Merge bitcoin/bitcoin#25001: Modernize util/strencodings and util/string: string_view and optional

fa7078d84f scripted-diff: Rename ValidAsCString to ContainsNoNUL (MacroFake)
e7d2fbda63 Use std::string_view throughout util strencodings/string (Pieter Wuille)
8ffbd1412d Make DecodeBase{32,64} take string_view arguments (Pieter Wuille)
1a72d62152 Generalize ConvertBits to permit transforming the input (Pieter Wuille)
78f3ac51b7 Make DecodeBase{32,64} return optional instead of taking bool* (Pieter Wuille)
a65931e3ce Make DecodeBase{32,64} always return vector, not string (Pieter Wuille)
a4377a0843 Reject incorrect base64 in HTTP auth (Pieter Wuille)
d648b5120b Make SanitizeString use string_view (Pieter Wuille)
963bc9b576 Make IsHexNumber use string_view (Pieter Wuille)
40062997f2 Make IsHex use string_view (Pieter Wuille)
c1d165a8c2 Make ParseHex use string_view (Pieter Wuille)

Pull request description:

  Make use of `std::string_view` and `std::optional` in the util/{strencodings, string} files.

  This avoids many temporary string/vector objects being created, while making the interface easier to read. Changes include:
  * Make all input arguments in functions in util/strencodings and util/string take `std::string_view` instead of `std::string`.
  * Add `RemovePrefixView` and `TrimStringView` which also *return* `std::string_view` objects (the corresponding `RemovePrefix` and `TrimString` keep returning an `std::string`, as that's needed in many call sites still).
  * Stop returning `std::string` from `DecodeBase32` and `DecodeBase64`, but return vectors. Base32/64 are fundamentally algorithms for encoding bytes as strings; returning `std::string` from those (especially doing it conditionally based on the input arguments/types) is just bizarre.
  * Stop taking a `bool* pf_invalid` output argument pointer in `DecodeBase32` and `DecodeBase64`; return an `std::optional` instead.
  * Make `DecodeBase32` and `DecodeBase64` more efficient by doing the conversion from characters to integer symbols on-the-fly rather than through a temporary vector.

ACKs for top commit:
  MarcoFalke:
    re-ACK fa7078d84f only change is rebase and adding a scripted-diff 🍲
  martinus:
    Code review ACK fa7078d84f, found no issue
  laanwj:
    Code review ACK  fa7078d84f
  sipa:
    utACK fa7078d84f (as far as the commit that isn't mine goes)

Tree-SHA512: 5cf02e541caef0bcd100466747664bdb828a68a05dae568cbcd0632a53dd3a4c4e85cd8c48ebbd168d4247d5c9666689c16005f1c8ad75b0f057d8683931f664

src/base58.cpp

@@ -126,7 +126,7 @@ std::string EncodeBase58(Span<const unsigned char> input)
 
 bool DecodeBase58(const std::string& str, std::vector<unsigned char>& vchRet, int max_ret_len)
 {
-    if (!ValidAsCString(str)) {
+    if (!ContainsNoNUL(str)) {
         return false;
     }
     return DecodeBase58(str.c_str(), vchRet, max_ret_len);
@@ -160,7 +160,7 @@ std::string EncodeBase58Check(Span<const unsigned char> input)
 
 bool DecodeBase58Check(const std::string& str, std::vector<unsigned char>& vchRet, int max_ret)
 {
-    if (!ValidAsCString(str)) {
+    if (!ContainsNoNUL(str)) {
         return false;
     }
     return DecodeBase58Check(str.c_str(), vchRet, max_ret);

src/bitcoin-tx.cpp

@@ -240,7 +240,7 @@ static void MutateTxRBFOptIn(CMutableTransaction& tx, const std::string& strInId
 template <typename T>
 static T TrimAndParse(const std::string& int_str, const std::string& err)
 {
-    const auto parsed{ToIntegral<T>(TrimString(int_str))};
+    const auto parsed{ToIntegral<T>(TrimStringView(int_str))};
     if (!parsed.has_value()) {
         throw std::runtime_error(err + " '" + int_str + "'");
     }

src/httprpc.cpp

@@ -131,8 +131,11 @@ static bool RPCAuthorized(const std::string& strAuth, std::string& strAuthUserna
         return false;
     if (strAuth.substr(0, 6) != "Basic ")
         return false;
-    std::string strUserPass64 = TrimString(strAuth.substr(6));
+    std::string_view strUserPass64 = TrimStringView(std::string_view{strAuth}.substr(6));
-    std::string strUserPass = DecodeBase64(strUserPass64);
+    auto userpass_data = DecodeBase64(strUserPass64);
+    std::string strUserPass;
+    if (!userpass_data) return false;
+    strUserPass.assign(userpass_data->begin(), userpass_data->end());
 
     if (strUserPass.find(':') != std::string::npos)
         strAuthUsernameOut = strUserPass.substr(0, strUserPass.find(':'));

src/i2p.cpp

@@ -69,12 +69,11 @@ static std::string SwapBase64(const std::string& from)
 static Binary DecodeI2PBase64(const std::string& i2p_b64)
 {
     const std::string& std_b64 = SwapBase64(i2p_b64);
-    bool invalid;
+    auto decoded = DecodeBase64(std_b64);
-    Binary decoded = DecodeBase64(std_b64.c_str(), &invalid);
+    if (!decoded) {
-    if (invalid) {
         throw std::runtime_error(strprintf("Cannot decode Base64: \"%s\"", i2p_b64));
     }
-    return decoded;
+    return std::move(*decoded);
 }
 
 /**

src/netaddress.cpp

@@ -210,7 +210,7 @@ static void Checksum(Span<const uint8_t> addr_pubkey, uint8_t (&checksum)[CHECKS
 
 bool CNetAddr::SetSpecial(const std::string& addr)
 {
-    if (!ValidAsCString(addr)) {
+    if (!ContainsNoNUL(addr)) {
         return false;
     }
 
@@ -234,17 +234,16 @@ bool CNetAddr::SetTor(const std::string& addr)
         return false;
     }
 
-    bool invalid;
+    auto input = DecodeBase32(std::string_view{addr}.substr(0, addr.size() - suffix_len));
-    const auto& input = DecodeBase32(addr.substr(0, addr.size() - suffix_len).c_str(), &invalid);
 
