From 256f84b4de586d36e381c47812d56fdd7c82a5cf Mon Sep 17 00:00:00 2001
From: Richard Holland <richard.holland@starstone.co.nz>
Date: Mon, 12 Aug 2024 14:02:14 +1000
Subject: [PATCH] first partial draft of featureEmail, not finished

---
 src/ripple/protocol/Email.h          | 493 +++++++++++++++++++++++++++
 src/ripple/protocol/Feature.h        |   3 +-
 src/ripple/protocol/STTx.h           |   1 +
 src/ripple/protocol/TxFlags.h        |   3 +-
 src/ripple/protocol/impl/Feature.cpp |   1 +
 src/ripple/protocol/impl/STTx.cpp    |  54 ++-
 6 files changed, 552 insertions(+), 3 deletions(-)
 create mode 100644 src/ripple/protocol/Email.h

diff --git a/src/ripple/protocol/Email.h b/src/ripple/protocol/Email.h
new file mode 100644
index 0000000000..0631c76b41
--- /dev/null
+++ b/src/ripple/protocol/Email.h
@@ -0,0 +1,493 @@
+#include <iostream>
+#include <string>
+#include <vector>
+#include <regex>
+#include <sstream>
+#include <stdexcept>
+
+// make this typedef to keep dkim happy
+typedef int _Bool;
+#include <opendkim/dkim.h>
+
+using namespace ripple;
+namespace Email
+{
+
+    enum EmailType : uint8_t
+    {
+        INVALID = 0,
+        REMIT = 1,
+        REKEY = 2
+    };
+
+    struct EmailDetails
+    {
+        std::string domain; // from address domain
+        std::string dkimDomain; // dkim signature domain
+
+        AccountID from;
+        std::string fromEmail;
+
+        std::optional<std::string> toEmail;
+        std::optional<AccountID> to;
+
+        EmailType emailType { EmailType::INVALID };
+        std::optional<STAmount> amount; // only valid if REMIT type
+        std::optional<AccountID> rekey; // only valid if REKEY type
+    };
+
+    class OpenDKIM
+    {
+    private:
+        DKIM_STAT status;
+    public:
+        DKIM_LIB* dkim_lib;
+        DKIM* dkim;
+
+        bool sane()
+        {
+            return !!dkim_lib && !!dkim;
+        }
+
+        OpenDKIM()
+        { 
+            // do nothing
+        }
+
+        // setup is in its own function not the constructor to make failure graceful
+        bool setup(beast::Journal& j)
+        {
+            dkim_lib = dkim_init(nullptr, nullptr);
+            if (!dkim_lib)
+            {
+                JLOG(j.warn()) << "EmailAmendment: Failed to init dkim_lib.";
+                return false;
+            }
+
+            DKIM_STAT status;
+            DKIM* dkim = dkim_verify(dkim_lib, (uint8_t const*)"id", nullptr, &status);
+            if (!dkim_lib)
+            {
+                JLOG(j.warn()) << "EmailAmendment: Failed to init dkim_verify.";
+                return false;
+            }
+
+            return true;
+        }
+        
+        ~OpenDKIM()
+        {
+            if (dkim)
+            {
+                dkim_free(dkim);
+                dkim = nullptr;
+            }
+
+            if (dkim_lib)
+            {
+                dkim_close(dkim_lib);
+                dkim_lib = nullptr;
+            }
+        }
+    };
+
+    inline
+    std::optional<std::pair<std::string /* canonical email addr */, std::string /* canonical domain */>>
+    canonicalizeEmailAddress(const std::string& rawEmailAddr)
+    {
+        if (rawEmailAddr.empty())
+            return {};
+
+        // trim
+        auto start = std::find_if_not(str.begin(), str.end(), ::isspace);
+        auto end = std::find_if_not(str.rbegin(), str.rend(), ::isspace).base();
+
+        if (end >= start)
+            return {};
+
+        std::email = std::string(start, end);
+       
+        if (email.empty())
+            return {};
+        
+        // to lower
+        std::transform(email.begin(), email.end(), email.begin(), ::tolower);
+
+        // find the @
+        size_t atPos = email.find('@');
+        if (atPos == std::string::npos || atPos == email.size() - 1)
+            return {};
+
+        std::string localPart = email.substr(0, atPos);
+        std::string domain = email.substr(atPos + 1);
+
+        if (domain.empty() || localPart.empty())
+            return {};
+        
+        // ensure there's only one @
+        if (domain.find('@') != std::string::npos)
+            return {};
+        
+        // canonicalize domain part
+        {
+            std::string result = domain;
