Satoshi Nakamoto - BTC Source Code :: base58.h

I'm just browsing the original Bitcoin source code archived at https://github.com/trottier/original-bitcoin and I have a couple questions about it so far:

1) Is it really necessary to define any function as inline? Because from my understanding, a function that is declared inline isn't guaranteed to be inline, due to today's compilers being free to choose whether or not to make a function inline (am I wrong on that point)?

2) The very first function defined in this file has a return type of string. But looking at the top of the file I don't see the string header being included, let alone any header? Also, the return type isn't prefixed with std:: and I don't see the using namespace std; statement anywhere. How is this possible?

3) Probably related to the second question, but there is also a user-defined-type being used; but again, I do not see any headers being included in that file. How is this possible?

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
// Copyright (c) 2009 Satoshi Nakamoto
// Distributed under the MIT/X11 software license, see the accompanying
// file license.txt or http://www.opensource.org/licenses/mit-license.php.


//
// Why base-58 instead of standard base-64 encoding?
// - Don't want 0OIl characters that look the same in some fonts and
//      could be used to create visually identical looking account numbers.
// - A string with non-alphanumeric characters is not as easily accepted as an account number.
// - E-mail usually won't line-break if there's no punctuation to break at.
// - Doubleclicking selects the whole number as one word if it's all alphanumeric.
//


static const char* pszBase58 = "123456789ABCDEFGHJKLMNPQRSTUVWXYZabcdefghijkmnopqrstuvwxyz";


inline string EncodeBase58(const unsigned char* pbegin, const unsigned char* pend)
{
    CAutoBN_CTX pctx;
    CBigNum bn58 = 58;
    CBigNum bn0 = 0;

    // Convert big endian data to little endian
    // Extra zero at the end make sure bignum will interpret as a positive number
    vector<unsigned char> vchTmp(pend-pbegin+1, 0);
    reverse_copy(pbegin, pend, vchTmp.begin());

    // Convert little endian data to bignum
    CBigNum bn;
    bn.setvch(vchTmp);

    // Convert bignum to string
    string str;
    str.reserve((pend - pbegin) * 138 / 100 + 1);
    CBigNum dv;
    CBigNum rem;
    while (bn > bn0)
    {
        if (!BN_div(&dv, &rem, &bn, &bn58, pctx))
            throw bignum_error("EncodeBase58 : BN_div failed");
        bn = dv;
        unsigned int c = rem.getulong();
        str += pszBase58[c];
    }

    // Leading zeroes encoded as base58 zeros
    for (const unsigned char* p = pbegin; p < pend && *p == 0; p++)
        str += pszBase58[0];

    // Convert little endian string to big endian
    reverse(str.begin(), str.end());
    return str;
}

inline string EncodeBase58(const vector<unsigned char>& vch)
{
    return EncodeBase58(&vch[0], &vch[0] + vch.size());
}

inline bool DecodeBase58(const char* psz, vector<unsigned char>& vchRet)
{
    CAutoBN_CTX pctx;
    vchRet.clear();
    CBigNum bn58 = 58;
    CBigNum bn = 0;
    CBigNum bnChar;
    while (isspace(*psz))
        psz++;

    // Convert big endian string to bignum
    for (const char* p = psz; *p; p++)
    {
        const char* p1 = strchr(pszBase58, *p);
        if (p1 == NULL)
        {
            while (isspace(*p))
                p++;
            if (*p != '\0')
                return false;
            break;
        }
        bnChar.setulong(p1 - pszBase58);
        if (!BN_mul(&bn, &bn, &bn58, pctx))
            throw bignum_error("DecodeBase58 : BN_mul failed");
        bn += bnChar;
    }

    // Get bignum as little endian data
    vector<unsigned char> vchTmp = bn.getvch();

    // Trim off sign byte if present
    if (vchTmp.size() >= 2 && vchTmp.end()[-1] == 0 && vchTmp.end()[-2] >= 0x80)
        vchTmp.erase(vchTmp.end()-1);

    // Restore leading zeros
    int nLeadingZeros = 0;
    for (const char* p = psz; *p == pszBase58[0]; p++)
        nLeadingZeros++;
    vchRet.assign(nLeadingZeros + vchTmp.size(), 0);