-    if (invalid) {
+    if (!input) {
         return false;
     }
 
-    if (input.size() == torv3::TOTAL_LEN) {
+    if (input->size() == torv3::TOTAL_LEN) {
-        Span<const uint8_t> input_pubkey{input.data(), ADDR_TORV3_SIZE};
+        Span<const uint8_t> input_pubkey{input->data(), ADDR_TORV3_SIZE};
-        Span<const uint8_t> input_checksum{input.data() + ADDR_TORV3_SIZE, torv3::CHECKSUM_LEN};
+        Span<const uint8_t> input_checksum{input->data() + ADDR_TORV3_SIZE, torv3::CHECKSUM_LEN};
-        Span<const uint8_t> input_version{input.data() + ADDR_TORV3_SIZE + torv3::CHECKSUM_LEN, sizeof(torv3::VERSION)};
+        Span<const uint8_t> input_version{input->data() + ADDR_TORV3_SIZE + torv3::CHECKSUM_LEN, sizeof(torv3::VERSION)};
 
         if (input_version != torv3::VERSION) {
             return false;
@@ -280,15 +279,14 @@ bool CNetAddr::SetI2P(const std::string& addr)
     // can decode it.
     const std::string b32_padded = addr.substr(0, b32_len) + "====";
 
-    bool invalid;
+    auto address_bytes = DecodeBase32(b32_padded);
-    const auto& address_bytes = DecodeBase32(b32_padded.c_str(), &invalid);
 
-    if (invalid || address_bytes.size() != ADDR_I2P_SIZE) {
+    if (!address_bytes || address_bytes->size() != ADDR_I2P_SIZE) {
         return false;
     }
 
     m_net = NET_I2P;
-    m_addr.assign(address_bytes.begin(), address_bytes.end());
+    m_addr.assign(address_bytes->begin(), address_bytes->end());
 
     return true;
 }

src/netbase.cpp

@@ -136,7 +136,7 @@ static bool LookupIntern(const std::string& name, std::vector<CNetAddr>& vIP, un
 {
     vIP.clear();
 
-    if (!ValidAsCString(name)) {
+    if (!ContainsNoNUL(name)) {
         return false;
     }
 
@@ -169,7 +169,7 @@ static bool LookupIntern(const std::string& name, std::vector<CNetAddr>& vIP, un
 
 bool LookupHost(const std::string& name, std::vector<CNetAddr>& vIP, unsigned int nMaxSolutions, bool fAllowLookup, DNSLookupFn dns_lookup_function)
 {
-    if (!ValidAsCString(name)) {
+    if (!ContainsNoNUL(name)) {
         return false;
     }
     std::string strHost = name;
@@ -184,7 +184,7 @@ bool LookupHost(const std::string& name, std::vector<CNetAddr>& vIP, unsigned in
 
 bool LookupHost(const std::string& name, CNetAddr& addr, bool fAllowLookup, DNSLookupFn dns_lookup_function)
 {
-    if (!ValidAsCString(name)) {
+    if (!ContainsNoNUL(name)) {
         return false;
     }
     std::vector<CNetAddr> vIP;
@@ -197,7 +197,7 @@ bool LookupHost(const std::string& name, CNetAddr& addr, bool fAllowLookup, DNSL
 
 bool Lookup(const std::string& name, std::vector<CService>& vAddr, uint16_t portDefault, bool fAllowLookup, unsigned int nMaxSolutions, DNSLookupFn dns_lookup_function)
 {
-    if (name.empty() || !ValidAsCString(name)) {
+    if (name.empty() || !ContainsNoNUL(name)) {
         return false;
     }
     uint16_t port{portDefault};
@@ -216,7 +216,7 @@ bool Lookup(const std::string& name, std::vector<CService>& vAddr, uint16_t port
 
 bool Lookup(const std::string& name, CService& addr, uint16_t portDefault, bool fAllowLookup, DNSLookupFn dns_lookup_function)
 {
-    if (!ValidAsCString(name)) {
+    if (!ContainsNoNUL(name)) {
         return false;
     }
     std::vector<CService> vService;
@@ -229,7 +229,7 @@ bool Lookup(const std::string& name, CService& addr, uint16_t portDefault, bool
 
 CService LookupNumeric(const std::string& name, uint16_t portDefault, DNSLookupFn dns_lookup_function)
 {
-    if (!ValidAsCString(name)) {
+    if (!ContainsNoNUL(name)) {
         return {};
     }
     CService addr;
@@ -684,7 +684,7 @@ bool ConnectThroughProxy(const Proxy& proxy, const std::string& strDest, uint16_
 
 bool LookupSubNet(const std::string& subnet_str, CSubNet& subnet_out)
 {
-    if (!ValidAsCString(subnet_str)) {
+    if (!ContainsNoNUL(subnet_str)) {
         return false;
     }
 

src/psbt.cpp

@@ -388,18 +388,17 @@ std::string PSBTRoleName(PSBTRole role) {
 
 bool DecodeBase64PSBT(PartiallySignedTransaction& psbt, const std::string& base64_tx, std::string& error)
 {
-    bool invalid;
+    auto tx_data = DecodeBase64(base64_tx);
-    std::string tx_data = DecodeBase64(base64_tx, &invalid);
+    if (!tx_data) {
-    if (invalid) {
         error = "invalid base64";
         return false;
     }
-    return DecodeRawPSBT(psbt, tx_data, error);
+    return DecodeRawPSBT(psbt, MakeByteSpan(*tx_data), error);
 }
 
-bool DecodeRawPSBT(PartiallySignedTransaction& psbt, const std::string& tx_data, std::string& error)
+bool DecodeRawPSBT(PartiallySignedTransaction& psbt, Span<const std::byte> tx_data, std::string& error)
 {
-    CDataStream ss_data(MakeByteSpan(tx_data), SER_NETWORK, PROTOCOL_VERSION);
+    CDataStream ss_data(tx_data, SER_NETWORK, PROTOCOL_VERSION);
     try {
         ss_data >> psbt;
         if (!ss_data.empty()) {

src/psbt.h

@@ -988,6 +988,6 @@ bool FinalizeAndExtractPSBT(PartiallySignedTransaction& psbtx, CMutableTransacti
 //! Decode a base64ed PSBT into a PartiallySignedTransaction
 [[nodiscard]] bool DecodeBase64PSBT(PartiallySignedTransaction& decoded_psbt, const std::string& base64_psbt, std::string& error);
 //! Decode a raw (binary blob) PSBT into a PartiallySignedTransaction
-[[nodiscard]] bool DecodeRawPSBT(PartiallySignedTransaction& decoded_psbt, const std::string& raw_psbt, std::string& error);
+[[nodiscard]] bool DecodeRawPSBT(PartiallySignedTransaction& decoded_psbt, Span<const std::byte> raw_psbt, std::string& error);
 