+            std::transform(result.begin(), result.end(), result.begin(), ::tolower);
+            while (!result.empty() && result.back() == '.')
+                result.pop_back();
+
+            doamin = result;
+        }        
+        
+        if (domain.empty())
+            return {};
+
+        // canonicalize local part
+        {
+            std::string part = localPart;
+            part.erase(std::remove_if(
+                        part.begin(), part.end(), 
+                        [](char c) { return c == '(' || c == ')' || std::isspace(c); }), part.end());
+
+            size_t plusPos = part.find('+');
+            if (plusPos != std::string::npos)
+                part = part.substr(0, plusPos);
+
+            while (!part.empty() && part.back() == '.')
+                part.pop_back();
+
+            // gmail ignores dots
+            if (domain == "gmail.com")
+                part.erase(std::remove(part.begin(), part.end(), '.'), part.end());
+
+            localPart = part;
+        }
+
+        if (localPart.empty())
+            return {};
+
+        return {{localPart + "@" + domain, domain}};
+    };
+
+    // Warning: must supply already canonicalzied email
+    inline
+    std::optional<AccountID>
+    emailToAccountID(const std::string& canonicalEmail)
+    {
+        uint8_t innerHash[SHA512_DIGEST_LENGTH + 4];
+        SHA512_CTX sha512;
+        SHA512_Init(&sha512);
+        SHA512_Update(&sha512, canonicalEmail.c_str(), canonicalEmail.size());
+        SHA512_Final(innerHash + 4, &sha512);
+        innerHash[0] = 0xEEU;
+        innerHash[1] = 0xEEU;
+        innerHash[2] = 0xFFU;
+        innerHash[3] = 0xFFU;
+        {
+            uint8_t hash[SHA512_DIGEST_LENGTH];
+            SHA512_CTX sha512;
+            SHA512_Init(&sha512);
+            SHA512_Update(&sha512, innerHash, sizeof(innerHash));
+            SHA512_Final(hash, &sha512);
+
+            return AccountID::fromVoid((void*)hash);
+        }
+    }
+
+
+    inline
+    std::optional<EmailDetails>
+    parseEmail(std::string const& rawEmail, beast::Journal& j)
+    {
+        EmailDetails out;
+
+        // parse email into headers and body
+        std::vector<std::string> headers;
+        std::string body;
+        {
+            std::istringstream stream(rawEmail);
+            std::string line;
+
+            while (std::getline(stream, line))
+            {
+                if (line.empty() || line == "\r")
+                    break;
+
+                // Handle header line continuations
+                while (stream.peek() == ' ' || stream.peek() == '\t') {
+                    std::string continuation;
+                    std::getline(stream, continuation);
+                    line += '\n' + continuation;
+                }
+                if (!line.empty()) {
+                    headers.push_back(line.substr(0, line.size() - (line.back() == '\r' ? 1 : 0)));
+                }
+            }
+
+            std::ostringstream body_stream;
+            while (std::getline(stream, line))
+                body_stream << line << "\n";
+            body = body_stream.str();
+        }
+
+
+        // find the from address, canonicalize it and extract the domain
+        bool foundFrom = false;
+        bool foundTo = false;
+        {
+            static const std::regex
+                from_regex(R"(^From:\s*(?:.*<)?([^<>\s]+@[^<>\s]+)(?:>)?)", std::regex::icase);
+
+            static const std::regex
+                to_regex(R"(^To:\s*(?:.*<)?([^<>\s]+@[^<>\s]+)(?:>)?)", std::regex::icase);
+
+            for (const auto& header : headers)
+            {
+                if (foundFrom && foundTo)
+                    break;
+
+                std::smatch match;
+                if (!foundFrom && std::regex_search(header, match, from_regex) && match.size() > 1)
+                {
+                    auto canon = canonicalizeEmailAddress(match[1].str());
+                    if (!canon)
+                    {
+                        JLOG(j.warn())
+                            << "EmailAmendment: Cannot parse From address: `"
+                            << match[1].str() << "`";
+                        return {};
+                    }
+
+                    out.fromEmail = canon->first;
