1/*
2   Copyright (C) 2000-2001 Dawit Alemayehu <adawit@kde.org>
3   Copyright (C) 2006 Alexey Proskuryakov <ap@webkit.org>
4   Copyright (C) 2007, 2008 Apple Inc. All rights reserved.
5   Copyright (C) 2010 Patrick Gansterer <paroga@paroga.com>
6
7   This program is free software; you can redistribute it and/or modify
8   it under the terms of the GNU Lesser General Public License (LGPL)
9   version 2 as published by the Free Software Foundation.
10
11   This program is distributed in the hope that it will be useful,
12   but WITHOUT ANY WARRANTY; without even the implied warranty of
13   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
14   GNU General Public License for more details.
15
16   You should have received a copy of the GNU Library General Public
17   License along with this program; if not, write to the Free Software
18   Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
19
20   This code is based on the java implementation in HTTPClient
21   package by Ronald Tschalär Copyright (C) 1996-1999.
22*/
23
24#include "config.h"
25#include "Base64.h"
26
27#include <limits.h>
28#include "wtf/StringExtras.h"
29
30namespace WTF {
31
32static const char base64EncMap[64] = {
33    0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, 0x48,
34    0x49, 0x4A, 0x4B, 0x4C, 0x4D, 0x4E, 0x4F, 0x50,
35    0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, 0x58,
36    0x59, 0x5A, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66,
37    0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E,
38    0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76,
39    0x77, 0x78, 0x79, 0x7A, 0x30, 0x31, 0x32, 0x33,
40    0x34, 0x35, 0x36, 0x37, 0x38, 0x39, 0x2B, 0x2F
41};
42
43static const char base64DecMap[128] = {
44    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
45    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
46    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
47    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
48    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
49    0x00, 0x00, 0x00, 0x3E, 0x00, 0x00, 0x00, 0x3F,
50    0x34, 0x35, 0x36, 0x37, 0x38, 0x39, 0x3A, 0x3B,
51    0x3C, 0x3D, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
52    0x00, 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06,
53    0x07, 0x08, 0x09, 0x0A, 0x0B, 0x0C, 0x0D, 0x0E,
54    0x0F, 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16,
55    0x17, 0x18, 0x19, 0x00, 0x00, 0x00, 0x00, 0x00,
56    0x00, 0x1A, 0x1B, 0x1C, 0x1D, 0x1E, 0x1F, 0x20,
57    0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, 0x28,
58    0x29, 0x2A, 0x2B, 0x2C, 0x2D, 0x2E, 0x2F, 0x30,
59    0x31, 0x32, 0x33, 0x00, 0x00, 0x00, 0x00, 0x00
60};
61
62String base64Encode(const char* data, unsigned length, Base64EncodePolicy policy)
63{
64    Vector<char> result;
65    base64Encode(data, length, result, policy);
66    return String(result.data(), result.size());
67}
68
69void base64Encode(const char* data, unsigned len, Vector<char>& out, Base64EncodePolicy policy)
70{
71    out.clear();
72    if (!len)
73        return;
74
75    // If the input string is pathologically large, just return nothing.
76    // Note: Keep this in sync with the "outLength" computation below.
77    // Rather than being perfectly precise, this is a bit conservative.
78    const unsigned maxInputBufferSize = UINT_MAX / 77 * 76 / 4 * 3 - 2;
79    if (len > maxInputBufferSize)
80        return;
81
82    unsigned sidx = 0;
83    unsigned didx = 0;
84
85    unsigned outLength = ((len + 2) / 3) * 4;
86
87    // Deal with the 76 character per line limit specified in RFC 2045.
88    bool insertLFs = (policy == Base64InsertLFs && outLength > 76);
89    if (insertLFs)
90        outLength += ((outLength - 1) / 76);
91
92    int count = 0;
93    out.grow(outLength);
94
95    // 3-byte to 4-byte conversion + 0-63 to ascii printable conversion
96    if (len > 1) {
97        while (sidx < len - 2) {
98            if (insertLFs) {
99                if (count && !(count % 76))
100                    out[didx++] = '\n';
101                count += 4;
102            }
103            out[didx++] = base64EncMap[(data[sidx] >> 2) & 077];
104            out[didx++] = base64EncMap[((data[sidx + 1] >> 4) & 017) | ((data[sidx] << 4) & 077)];
105            out[didx++] = base64EncMap[((data[sidx + 2] >> 6) & 003) | ((data[sidx + 1] << 2) & 077)];
106            out[didx++] = base64EncMap[data[sidx + 2] & 077];
107            sidx += 3;
108        }
109    }
110
111    if (sidx < len) {
112        if (insertLFs && (count > 0) && !(count % 76))
113           out[didx++] = '\n';
114
115        out[didx++] = base64EncMap[(data[sidx] >> 2) & 077];
116        if (sidx < len - 1) {
117            out[didx++] = base64EncMap[((data[sidx + 1] >> 4) & 017) | ((data[sidx] << 4) & 077)];
118            out[didx++] = base64EncMap[(data[sidx + 1] << 2) & 077];
119        } else
120            out[didx++] = base64EncMap[(data[sidx] << 4) & 077];
121    }
122
123    // Add padding
124    while (didx < out.size()) {
125        out[didx] = '=';
126        ++didx;
127    }
128}
129
130bool base64Decode(const Vector<char>& in, Vector<char>& out, CharacterMatchFunctionPtr shouldIgnoreCharacter, Base64DecodePolicy policy)
131{
132    out.clear();
133
134    // If the input string is pathologically large, just return nothing.