 #endif // BITCOIN_PSBT_H

src/qt/walletframe.cpp

@@ -194,16 +194,16 @@ void WalletFrame::gotoVerifyMessageTab(QString addr)
 
 void WalletFrame::gotoLoadPSBT(bool from_clipboard)
 {
-    std::string data;
+    std::vector<unsigned char> data;
 
     if (from_clipboard) {
         std::string raw = QApplication::clipboard()->text().toStdString();
-        bool invalid;
+        auto result = DecodeBase64(raw);
-        data = DecodeBase64(raw, &invalid);
+        if (!result) {
-        if (invalid) {
             Q_EMIT message(tr("Error"), tr("Unable to decode PSBT from clipboard (invalid base64)"), CClientUIInterface::MSG_ERROR);
             return;
         }
+        data = std::move(*result);
     } else {
         QString filename = GUIUtil::getOpenFileName(this,
             tr("Load Transaction Data"), QString(),
@@ -214,12 +214,12 @@ void WalletFrame::gotoLoadPSBT(bool from_clipboard)
             return;
         }
         std::ifstream in{filename.toLocal8Bit().data(), std::ios::binary};
-        data = std::string(std::istreambuf_iterator<char>{in}, {});
+        data.assign(std::istream_iterator<unsigned char>{in}, {});
     }
 
     std::string error;
     PartiallySignedTransaction psbtx;
-    if (!DecodeRawPSBT(psbtx, data, error)) {
+    if (!DecodeRawPSBT(psbtx, MakeByteSpan(data), error)) {
         Q_EMIT message(tr("Error"), tr("Unable to decode PSBT") + "\n" + QString::fromStdString(error), CClientUIInterface::MSG_ERROR);
         return;
     }

src/test/base32_tests.cpp

@@ -22,20 +22,16 @@ BOOST_AUTO_TEST_CASE(base32_testvectors)
         BOOST_CHECK_EQUAL(strEnc, vstrOut[i]);
         strEnc = EncodeBase32(vstrIn[i], false);
         BOOST_CHECK_EQUAL(strEnc, vstrOutNoPadding[i]);
-        std::string strDec = DecodeBase32(vstrOut[i]);
+        auto dec = DecodeBase32(vstrOut[i]);
-        BOOST_CHECK_EQUAL(strDec, vstrIn[i]);
+        BOOST_REQUIRE(dec);
+        BOOST_CHECK_MESSAGE(MakeByteSpan(*dec) == MakeByteSpan(vstrIn[i]), vstrOut[i]);
     }
 
     // Decoding strings with embedded NUL characters should fail
-    bool failure;
+    BOOST_CHECK(!DecodeBase32("invalid\0"s)); // correct size, invalid due to \0
-    (void)DecodeBase32("invalid\0"s, &failure); // correct size, invalid due to \0
+    BOOST_CHECK(DecodeBase32("AWSX3VPP"s)); // valid
-    BOOST_CHECK(failure);
+    BOOST_CHECK(!DecodeBase32("AWSX3VPP\0invalid"s)); // correct size, invalid due to \0
-    (void)DecodeBase32("AWSX3VPP"s, &failure); // valid
+    BOOST_CHECK(!DecodeBase32("AWSX3VPPinvalid"s)); // invalid size
-    BOOST_CHECK(!failure);
-    (void)DecodeBase32("AWSX3VPP\0invalid"s, &failure); // correct size, invalid due to \0
-    BOOST_CHECK(failure);
-    (void)DecodeBase32("AWSX3VPPinvalid"s, &failure); // invalid size
-    BOOST_CHECK(failure);
 }
 
 BOOST_AUTO_TEST_SUITE_END()

src/test/base64_tests.cpp

@@ -19,8 +19,9 @@ BOOST_AUTO_TEST_CASE(base64_testvectors)
     {
         std::string strEnc = EncodeBase64(vstrIn[i]);
         BOOST_CHECK_EQUAL(strEnc, vstrOut[i]);
-        std::string strDec = DecodeBase64(strEnc);
+        auto dec = DecodeBase64(strEnc);
-        BOOST_CHECK_EQUAL(strDec, vstrIn[i]);
+        BOOST_REQUIRE(dec);
+        BOOST_CHECK_MESSAGE(MakeByteSpan(*dec) == MakeByteSpan(vstrIn[i]), vstrOut[i]);
     }
 
     {
@@ -34,15 +35,10 @@ BOOST_AUTO_TEST_CASE(base64_testvectors)
     }
 
     // Decoding strings with embedded NUL characters should fail
-    bool failure;
+    BOOST_CHECK(!DecodeBase64("invalid\0"s));
-    (void)DecodeBase64("invalid\0"s, &failure);
+    BOOST_CHECK(DecodeBase64("nQB/pZw="s));
-    BOOST_CHECK(failure);
+    BOOST_CHECK(!DecodeBase64("nQB/pZw=\0invalid"s));
-    (void)DecodeBase64("nQB/pZw="s, &failure);
+    BOOST_CHECK(!DecodeBase64("nQB/pZw=invalid\0"s));
-    BOOST_CHECK(!failure);
-    (void)DecodeBase64("nQB/pZw=\0invalid"s, &failure);
-    BOOST_CHECK(failure);
-    (void)DecodeBase64("nQB/pZw=invalid\0"s, &failure);
-    BOOST_CHECK(failure);
 }
 
 BOOST_AUTO_TEST_SUITE_END()

src/test/fuzz/base_encode_decode.cpp

@@ -26,7 +26,7 @@ FUZZ_TARGET_INIT(base_encode_decode, initialize_base_encode_decode)
     std::vector<unsigned char> decoded;
     if (DecodeBase58(random_encoded_string, decoded, 100)) {
         const std::string encoded_string = EncodeBase58(decoded);
-        assert(encoded_string == TrimString(encoded_string));
+        assert(encoded_string == TrimStringView(encoded_string));
         assert(ToLower(encoded_string) == ToLower(TrimString(random_encoded_string)));
     }
 
@@ -36,17 +36,16 @@ FUZZ_TARGET_INIT(base_encode_decode, initialize_base_encode_decode)
         assert(ToLower(encoded_string) == ToLower(TrimString(random_encoded_string)));
     }
 
-    bool pf_invalid;
+    auto result = DecodeBase32(random_encoded_string);
-    std::string decoded_string = DecodeBase32(random_encoded_string, &pf_invalid);
+    if (result) {
-    if (!pf_invalid) {
+        const std::string encoded_string = EncodeBase32(*result);
-        const std::string encoded_string = EncodeBase32(decoded_string);
+        assert(encoded_string == TrimStringView(encoded_string));
-        assert(encoded_string == TrimString(encoded_string));
         assert(ToLower(encoded_string) == ToLower(TrimString(random_encoded_string)));
     }
 