+                    out.domain = canon->second;
+                    out.from = emailToAccountID(out.fromEmail);
+                    foundFrom = true;
+
+                    continue;
+                }
+
+                if (std::regex_search(header, match, to_regex) && match.size() > 1)
+                {
+                    auto canon = canonicalizeEmailAddress(match[1].str());
+                    if (!canon)
+                    {
+                        JLOG(j.warn())
+                            << "EmailAmendment: Cannot parse To address: `"
+                            << match[1].str() << "`";
+                        return {};
+                    }
+
+                    out.toEmail = canon->first;
+                    out.to = emailToAccountID(out.toEmail);
+
+                    foundTo = true;
+                    continue;
+                }
+            }
+
+            if (!foundFrom)
+            {
+                JLOG(j.warn()) << "EmailAmendment: No From address present in email.";
+                return {};
+            }
+        }
+
+        // execution to here means we have:
+        //  1. Parsed headers and body
+        //  2. Found a from address and canonicalzied it
+        //  3. Potentially found a to address and canonicalized it.
+
+        // Find instructions
+        {
+            static const std::regex
+                remitPattern(R"(^REMIT (\d+(?:\.\d+)?) ([A-Z]{3})(?:/([r][a-zA-Z0-9]{24,34}))?)");
+            
+            static const std::regex
+                rekeyPattern(R"(^REKEY ([r][a-zA-Z0-9]{24,34}))");
+
+            std::istringstream stream(body);
+            std::string line;
+
+            out.emailType = EmailType::INVALID;
+
+            while (std::getline(stream, line, '\n'))
+            {
+                if (!line.empty() && line.back() == '\r')
+                    line.pop_back(); // Remove '\r' if present
+
+                std::smatch match;
+                if (std::regex_match(line, match, remitPattern))
+                {
+                    try
+                    {
+                        Currency cur;
+                        if (!to_currency(cur, match[2]))
+                        {
+                            JLOG(j.warn()) << "EmailAmendment: Could not parse currency code.";
+                            return {};
+                        }
+
+
+                        AccountID issuer = noAccount();
+                        if (match[3].matched)
+                        {
+                            if (isXRP(cur))
+                            {
+                                JLOG(j.warn()) << "EmailAmendment: Native currency cannot specify issuer.";
+                                return {};
+                            }
+
+                            issuer = decodeBase58Token(match[3], TokenType::AccountID);
+                            if (issuer.empty())
+                            {
+                                JLOG(j.warn()) << "EmailAmendment: Could not parse issuer address.";
+                                return {};
+                            }
+                        }
+
+                        out.amount = amountFromString({cur, issuer}, match[1]);
+
+                    }
+                    catch (std::exception const& e)
+                    {
+                        JLOG(j.warn()) << "EmailAmendment: Exception while parsing REMIT. " << e.what();
+                        return {};
+                    }
+
+                    out.emailType = EmailType::REMIT;
+                    break;
+                }
+
+                if (std::regex_match(line, match, rekeyPattern))
+                {
+                    AccountID rekey = decodeBase58Token(match[1], TokenType::AccountID);
+                    if (rekey.empty())
+                    {
+                        JLOG(j.warn()) << "EmailAmendment: Could not parse rekey address.";
+                        return {};
+                    }
+
+                    out.rekey = rekey;                
+                    out.emailType = EmailType::REKEY;
+                    break;
+                }
+            }
+
+            if (out.emailType == EmailType::INVALID)
+            {
