                            )
+                        ) {
+                            return false;
+                        }
+                    } break;
+
+                    case 8: { // illegal to have anything else, unless it's a colon,
+                              // in which case it's a port delimiter
+                        if (c == ':') {
+                            decoderState = 9;
+                        } else {
+                            return false;
+                        }
+                    } break;
+
+                    case 9: { // port
+                        portString.push_back(c);
+                    } break;
+                }
+            }
+            if (portString.empty()) {
                 hasPort = false;
             } else {
-                host = hostPortString.substr(0, portDelimiter);
-                const auto portString = hostPortString.substr(portDelimiter + 1);
                 if (!ParseUint16(portString, port)) {
                     return false;
                 }