-    decoded_string = DecodeBase64(random_encoded_string, &pf_invalid);
+    result = DecodeBase64(random_encoded_string);
-    if (!pf_invalid) {
+    if (result) {
-        const std::string encoded_string = EncodeBase64(decoded_string);
+        const std::string encoded_string = EncodeBase64(*result);
         assert(encoded_string == TrimString(encoded_string));
         assert(ToLower(encoded_string) == ToLower(TrimString(random_encoded_string)));
     }

src/test/fuzz/http_request.cpp

@@ -39,7 +39,7 @@ FUZZ_TARGET(http_request)
     // and is a consequence of our hacky but necessary use of the internal function evhttp_parse_firstline_ in
     // this fuzzing harness. The workaround is not aesthetically pleasing, but it successfully avoids the troublesome
     // code path. " http:// HTTP/1.1\n" was a crashing input prior to this workaround.
-    const std::string http_buffer_str = ToLower({http_buffer.begin(), http_buffer.end()});
+    const std::string http_buffer_str = ToLower(std::string{http_buffer.begin(), http_buffer.end()});
     if (http_buffer_str.find(" http://") != std::string::npos || http_buffer_str.find(" https://") != std::string::npos ||
         evhttp_parse_firstline_(evreq, evbuf) != 1 || evhttp_parse_headers_(evreq, evbuf) != 1) {
         evbuffer_free(evbuf);

src/test/fuzz/psbt.cpp

@@ -32,7 +32,8 @@ FUZZ_TARGET_INIT(psbt, initialize_psbt)
     FuzzedDataProvider fuzzed_data_provider{buffer.data(), buffer.size()};
     PartiallySignedTransaction psbt_mut;
     std::string error;
-    if (!DecodeRawPSBT(psbt_mut, fuzzed_data_provider.ConsumeRandomLengthString(), error)) {
+    auto str = fuzzed_data_provider.ConsumeRandomLengthString();
+    if (!DecodeRawPSBT(psbt_mut, MakeByteSpan(str), error)) {
         return;
     }
     const PartiallySignedTransaction psbt = psbt_mut;
@@ -79,7 +80,8 @@ FUZZ_TARGET_INIT(psbt, initialize_psbt)
     }
 
     PartiallySignedTransaction psbt_merge;
-    if (!DecodeRawPSBT(psbt_merge, fuzzed_data_provider.ConsumeRandomLengthString(), error)) {
+    str = fuzzed_data_provider.ConsumeRandomLengthString();
+    if (!DecodeRawPSBT(psbt_merge, MakeByteSpan(str), error)) {
         psbt_merge = psbt;
     }
     psbt_mut = psbt;

src/test/fuzz/string.cpp

@@ -42,7 +42,7 @@ bool LegacyParsePrechecks(const std::string& str)
         return false;
     if (str.size() >= 1 && (IsSpace(str[0]) || IsSpace(str[str.size() - 1]))) // No padding allowed
         return false;
-    if (!ValidAsCString(str)) // No embedded NUL characters allowed
+    if (!ContainsNoNUL(str)) // No embedded NUL characters allowed
         return false;
     return true;
 }
@@ -188,7 +188,7 @@ FUZZ_TARGET(string)
     (void)TrimString(random_string_1);
     (void)TrimString(random_string_1, random_string_2);
     (void)urlDecode(random_string_1);
-    (void)ValidAsCString(random_string_1);
+    (void)ContainsNoNUL(random_string_1);
     (void)_(random_string_1.c_str());
     try {
         throw scriptnum_error{random_string_1};

src/test/util_tests.cpp

@@ -226,17 +226,17 @@ BOOST_AUTO_TEST_CASE(util_Join)
 BOOST_AUTO_TEST_CASE(util_TrimString)
 {
     BOOST_CHECK_EQUAL(TrimString(" foo bar "), "foo bar");
-    BOOST_CHECK_EQUAL(TrimString("\t \n  \n \f\n\r\t\v\tfoo \n \f\n\r\t\v\tbar\t  \n \f\n\r\t\v\t\n "), "foo \n \f\n\r\t\v\tbar");
+    BOOST_CHECK_EQUAL(TrimStringView("\t \n  \n \f\n\r\t\v\tfoo \n \f\n\r\t\v\tbar\t  \n \f\n\r\t\v\t\n "), "foo \n \f\n\r\t\v\tbar");
     BOOST_CHECK_EQUAL(TrimString("\t \n foo \n\tbar\t \n "), "foo \n\tbar");
-    BOOST_CHECK_EQUAL(TrimString("\t \n foo \n\tbar\t \n ", "fobar"), "\t \n foo \n\tbar\t \n ");
+    BOOST_CHECK_EQUAL(TrimStringView("\t \n foo \n\tbar\t \n ", "fobar"), "\t \n foo \n\tbar\t \n ");
     BOOST_CHECK_EQUAL(TrimString("foo bar"), "foo bar");
-    BOOST_CHECK_EQUAL(TrimString("foo bar", "fobar"), " ");
+    BOOST_CHECK_EQUAL(TrimStringView("foo bar", "fobar"), " ");
     BOOST_CHECK_EQUAL(TrimString(std::string("\0 foo \0 ", 8)), std::string("\0 foo \0", 7));
-    BOOST_CHECK_EQUAL(TrimString(std::string(" foo ", 5)), std::string("foo", 3));
+    BOOST_CHECK_EQUAL(TrimStringView(std::string(" foo ", 5)), std::string("foo", 3));
     BOOST_CHECK_EQUAL(TrimString(std::string("\t\t\0\0\n\n", 6)), std::string("\0\0", 2));
-    BOOST_CHECK_EQUAL(TrimString(std::string("\x05\x04\x03\x02\x01\x00", 6)), std::string("\x05\x04\x03\x02\x01\x00", 6));
+    BOOST_CHECK_EQUAL(TrimStringView(std::string("\x05\x04\x03\x02\x01\x00", 6)), std::string("\x05\x04\x03\x02\x01\x00", 6));
     BOOST_CHECK_EQUAL(TrimString(std::string("\x05\x04\x03\x02\x01\x00", 6), std::string("\x05\x04\x03\x02\x01", 5)), std::string("\0", 1));
-    BOOST_CHECK_EQUAL(TrimString(std::string("\x05\x04\x03\x02\x01\x00", 6), std::string("\x05\x04\x03\x02\x01\x00", 6)), "");
+    BOOST_CHECK_EQUAL(TrimStringView(std::string("\x05\x04\x03\x02\x01\x00", 6), std::string("\x05\x04\x03\x02\x01\x00", 6)), "");
 }
 