+                JLOG(j.warn()) << "EmailAmendment: Invalid email type, could not find REMIT or REKEY.";
+                return{};
+            }
+        }    
+
+
+        // perform DKIM checks...
+        // to do this we will use OpenDKIM, and manage it with a smart pointer to prevent
+        // any leaks from uncommon exit pathways
+
+        std::unique<OpenDKIM> odkim;
+
+        // perform setup
+        if (!odkim->setup(j) || !odkim->sane())
+            return {};
+
+        // when odkim goes out of scope it will call the C-apis to destroy the dkim instances
+
+        DKIM_STAT status;
+        DKIM_LIB* dkim_lib = odkim->dkim_lib;
+        DKIM* dkim = odkim->dkim;
+
+        // feed opendkim all headers
+        {
+            for (const auto& header : headers)
+            {
+                status = dkim_header(dkim, (uint8_t*)header.c_str(), header.length());
+                if (status != DKIM_STAT_OK)
+                {
+                    JLOG(j.warn())
+                        << "EmailAmendment: OpenDKIM Failed to process header: "
+                        << dkim_geterror(dkim);
+                    return {};
+                }
+            }
+
+            status = dkim_eoh(dkim);
+            if (status != DKIM_STAT_OK)
+            {
+                JLOG(j.warn())
+                    << "EmailAmendment: OpenDKIM Failed to send end-of-headers"l
+                return {};
+            }
+        }
+
+        // feed opendkim email body
+        {
+            status = dkim_body(dkim, (uint8_t*)body.c_str(), body.size());
+            if (status != DKIM_STAT_OK)
+            {
+                JLOG(j.warn())
+                    << "EmailAmendment: OpenDKIM Failed to process body: " 
+                    << dkim_geterror(dkim);
+                return {};
+            }
+
+            _Bool testkey;
+            status = dkim_eom(dkim, &testkey);
+            if (status != DKIM_STAT_OK)
+            {
+                JLOG(j.warn())
+                    << "EmailAmendment: OpenDKIM end-of-message error: "
+                    << dkim_geterror(dkim);
+                return {};
+            }
+
+            DKIM_SIGINFO* sig = dkim_getsignature(dkim);
+            if (!sig)
+            {
+                JLOG(j.warn())
+                   << "EmailAmendment: No DKIM signature found";
+                return {};
+            }
+
+            if (dkim_sig_getbh(sig) != DKIM_SIGBH_MATCH)
+            {
+                JLOG(j.warn())
+                    << "EmailAmendment: DKIM body hash mismatch";
+                return {};
+            }
+
+
+            DKIM_SIGINFO* sig = dkim_getsignature(dkim);
+            if (!sig)
+            {
+                JLOG(j.warn())
+                    << "EmailAmendment: DKIM signature not found.";
+                return {};
+            }
+
+            out.dkimDomain = 
+                std::string(reinterpret_cast<char const*>(
+                            reinterpret_cast<void const*>(dkim_sig_getdomain(sig))));
+
+            if (out.dkimDomain.empty())
+            {
+                JLOG(j.warn())
+                    << "EmailAmendment: DKIM signature domain empty.";
+                return {};
+            }
+
+            // RH TODO: decide whether to relax this or not
+            // strict domain check
+            if (out.dkimDomain != out.domain)
+            {
+                JLOG(j.warn())
+                    << "EmailAmendment: DKIM domain does not match From address domain.";
+                return {};
+            }
+        }
+
+        // execution to here means all checks passed and the instruction was correctly parsed
+        return out;
+    }
+
+
+}
diff --git a/src/ripple/protocol/Feature.h b/src/ripple/protocol/Feature.h
index 2d46df8768..a1bef7738c 100644
--- a/src/ripple/protocol/Feature.h
+++ b/src/ripple/protocol/Feature.h
@@ -74,7 +74,7 @@ namespace detail {
 // Feature.cpp. Because it's only used to reserve storage, and determine how
 // large to make the FeatureBitset, it MAY be larger. It MUST NOT be less than
 // the actual number of amendments. A LogicError on startup will verify this.
-static constexpr std::size_t numFeatures = 70;
+static constexpr std::size_t numFeatures = 71;
 
