Inside Bitcoin

1// Copyright (c) 2009-2010 Satoshi Nakamoto
2// Copyright (c) 2009-present The Bitcoin Core developers
3// Distributed under the MIT software license, see the accompanying
4// file COPYING or http://www.opensource.org/licenses/mit-license.php.
5
6/**
7 * Utilities for converting data from/to strings.
8 */
9#ifndef BITCOIN_UTIL_STRENCODINGS_H
10#define BITCOIN_UTIL_STRENCODINGS_H
11
12#include <span.h>
13#include <util/string.h>
14
15#include <array>
16#include <bit>
17#include <charconv>
18#include <cstddef>
19#include <cstdint>
20#include <functional>
21#include <limits>
22#include <optional>
23#include <span>
24#include <string>
25#include <string_view>
26#include <system_error>
27#include <type_traits>
28#include <vector>
29
30/** Used by SanitizeString() */
31enum SafeChars
32{
33    SAFE_CHARS_DEFAULT, //!< The full set of allowed chars
34    SAFE_CHARS_UA_COMMENT, //!< BIP-0014 subset
35    SAFE_CHARS_FILENAME, //!< Chars allowed in filenames
36    SAFE_CHARS_URI, //!< Chars allowed in URIs (RFC 3986)
37};
38
39/**
40 * Used by ParseByteUnits()
41 * Lowercase base 1000
42 * Uppercase base 1024
43*/
44enum class ByteUnit : uint64_t {
45    NOOP = 1ULL,
46    k = 1000ULL,
47    K = 1024ULL,
48    m = 1'000'000ULL,
49    M = 1ULL << 20,
50    g = 1'000'000'000ULL,
51    G = 1ULL << 30,
52    t = 1'000'000'000'000ULL,
53    T = 1ULL << 40,
54};
55
56/**
57* Remove unsafe chars. Safe chars chosen to allow simple messages/URLs/email
58* addresses, but avoid anything even possibly remotely dangerous like & or >
59* @param[in] str    The string to sanitize
60* @param[in] rule   The set of safe chars to choose (default: least restrictive)
61* @return           A new string without unsafe chars
62*/
63std::string SanitizeString(std::string_view str, int rule = SAFE_CHARS_DEFAULT);
64/** Parse the hex string into bytes (uint8_t or std::byte). Ignores whitespace. Returns nullopt on invalid input. */
65template <typename Byte = std::byte>
66std::optional<std::vector<Byte>> TryParseHex(std::string_view str);
67/** Like TryParseHex, but returns an empty vector on invalid input. */
68template <typename Byte = uint8_t>
69std::vector<Byte> ParseHex(std::string_view hex_str)
70{
71    return TryParseHex<Byte>(hex_str).value_or(std::vector<Byte>{});
72}
73/* Returns true if each character in str is a hex character, and has an even
74 * number of hex digits.*/
75bool IsHex(std::string_view str);
76std::optional<std::vector<unsigned char>> DecodeBase64(std::string_view str);
77std::string EncodeBase64(std::span<const unsigned char> input);
78inline std::string EncodeBase64(std::span<const std::byte> input) { return EncodeBase64(MakeUCharSpan(input)); }
79inline std::string EncodeBase64(std::string_view str) { return EncodeBase64(MakeUCharSpan(str)); }
80std::optional<std::vector<unsigned char>> DecodeBase32(std::string_view str);
81
82/**
83 * Base32 encode.
84 * If `pad` is true, then the output will be padded with '=' so that its length
85 * is a multiple of 8.
86 */
87std::string EncodeBase32(std::span<const unsigned char> input, bool pad = true);
88
89/**
90 * Base32 encode.
91 * If `pad` is true, then the output will be padded with '=' so that its length
92 * is a multiple of 8.
93 */
94std::string EncodeBase32(std::string_view str, bool pad = true);
95
96/**
97 * Splits socket address string into host string and port value.
98 * Validates port value.
99 *
100 * @param[in] in        The socket address string to split.
101 * @param[out] portOut  Port-portion of the input, if found and parsable.
102 * @param[out] hostOut  Host-portion of the input, if found.
103 * @return              true if port-portion is absent or within its allowed range, otherwise false
104 */
105bool SplitHostPort(std::string_view in, uint16_t& portOut, std::string& hostOut);
106
107// LocaleIndependentAtoi is provided for backwards compatibility reasons.
108//
109// New code should use ToIntegral.
110//
111// The goal of LocaleIndependentAtoi is to replicate the defined behaviour of
112// std::atoi as it behaves under the "C" locale, and remove some undefined
113// behavior. If the parsed value is bigger than the integer type's maximum
114// value, or smaller than the integer type's minimum value, std::atoi has
115// undefined behavior, while this function returns the maximum or minimum
116// values, respectively.
117template <typename T>
118T LocaleIndependentAtoi(std::string_view str)
119{
120    static_assert(std::is_integral_v<T>);
121    T result;
122    // Emulate atoi(...) handling of white space and leading +/-.