 BOOST_AUTO_TEST_CASE(util_FormatParseISO8601DateTime)
@@ -2618,13 +2618,13 @@ BOOST_AUTO_TEST_CASE(message_hash)
 BOOST_AUTO_TEST_CASE(remove_prefix)
 {
     BOOST_CHECK_EQUAL(RemovePrefix("./util/system.h", "./"), "util/system.h");
-    BOOST_CHECK_EQUAL(RemovePrefix("foo", "foo"), "");
+    BOOST_CHECK_EQUAL(RemovePrefixView("foo", "foo"), "");
     BOOST_CHECK_EQUAL(RemovePrefix("foo", "fo"), "o");
-    BOOST_CHECK_EQUAL(RemovePrefix("foo", "f"), "oo");
+    BOOST_CHECK_EQUAL(RemovePrefixView("foo", "f"), "oo");
     BOOST_CHECK_EQUAL(RemovePrefix("foo", ""), "foo");
-    BOOST_CHECK_EQUAL(RemovePrefix("fo", "foo"), "fo");
+    BOOST_CHECK_EQUAL(RemovePrefixView("fo", "foo"), "fo");
     BOOST_CHECK_EQUAL(RemovePrefix("f", "foo"), "f");
-    BOOST_CHECK_EQUAL(RemovePrefix("", "foo"), "");
+    BOOST_CHECK_EQUAL(RemovePrefixView("", "foo"), "");
     BOOST_CHECK_EQUAL(RemovePrefix("", ""), "");
 }
 

src/util/message.cpp

@@ -35,14 +35,13 @@ MessageVerificationResult MessageVerify(
         return MessageVerificationResult::ERR_ADDRESS_NO_KEY;
     }
 
-    bool invalid = false;
+    auto signature_bytes = DecodeBase64(signature);
-    std::vector<unsigned char> signature_bytes = DecodeBase64(signature.c_str(), &invalid);
+    if (!signature_bytes) {
-    if (invalid) {
         return MessageVerificationResult::ERR_MALFORMED_SIGNATURE;
     }
 
     CPubKey pubkey;
-    if (!pubkey.RecoverCompact(MessageHash(message), signature_bytes)) {
+    if (!pubkey.RecoverCompact(MessageHash(message), *signature_bytes)) {
         return MessageVerificationResult::ERR_PUBKEY_NOT_RECOVERED;
     }
 

src/util/moneystr.cpp

@@ -40,7 +40,7 @@ std::string FormatMoney(const CAmount n)
 
 std::optional<CAmount> ParseMoney(const std::string& money_string)
 {
-    if (!ValidAsCString(money_string)) {
+    if (!ContainsNoNUL(money_string)) {
         return std::nullopt;
     }
     const std::string str = TrimString(money_string);

src/util/strencodings.cpp

@@ -24,15 +24,15 @@ static const std::string SAFE_CHARS[] =
     CHARS_ALPHA_NUM + "!*'();:@&=+$,/?#[]-_.~%", // SAFE_CHARS_URI
 };
 
-std::string SanitizeString(const std::string& str, int rule)
+std::string SanitizeString(std::string_view str, int rule)
 {
-    std::string strResult;
+    std::string result;
-    for (std::string::size_type i = 0; i < str.size(); i++)
+    for (char c : str) {
-    {
+        if (SAFE_CHARS[rule].find(c) != std::string::npos) {
-        if (SAFE_CHARS[rule].find(str[i]) != std::string::npos)
+            result.push_back(c);
-            strResult.push_back(str[i]);
+        }
     }
-    return strResult;
+    return result;
 }
 
 const signed char p_util_hexdigit[256] =
@@ -58,56 +58,43 @@ signed char HexDigit(char c)
     return p_util_hexdigit[(unsigned char)c];
 }
 
-bool IsHex(const std::string& str)
+bool IsHex(std::string_view str)
 {
-    for(std::string::const_iterator it(str.begin()); it != str.end(); ++it)
+    for (char c : str) {
-    {
+        if (HexDigit(c) < 0) return false;
-        if (HexDigit(*it) < 0)
-            return false;
     }
     return (str.size() > 0) && (str.size()%2 == 0);
 }
 
-bool IsHexNumber(const std::string& str)
+bool IsHexNumber(std::string_view str)
 {
-    size_t starting_location = 0;
+    if (str.substr(0, 2) == "0x") str.remove_prefix(2);
-    if (str.size() > 2 && *str.begin() == '0' && *(str.begin()+1) == 'x') {
+    for (char c : str) {
-        starting_location = 2;
-    }
-    for (const char c : str.substr(starting_location)) {
         if (HexDigit(c) < 0) return false;
     }
     // Return false for empty string or "0x".
-    return (str.size() > starting_location);
+    return str.size() > 0;
 }
 
-std::vector<unsigned char> ParseHex(const char* psz)
+std::vector<unsigned char> ParseHex(std::string_view str)
 {
     // convert hex dump to vector
     std::vector<unsigned char> vch;
-    while (true)
+    auto it = str.begin();
-    {
+    while (it != str.end() && it + 1 != str.end()) {
-        while (IsSpace(*psz))
+        if (IsSpace(*it)) {
-            psz++;
+            ++it;
-        signed char c = HexDigit(*psz++);
+            continue;
-        if (c == (signed char)-1)
+        }
-            break;
+        auto c1 = HexDigit(*(it++));
-        auto n{uint8_t(c << 4)};
+        auto c2 = HexDigit(*(it++));
-        c = HexDigit(*psz++);
+        if (c1 < 0 || c2 < 0) break;
-        if (c == (signed char)-1)
+        vch.push_back(uint8_t(c1 << 4) | c2);
-            break;
-        n |= c;
-        vch.push_back(n);
     }
     return vch;
 }
 
-std::vector<unsigned char> ParseHex(const std::string& str)
+void SplitHostPort(std::string_view in, uint16_t& portOut, std::string& hostOut)
-{
-    return ParseHex(str.c_str());
-}
-
-void SplitHostPort(std::string in, uint16_t& portOut, std::string& hostOut)
 {
     size_t colon = in.find_last_of(':');
     // if a : is found, and it either follows a [...], or no other : is in the string, treat it as port separator
@@ -139,7 +126,7 @@ std::string EncodeBase64(Span<const unsigned char> input)
     return str;
 }
 
-std::vector<unsigned char> DecodeBase64(const char* p, bool* pf_invalid)
+std::optional<std::vector<unsigned char>> DecodeBase64(std::string_view str)
 {
     static const int8_t decode64_table[256]{
         -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
@@ -157,46 +144,23 @@ std::vector<unsigned char> DecodeBase64(const char* p, bool* pf_invalid)
         -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1
     };
 