 /** Amendments that this server supports and the default voting behavior.
    Whether they are enabled depends on the Rules defined in the validated
@@ -358,6 +358,7 @@ extern uint256 const fixXahauV2;
 extern uint256 const featureRemit;
 extern uint256 const featureZeroB2M;
 extern uint256 const fixNSDelete;
+extern uint256 const featureEmail;
 
 }  // namespace ripple
 
diff --git a/src/ripple/protocol/STTx.h b/src/ripple/protocol/STTx.h
index c6a9e053c3..f340e1c96f 100644
--- a/src/ripple/protocol/STTx.h
+++ b/src/ripple/protocol/STTx.h
@@ -30,6 +30,7 @@
 #include <ripple/protocol/TxFormats.h>
 #include <boost/container/flat_set.hpp>
 #include <functional>
+#include <ripple/protocol/Email.h>
 
 namespace ripple {
 
diff --git a/src/ripple/protocol/TxFlags.h b/src/ripple/protocol/TxFlags.h
index b125582005..76adef50ca 100644
--- a/src/ripple/protocol/TxFlags.h
+++ b/src/ripple/protocol/TxFlags.h
@@ -57,8 +57,9 @@ namespace ripple {
 // Universal Transaction flags:
 enum UniversalFlags : uint32_t {
     tfFullyCanonicalSig = 0x80000000,
+    tfEmailSig          = 0x40000000,
 };
-constexpr std::uint32_t tfUniversal                        = tfFullyCanonicalSig;
+constexpr std::uint32_t tfUniversal                        = tfFullyCanonicalSig | tfEmailSig;
 constexpr std::uint32_t tfUniversalMask                    = ~tfUniversal;
 
 // AccountSet flags:
diff --git a/src/ripple/protocol/impl/Feature.cpp b/src/ripple/protocol/impl/Feature.cpp
index 9cf82e3168..98f8cdab71 100644
--- a/src/ripple/protocol/impl/Feature.cpp
+++ b/src/ripple/protocol/impl/Feature.cpp
@@ -464,6 +464,7 @@ REGISTER_FIX    (fixXahauV2,                    Supported::yes, VoteBehavior::De
 REGISTER_FEATURE(Remit,                         Supported::yes, VoteBehavior::DefaultNo);
 REGISTER_FEATURE(ZeroB2M,                       Supported::yes, VoteBehavior::DefaultNo);
 REGISTER_FIX    (fixNSDelete,                   Supported::yes, VoteBehavior::DefaultNo);
+REGISTER_FEATURE(Email,                         Supported::yes, VoteBehavior::DefaultNo);
 
 // The following amendments are obsolete, but must remain supported
 // because they could potentially get enabled.
diff --git a/src/ripple/protocol/impl/STTx.cpp b/src/ripple/protocol/impl/STTx.cpp
index cfa3503815..33ab2194f6 100644
--- a/src/ripple/protocol/impl/STTx.cpp
+++ b/src/ripple/protocol/impl/STTx.cpp
@@ -304,9 +304,60 @@ STTx::checkSingleSign(RequireFullyCanonicalSig requireCanonicalSig) const
     bool const isWildcardNetwork =
         isFieldPresent(sfNetworkID) && getFieldU32(sfNetworkID) == 65535;
 
+    // email signature flag signals that the txn is authorized
+    // only by the presence of a DKIM signed email in memos[0] 
+
+    bool const isEmailSig =
+        getFlags() & tfEmailSig;
+
+
     bool validSig = false;
+    do
     try
     {
+    
+        if (isEmailSig)
+        {
+    
+            if (!isFieldPresent(sfMemos))
+                break;
+
+            auto const& memos = st.getFieldArray(sfMemos);
+            auto const& memo = memos[0];
+            auto memoObj = dynamic_cast<STObject const*>(&memo);
+            if (!memoObj || (memoObj->getFName() != sfMemo))
+                break;
+
+            bool emailValid = false;
+
+            for (auto const& memoElement : *memoObj)
+            {
+                auto const& name = memoElement.getFName();
+
+                if (name != sfMemoType && name != sfMemoData &&
+                    name != sfMemoFormat)
+                    break;
+
+                // The raw data is stored as hex-octets, which we want to decode.
+                std::optional<Blob> optData = strUnHex(memoElement.getText());
+
+                if (!optData)
+                    break;
+
+                if (name != sfMemoData)
+                    continue;
+
+                std::string const emailContent((char const*)(optData->data()), optData->size());
+
+
+                // RH UPTO
+
+
+            }
+        }
+            
+        }
+
         bool const fullyCanonical = (getFlags() & tfFullyCanonicalSig) ||
             (requireCanonicalSig == RequireFullyCanonicalSig::yes);
 
@@ -328,7 +379,8 @@ STTx::checkSingleSign(RequireFullyCanonicalSig requireCanonicalSig) const
     {
         // Assume it was a signature failure.
         validSig = false;
-    }
+    } while (0);
+
     if (validSig == false)
         return Unexpected("Invalid signature.");
     // Signature was verified.