123    std::string_view s = util::TrimStringView(str);
124    if (!s.empty() && s[0] == '+') {
125        if (s.length() >= 2 && s[1] == '-') {
126            return 0;
127        }
128        s = s.substr(1);
129    }
130    auto [_, error_condition] = std::from_chars(s.data(), s.data() + s.size(), result);
131    if (error_condition == std::errc::result_out_of_range) {
132        if (s.length() >= 1 && s[0] == '-') {
133            // Saturate underflow, per strtoll's behavior.
134            return std::numeric_limits<T>::min();
135        } else {
136            // Saturate overflow, per strtoll's behavior.
137            return std::numeric_limits<T>::max();
138        }
139    } else if (error_condition != std::errc{}) {
140        return 0;
141    }
142    return result;
143}
144
145/**
146 * Tests if the given character is a decimal digit.
147 * @param[in] c     character to test
148 * @return          true if the argument is a decimal digit; otherwise false.
149 */
150constexpr bool IsDigit(char c)
151{
152    return c >= '0' && c <= '9';
153}
154
155/**
156 * Tests if the given character is a whitespace character. The whitespace characters
157 * are: space, form-feed ('\f'), newline ('\n'), carriage return ('\r'), horizontal
158 * tab ('\t'), and vertical tab ('\v').
159 *
160 * This function is locale independent. Under the C locale this function gives the
161 * same result as std::isspace.
162 *
163 * @param[in] c     character to test
164 * @return          true if the argument is a whitespace character; otherwise false
165 */
166constexpr inline bool IsSpace(char c) noexcept {
167    return c == ' ' || c == '\f' || c == '\n' || c == '\r' || c == '\t' || c == '\v';
168}
169
170/**
171 * Convert string to integral type T. Leading whitespace, a leading +, or any
172 * trailing character fail the parsing. The required format expressed as regex
173 * is `-?[0-9]+` by default (or `-?[0-9a-fA-F]+` if base = 16).
174 * The minus sign is only permitted for signed integer types.
175 *
176 * @returns std::nullopt if the entire string could not be parsed, or if the
177 *   parsed value is not in the range representable by the type T.
178 */
179template <typename T>
180std::optional<T> ToIntegral(std::string_view str, size_t base = 10)
181{
182    static_assert(std::is_integral_v<T>);
183    T result;
184    const auto [first_nonmatching, error_condition] = std::from_chars(str.data(), str.data() + str.size(), result, base);
185    if (first_nonmatching != str.data() + str.size() || error_condition != std::errc{}) {
186        return std::nullopt;
187    }
188    return result;
189}
190
191/**
192 * Format a paragraph of text to a fixed width, adding spaces for
193 * indentation to any added line.
194 */
195std::string FormatParagraph(std::string_view in, size_t width = 79, size_t indent = 0);
196
197/**
198 * Timing-attack-resistant comparison.
199 * Takes time proportional to length
200 * of first argument.
201 */
202template <typename T>
203bool TimingResistantEqual(const T& a, const T& b)
204{
205    if (b.size() == 0) return a.size() == 0;
206    size_t accumulator = a.size() ^ b.size();
207    for (size_t i = 0; i < a.size(); i++)
208        accumulator |= size_t(a[i] ^ b[i%b.size()]);
209    return accumulator == 0;
210}
211
212/** Parse number as fixed point according to JSON number syntax.
213 * @returns true on success, false on error.
214 * @note The result must be in the range (-10^18,10^18), otherwise an overflow error will trigger.
215 */
216[[nodiscard]] bool ParseFixedPoint(std::string_view, int decimals, int64_t *amount_out);
217
218/** Convert from one power-of-2 number base to another. */
219template <int frombits, int tobits, bool pad, typename O, typename It, typename I = std::identity>
220bool ConvertBits(O outfn, It it, It end, I infn = {})
221{
222    size_t acc = 0;
223    size_t bits = 0;
224    constexpr size_t maxv = (1 << tobits) - 1;
225    constexpr size_t max_acc = (1 << (frombits + tobits - 1)) - 1;
226    while (it != end) {
227        int v = infn(*it);
228        if (v < 0) return false;
229        acc = ((acc << frombits) | v) & max_acc;
230        bits += frombits;
231        while (bits >= tobits) {
232            bits -= tobits;
233            outfn((acc >> bits) & maxv);
234        }
235        ++it;
236    }
237    if (pad) {
238        if (bits) outfn((acc << (tobits - bits)) & maxv);
239    } else if (bits >= frombits || ((acc << (tobits - bits)) & maxv)) {
240        return false;
241    }
242    return true;
243}
244
245/**
246 * Converts the given character to its lowercase equivalent.
247 * This function is locale independent. It only converts uppercase
248 * characters in the standard 7-bit ASCII range.
249 * This is a feature, not a limitation.
250 *