-    const char* e = p;
+    if (str.size() % 4 != 0) return {};
-    std::vector<uint8_t> val;
+    /* One or two = characters at the end are permitted. */
-    val.reserve(strlen(p));
+    if (str.size() >= 1 && str.back() == '=') str.remove_suffix(1);
-    while (*p != 0) {
+    if (str.size() >= 1 && str.back() == '=') str.remove_suffix(1);
-        int x = decode64_table[(unsigned char)*p];
-        if (x == -1) break;
-        val.push_back(uint8_t(x));
-        ++p;
-    }
 
     std::vector<unsigned char> ret;
-    ret.reserve((val.size() * 3) / 4);
+    ret.reserve((str.size() * 3) / 4);
-    bool valid = ConvertBits<6, 8, false>([&](unsigned char c) { ret.push_back(c); }, val.begin(), val.end());
+    bool valid = ConvertBits<6, 8, false>(
-
+        [&](unsigned char c) { ret.push_back(c); },
-    const char* q = p;
+        str.begin(), str.end(),
-    while (valid && *p != 0) {
+        [](char c) { return decode64_table[uint8_t(c)]; }
-        if (*p != '=') {
+    );
-            valid = false;
+    if (!valid) return {};
-            break;
-        }
-        ++p;
-    }
-    valid = valid && (p - e) % 4 == 0 && p - q < 4;
-    if (pf_invalid) *pf_invalid = !valid;
 
     return ret;
 }
 
-std::string DecodeBase64(const std::string& str, bool* pf_invalid)
-{
-    if (!ValidAsCString(str)) {
-        if (pf_invalid) {
-            *pf_invalid = true;
-        }
-        return {};
-    }
-    std::vector<unsigned char> vchRet = DecodeBase64(str.c_str(), pf_invalid);
-    return std::string((const char*)vchRet.data(), vchRet.size());
-}
-
 std::string EncodeBase32(Span<const unsigned char> input, bool pad)
 {
     static const char *pbase32 = "abcdefghijklmnopqrstuvwxyz234567";
@@ -212,12 +176,12 @@ std::string EncodeBase32(Span<const unsigned char> input, bool pad)
     return str;
 }
 
-std::string EncodeBase32(const std::string& str, bool pad)
+std::string EncodeBase32(std::string_view str, bool pad)
 {
     return EncodeBase32(MakeUCharSpan(str), pad);
 }
 
-std::vector<unsigned char> DecodeBase32(const char* p, bool* pf_invalid)
+std::optional<std::vector<unsigned char>> DecodeBase32(std::string_view str)
 {
     static const int8_t decode32_table[256]{
         -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
@@ -235,49 +199,29 @@ std::vector<unsigned char> DecodeBase32(const char* p, bool* pf_invalid)
         -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1
     };
 
-    const char* e = p;
+    if (str.size() % 8 != 0) return {};
-    std::vector<uint8_t> val;
+    /* 1, 3, 4, or 6 padding '=' suffix characters are permitted. */
-    val.reserve(strlen(p));
+    if (str.size() >= 1 && str.back() == '=') str.remove_suffix(1);
-    while (*p != 0) {
+    if (str.size() >= 2 && str.substr(str.size() - 2) == "==") str.remove_suffix(2);
-        int x = decode32_table[(unsigned char)*p];
+    if (str.size() >= 1 && str.back() == '=') str.remove_suffix(1);
-        if (x == -1) break;
+    if (str.size() >= 2 && str.substr(str.size() - 2) == "==") str.remove_suffix(2);
-        val.push_back(uint8_t(x));
-        ++p;
-    }
 
     std::vector<unsigned char> ret;
-    ret.reserve((val.size() * 5) / 8);
+    ret.reserve((str.size() * 5) / 8);
-    bool valid = ConvertBits<5, 8, false>([&](unsigned char c) { ret.push_back(c); }, val.begin(), val.end());
+    bool valid = ConvertBits<5, 8, false>(
+        [&](unsigned char c) { ret.push_back(c); },
+        str.begin(), str.end(),
+        [](char c) { return decode32_table[uint8_t(c)]; }
+    );
 
-    const char* q = p;
+    if (!valid) return {};
-    while (valid && *p != 0) {
-        if (*p != '=') {
-            valid = false;
-            break;
-        }
-        ++p;
-    }
-    valid = valid && (p - e) % 8 == 0 && p - q < 8;
-    if (pf_invalid) *pf_invalid = !valid;
 
     return ret;
 }
 
-std::string DecodeBase32(const std::string& str, bool* pf_invalid)
-{
-    if (!ValidAsCString(str)) {
-        if (pf_invalid) {
-            *pf_invalid = true;
-        }
-        return {};
-    }
-    std::vector<unsigned char> vchRet = DecodeBase32(str.c_str(), pf_invalid);
-    return std::string((const char*)vchRet.data(), vchRet.size());
-}
-
 namespace {
 template <typename T>
-bool ParseIntegral(const std::string& str, T* out)
+bool ParseIntegral(std::string_view str, T* out)
 {
     static_assert(std::is_integral<T>::value);
     // Replicate the exact behavior of strtol/strtoll/strtoul/strtoull when
@@ -296,37 +240,37 @@ bool ParseIntegral(const std::string& str, T* out)
 }
 }; // namespace
 
-bool ParseInt32(const std::string& str, int32_t* out)
+bool ParseInt32(std::string_view str, int32_t* out)
 {
     return ParseIntegral<int32_t>(str, out);
 }
 
-bool ParseInt64(const std::string& str, int64_t* out)
+bool ParseInt64(std::string_view str, int64_t* out)
 {
     return ParseIntegral<int64_t>(str, out);
 }
 
-bool ParseUInt8(const std::string& str, uint8_t* out)
+bool ParseUInt8(std::string_view str, uint8_t* out)
 {
     return ParseIntegral<uint8_t>(str, out);
 }
 
-bool ParseUInt16(const std::string& str, uint16_t* out)
+bool ParseUInt16(std::string_view str, uint16_t* out)
 {
     return ParseIntegral<uint16_t>(str, out);
 }
 
-bool ParseUInt32(const std::string& str, uint32_t* out)
+bool ParseUInt32(std::string_view str, uint32_t* out)
 {
     return ParseIntegral<uint32_t>(str, out);
 }
 
-bool ParseUInt64(const std::string& str, uint64_t* out)
+bool ParseUInt64(std::string_view str, uint64_t* out)
 {
     return ParseIntegral<uint64_t>(str, out);
 }
 
-std::string FormatParagraph(const std::string& in, size_t width, size_t indent)
+std::string FormatParagraph(std::string_view in, size_t width, size_t indent)
 {
     assert(width >= indent);
     std::stringstream out;
@@ -395,7 +339,7 @@ static inline bool ProcessMantissaDigit(char ch, int64_t &mantissa, int &mantiss
     return true;
 }
 
-bool ParseFixedPoint(const std::string &val, int decimals, int64_t *amount_out)
+bool ParseFixedPoint(std::string_view val, int decimals, int64_t *amount_out)
 {
     int64_t mantissa = 0;
     int64_t exponent = 0;
@@ -487,14 +431,14 @@ bool ParseFixedPoint(const std::string &val, int decimals, int64_t *amount_out)
     return true;
 }
 