    // Convert little endian data to big endian
    reverse_copy(vchTmp.begin(), vchTmp.end(), vchRet.end() - vchTmp.size());
    return true;
}

inline bool DecodeBase58(const string& str, vector<unsigned char>& vchRet)
{
    return DecodeBase58(str.c_str(), vchRet);
}





inline string EncodeBase58Check(const vector<unsigned char>& vchIn)
{
    // add 4-byte hash check to the end
    vector<unsigned char> vch(vchIn);
    uint256 hash = Hash(vch.begin(), vch.end());
    vch.insert(vch.end(), (unsigned char*)&hash, (unsigned char*)&hash + 4);
    return EncodeBase58(vch);
}

inline bool DecodeBase58Check(const char* psz, vector<unsigned char>& vchRet)
{
    if (!DecodeBase58(psz, vchRet))
        return false;
    if (vchRet.size() < 4)
    {
        vchRet.clear();
        return false;
    }
    uint256 hash = Hash(vchRet.begin(), vchRet.end()-4);
    if (memcmp(&hash, &vchRet.end()[-4], 4) != 0)
    {
        vchRet.clear();
        return false;
    }
    vchRet.resize(vchRet.size()-4);
    return true;
}

inline bool DecodeBase58Check(const string& str, vector<unsigned char>& vchRet)
{
    return DecodeBase58Check(str.c_str(), vchRet);
}






static const unsigned char ADDRESSVERSION = 0;

inline string Hash160ToAddress(uint160 hash160)
{
    // add 1-byte version number to the front
    vector<unsigned char> vch(1, ADDRESSVERSION);
    vch.insert(vch.end(), UBEGIN(hash160), UEND(hash160));
    return EncodeBase58Check(vch);
}

inline bool AddressToHash160(const char* psz, uint160& hash160Ret)
{
    vector<unsigned char> vch;
    if (!DecodeBase58Check(psz, vch))
        return false;
    if (vch.empty())
        return false;
    unsigned char nVersion = vch[0];
    if (vch.size() != sizeof(hash160Ret) + 1)
        return false;
    memcpy(&hash160Ret, &vch[1], sizeof(hash160Ret));
    return (nVersion <= ADDRESSVERSION);
}

inline bool AddressToHash160(const string& str, uint160& hash160Ret)
{
    return AddressToHash160(str.c_str(), hash160Ret);
}

inline bool IsValidBitcoinAddress(const char* psz)
{
    uint160 hash160;
    return AddressToHash160(psz, hash160);
}

inline bool IsValidBitcoinAddress(const string& str)
{
    return IsValidBitcoinAddress(str.c_str());
}




inline string PubKeyToAddress(const vector<unsigned char>& vchPubKey)
{
    return Hash160ToAddress(Hash160(vchPubKey));
}
Last edited on
This file, base58.h , is mentioned only in a single place; https://github.com/trottier/original-bitcoin/blob/92ee8d9a994391d148733da77e2bbc2f4acc43cd/src/headers.h

You can see there that when the compiler sees this file, it will already have seen string.h and all sorts of other files. That's all that matters. When the compiler sees "string" above on line 19, it must have already seen a defintion or declaration of string.


> Is it really necessary to define any function as inline?

These functions are defined in a header file; this implies that they may get defined (identically) in multiple translation units (every translation unit in which this header is included). In modern C++, the semantic meaning of the keyword inline applied to a function with external linkage has become: "multiple (identical) definitions of this function is allowed, as long as it is not more than one definition per translation unit".
See: One Definition Rule http://en.cppreference.com/w/cpp/language/definition


Re. 2) and 3)

This is a classic example of how not to write a header file.
It would (should) never have passed a (mildly professional) code review.

23. Make header files self-sufficient.

Summary
Behave responsibly: Ensure that each header you write is compilable standalone, by having it include any headers its contents depend upon.

Discussion
If one header file won't work unless the file that includes it also includes another header, that's gauche and puts unnecessary burden on that header file's users.
...
Consider this technique to help enforce header self-sufficiency: In your build, compile each header in isolation and validate that there are no errors or warnings.

Alexandrescu and Sutter in 'C++ Coding Standards: 101 Rules, Guidelines, and Best Practices'
Ok good stuff.
Topic archived. No new replies allowed.