135    if (in.size() > UINT_MAX)
136        return false;
137
138    return base64Decode(in.data(), in.size(), out, shouldIgnoreCharacter, policy);
139}
140
141template<typename T>
142static inline bool base64DecodeInternal(const T* data, unsigned length, Vector<char>& out, CharacterMatchFunctionPtr shouldIgnoreCharacter, Base64DecodePolicy policy)
143{
144    out.clear();
145    if (!length)
146        return true;
147
148    out.grow(length);
149
150    unsigned equalsSignCount = 0;
151    unsigned outLength = 0;
152    bool hadError = false;
153    for (unsigned idx = 0; idx < length; ++idx) {
154        unsigned ch = data[idx];
155        if (ch == '=') {
156            ++equalsSignCount;
157            // There should never be more than 2 padding characters.
158            if (policy == Base64ValidatePadding && equalsSignCount > 2) {
159                hadError = true;
160                break;
161            }
162        } else if (('0' <= ch && ch <= '9') || ('A' <= ch && ch <= 'Z') || ('a' <= ch && ch <= 'z') || ch == '+' || ch == '/') {
163            if (equalsSignCount) {
164                hadError = true;
165                break;
166            }
167            out[outLength++] = base64DecMap[ch];
168        } else if (!shouldIgnoreCharacter || !shouldIgnoreCharacter(ch)) {
169            hadError = true;
170            break;
171        }
172    }
173
174    if (outLength < out.size())
175        out.shrink(outLength);
176
177    if (hadError)
178        return false;
179
180    if (!outLength)
181        return !equalsSignCount;
182
183    // There should be no padding if length is a multiple of 4.
184    // We use (outLength + equalsSignCount) instead of length because we don't want to account for ignored characters.
185    if (policy == Base64ValidatePadding && equalsSignCount && (outLength + equalsSignCount) % 4)
186        return false;
187
188    // Valid data is (n * 4 + [0,2,3]) characters long.
189    if ((outLength % 4) == 1)
190        return false;
191
192    // 4-byte to 3-byte conversion
193    outLength -= (outLength + 3) / 4;
194    if (!outLength)
195        return false;
196
197    unsigned sidx = 0;
198    unsigned didx = 0;
199    if (outLength > 1) {
200        while (didx < outLength - 2) {
201            out[didx] = (((out[sidx] << 2) & 255) | ((out[sidx + 1] >> 4) & 003));
202            out[didx + 1] = (((out[sidx + 1] << 4) & 255) | ((out[sidx + 2] >> 2) & 017));
203            out[didx + 2] = (((out[sidx + 2] << 6) & 255) | (out[sidx + 3] & 077));
204            sidx += 4;
205            didx += 3;
206        }
207    }
208
209    if (didx < outLength)
210        out[didx] = (((out[sidx] << 2) & 255) | ((out[sidx + 1] >> 4) & 003));
211
212    if (++didx < outLength)
213        out[didx] = (((out[sidx + 1] << 4) & 255) | ((out[sidx + 2] >> 2) & 017));
214
215    if (outLength < out.size())
216        out.shrink(outLength);
217
218    return true;
219}
220
221bool base64Decode(const char* data, unsigned length, Vector<char>& out, CharacterMatchFunctionPtr shouldIgnoreCharacter, Base64DecodePolicy policy)
222{
223    return base64DecodeInternal<LChar>(reinterpret_cast<const LChar*>(data), length, out, shouldIgnoreCharacter, policy);
224}
225
226bool base64Decode(const UChar* data, unsigned length, Vector<char>& out, CharacterMatchFunctionPtr shouldIgnoreCharacter, Base64DecodePolicy policy)
227{
228    return base64DecodeInternal<UChar>(data, length, out, shouldIgnoreCharacter, policy);
229}
230
231bool base64Decode(const String& in, Vector<char>& out, CharacterMatchFunctionPtr shouldIgnoreCharacter, Base64DecodePolicy policy)
232{
233    if (in.isEmpty())
234        return base64DecodeInternal<LChar>(0, 0, out, shouldIgnoreCharacter, policy);
235    if (in.is8Bit())
236        return base64DecodeInternal<LChar>(in.characters8(), in.length(), out, shouldIgnoreCharacter, policy);
237    return base64DecodeInternal<UChar>(in.characters16(), in.length(), out, shouldIgnoreCharacter, policy);
238}
239
240} // namespace WTF
241