-std::string ToLower(const std::string& str)
+std::string ToLower(std::string_view str)
 {
     std::string r;
     for (auto ch : str) r += ToLower(ch);
     return r;
 }
 
-std::string ToUpper(const std::string& str)
+std::string ToUpper(std::string_view str)
 {
     std::string r;
     for (auto ch : str) r += ToUpper(ch);
@@ -522,7 +466,7 @@ std::string HexStr(const Span<const uint8_t> s)
     return rv;
 }
 
-std::optional<uint64_t> ParseByteUnits(const std::string& str, ByteUnit default_multiplier)
+std::optional<uint64_t> ParseByteUnits(std::string_view str, ByteUnit default_multiplier)
 {
     if (str.empty()) {
         return std::nullopt;

src/util/strencodings.h

@@ -54,24 +54,21 @@ enum class ByteUnit : uint64_t {
 * @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 SanitizeString(std::string_view str, int rule = SAFE_CHARS_DEFAULT);
-std::vector<unsigned char> ParseHex(const char* psz);
+std::vector<unsigned char> ParseHex(std::string_view str);
-std::vector<unsigned char> ParseHex(const std::string& str);
 signed char HexDigit(char c);
 /* Returns true if each character in str is a hex character, and has an even
  * number of hex digits.*/
-bool IsHex(const std::string& str);
+bool IsHex(std::string_view str);
 /**
 * Return true if the string is a hex number, optionally prefixed with "0x"
 */
-bool IsHexNumber(const std::string& str);
+bool IsHexNumber(std::string_view str);
-std::vector<unsigned char> DecodeBase64(const char* p, bool* pf_invalid = nullptr);
+std::optional<std::vector<unsigned char>> DecodeBase64(std::string_view str);
-std::string DecodeBase64(const std::string& str, bool* pf_invalid = nullptr);
 std::string EncodeBase64(Span<const unsigned char> input);
 inline std::string EncodeBase64(Span<const std::byte> input) { return EncodeBase64(MakeUCharSpan(input)); }
-inline std::string EncodeBase64(const std::string& str) { return EncodeBase64(MakeUCharSpan(str)); }
+inline std::string EncodeBase64(std::string_view str) { return EncodeBase64(MakeUCharSpan(str)); }
-std::vector<unsigned char> DecodeBase32(const char* p, bool* pf_invalid = nullptr);
+std::optional<std::vector<unsigned char>> DecodeBase32(std::string_view str);
-std::string DecodeBase32(const std::string& str, bool* pf_invalid = nullptr);
 
 /**
  * Base32 encode.
@@ -85,9 +82,9 @@ std::string EncodeBase32(Span<const unsigned char> input, bool pad = true);
  * If `pad` is true, then the output will be padded with '=' so that its length
  * is a multiple of 8.
  */
-std::string EncodeBase32(const std::string& str, bool pad = true);
+std::string EncodeBase32(std::string_view str, bool pad = true);
 
-void SplitHostPort(std::string in, uint16_t& portOut, std::string& hostOut);
+void SplitHostPort(std::string_view in, uint16_t& portOut, std::string& hostOut);
 
 // LocaleIndependentAtoi is provided for backwards compatibility reasons.
 //
@@ -101,12 +98,12 @@ void SplitHostPort(std::string in, uint16_t& portOut, std::string& hostOut);
 // undefined behavior, while this function returns the maximum or minimum
 // values, respectively.
 template <typename T>
-T LocaleIndependentAtoi(const std::string& str)
+T LocaleIndependentAtoi(std::string_view str)
 {
     static_assert(std::is_integral<T>::value);
     T result;
     // Emulate atoi(...) handling of white space and leading +/-.
-    std::string s = TrimString(str);
+    std::string_view s = TrimStringView(str);
     if (!s.empty() && s[0] == '+') {
         if (s.length() >= 2 && s[1] == '-') {
             return 0;
@@ -162,7 +159,7 @@ constexpr inline bool IsSpace(char c) noexcept {
  *   parsed value is not in the range representable by the type T.
  */
 template <typename T>
-std::optional<T> ToIntegral(const std::string& str)
+std::optional<T> ToIntegral(std::string_view str)
 {
     static_assert(std::is_integral<T>::value);
     T result;
@@ -178,42 +175,42 @@ std::optional<T> ToIntegral(const std::string& str)
  * @returns true if the entire string could be parsed as valid integer,
  *   false if not the entire string could be parsed or when overflow or underflow occurred.
  */
-[[nodiscard]] bool ParseInt32(const std::string& str, int32_t *out);
+[[nodiscard]] bool ParseInt32(std::string_view str, int32_t *out);
 
 /**
  * Convert string to signed 64-bit integer with strict parse error feedback.
  * @returns true if the entire string could be parsed as valid integer,
  *   false if not the entire string could be parsed or when overflow or underflow occurred.
  */
-[[nodiscard]] bool ParseInt64(const std::string& str, int64_t *out);
+[[nodiscard]] bool ParseInt64(std::string_view str, int64_t *out);
 
 /**
  * Convert decimal string to unsigned 8-bit integer with strict parse error feedback.
  * @returns true if the entire string could be parsed as valid integer,
  *   false if not the entire string could be parsed or when overflow or underflow occurred.
  */
-[[nodiscard]] bool ParseUInt8(const std::string& str, uint8_t *out);
+[[nodiscard]] bool ParseUInt8(std::string_view str, uint8_t *out);
 
 /**
  * Convert decimal string to unsigned 16-bit integer with strict parse error feedback.
  * @returns true if the entire string could be parsed as valid integer,
  *   false if the entire string could not be parsed or if overflow or underflow occurred.
  */
-[[nodiscard]] bool ParseUInt16(const std::string& str, uint16_t* out);
+[[nodiscard]] bool ParseUInt16(std::string_view str, uint16_t* out);
 
 /**
  * Convert decimal string to unsigned 32-bit integer with strict parse error feedback.
  * @returns true if the entire string could be parsed as valid integer,
  *   false if not the entire string could be parsed or when overflow or underflow occurred.
  */
-[[nodiscard]] bool ParseUInt32(const std::string& str, uint32_t *out);
+[[nodiscard]] bool ParseUInt32(std::string_view str, uint32_t *out);
 
 /**
  * Convert decimal string to unsigned 64-bit integer with strict parse error feedback.
  * @returns true if the entire string could be parsed as valid integer,
  *   false if not the entire string could be parsed or when overflow or underflow occurred.
  */
-[[nodiscard]] bool ParseUInt64(const std::string& str, uint64_t *out);
+[[nodiscard]] bool ParseUInt64(std::string_view str, uint64_t *out);
 
 /**
  * Convert a span of bytes to a lower-case hexadecimal string.
@@ -226,7 +223,7 @@ inline std::string HexStr(const Span<const std::byte> s) { return HexStr(MakeUCh
  * Format a paragraph of text to a fixed width, adding spaces for
  * indentation to any added line.
  */
-std::string FormatParagraph(const std::string& in, size_t width = 79, size_t indent = 0);
+std::string FormatParagraph(std::string_view in, size_t width = 79, size_t indent = 0);
 
 /**
  * Timing-attack-resistant comparison.
@@ -248,17 +245,28 @@ bool TimingResistantEqual(const T& a, const T& b)
  * @returns true on success, false on error.
  * @note The result must be in the range (-10^18,10^18), otherwise an overflow error will trigger.
  */
-[[nodiscard]] bool ParseFixedPoint(const std::string &val, int decimals, int64_t *amount_out);
+[[nodiscard]] bool ParseFixedPoint(std::string_view, int decimals, int64_t *amount_out);
+
+namespace {
+/** Helper class for the default infn argument to ConvertBits (just returns the input). */
+struct IntIdentity
+{
+    [[maybe_unused]] int operator()(int x) const { return x; }
+};
+
+} // namespace
 
 /** Convert from one power-of-2 number base to another. */
-template<int frombits, int tobits, bool pad, typename O, typename I>
+template<int frombits, int tobits, bool pad, typename O, typename It, typename I = IntIdentity>
-bool ConvertBits(const O& outfn, I it, I end) {
+bool ConvertBits(O outfn, It it, It end, I infn = {}) {
     size_t acc = 0;
     size_t bits = 0;
     constexpr size_t maxv = (1 << tobits) - 1;
     constexpr size_t max_acc = (1 << (frombits + tobits - 1)) - 1;
     while (it != end) {
-        acc = ((acc << frombits) | *it) & max_acc;
+        int v = infn(*it);
+        if (v < 0) return false;
+        acc = ((acc << frombits) | v) & max_acc;
         bits += frombits;
         while (bits >= tobits) {
             bits -= tobits;
@@ -298,7 +306,7 @@ constexpr char ToLower(char c)
  * @param[in] str   the string to convert to lowercase.
  * @returns         lowercased equivalent of str
  */
-std::string ToLower(const std::string& str);
+std::string ToLower(std::string_view str);
 
 /**
  * Converts the given character to its uppercase equivalent.
@@ -324,7 +332,7 @@ constexpr char ToUpper(char c)
  * @param[in] str   the string to convert to uppercase.
  * @returns         UPPERCASED EQUIVALENT OF str
  */
-std::string ToUpper(const std::string& str);
+std::string ToUpper(std::string_view str);
 
 /**
  * Capitalizes the first character of the given string.
@@ -348,6 +356,6 @@ std::string Capitalize(std::string str);
  * @returns                        optional uint64_t bytes from str or nullopt
  *                                 if ToIntegral is false, str is empty, trailing whitespace or overflow
  */
-std::optional<uint64_t> ParseByteUnits(const std::string& str, ByteUnit default_multiplier);
+std::optional<uint64_t> ParseByteUnits(std::string_view str, ByteUnit default_multiplier);
 
 #endif // BITCOIN_UTIL_STRENCODINGS_H

src/util/string.h

@@ -21,17 +21,22 @@
     return spanparsing::Split<std::string>(str, sep);
 }
 
-[[nodiscard]] inline std::string TrimString(const std::string& str, const std::string& pattern = " \f\n\r\t\v")
+[[nodiscard]] inline std::string_view TrimStringView(std::string_view str, std::string_view pattern = " \f\n\r\t\v")
 {
     std::string::size_type front = str.find_first_not_of(pattern);
     if (front == std::string::npos) {
-        return std::string();
+        return {};
     }
     std::string::size_type end = str.find_last_not_of(pattern);
     return str.substr(front, end - front + 1);
 }
 
-[[nodiscard]] inline std::string RemovePrefix(const std::string& str, const std::string& prefix)
+[[nodiscard]] inline std::string TrimString(std::string_view str, std::string_view pattern = " \f\n\r\t\v")
+{
+    return std::string(TrimStringView(str, pattern));
+}
+
+[[nodiscard]] inline std::string_view RemovePrefixView(std::string_view str, std::string_view prefix)
 {
     if (str.substr(0, prefix.size()) == prefix) {
         return str.substr(prefix.size());
@@ -39,6 +44,11 @@
     return str;
 }
 
+[[nodiscard]] inline std::string RemovePrefix(std::string_view str, std::string_view prefix)
+{
+    return std::string(RemovePrefixView(str, prefix));
+}
+
 /**
  * Join a list of items
  *
@@ -58,14 +68,14 @@ auto Join(const std::vector<T>& list, const BaseType& separator, UnaryOp unary_o
     return ret;
 }
 
-template <typename T>
+template <typename T, typename T2>
-T Join(const std::vector<T>& list, const T& separator)
+T Join(const std::vector<T>& list, const T2& separator)
 {
     return Join(list, separator, [](const T& i) { return i; });
 }
 
 // Explicit overload needed for c_str arguments, which would otherwise cause a substitution failure in the template above.
-inline std::string Join(const std::vector<std::string>& list, const std::string& separator)
+inline std::string Join(const std::vector<std::string>& list, std::string_view separator)
 {
     return Join<std::string>(list, separator);
 }
@@ -81,9 +91,12 @@ inline std::string MakeUnorderedList(const std::vector<std::string>& items)
 /**
  * Check if a string does not contain any embedded NUL (\0) characters
  */
-[[nodiscard]] inline bool ValidAsCString(const std::string& str) noexcept
+[[nodiscard]] inline bool ContainsNoNUL(std::string_view str) noexcept
 {
-    return str.size() == strlen(str.c_str());
+    for (auto c : str) {
+        if (c == 0) return false;
+    }
+    return true;
 }
 
 /**

src/util/system.cpp

@@ -853,8 +853,8 @@ static bool GetConfigOptions(std::istream& stream, const std::string& filepath,
                 error = strprintf("parse error on line %i: %s, options in configuration file must be specified without leading -", linenr, str);
                 return false;
             } else if ((pos = str.find('=')) != std::string::npos) {
-                std::string name = prefix + TrimString(str.substr(0, pos), pattern);
+                std::string name = prefix + TrimString(std::string_view{str}.substr(0, pos), pattern);
-                std::string value = TrimString(str.substr(pos + 1), pattern);
+                std::string_view value = TrimStringView(std::string_view{str}.substr(pos + 1), pattern);
                 if (used_hash && name.find("rpcpassword") != std::string::npos) {
                     error = strprintf("parse error on line %i, using # in rpcpassword can be ambiguous and should be avoided", linenr);
                     return false;