1 /*********************************************************************\ 2 3 MODULE NAME: b64.c 4 5 AUTHOR: Bob Trower 08/04/01 6 7 PROJECT: Crypt Data Packaging 8 9 COPYRIGHT: Copyright (c) Trantor Standard Systems Inc., 2001 10 11 NOTE: This source code may be used as you wish, subject to 12 the MIT license. See the LICENCE section below. 13 14 DESCRIPTION: 15 This little utility implements the Base64 16 Content-Transfer-Encoding standard described in 17 RFC1113 (http://www.faqs.org/rfcs/rfc1113.html). 18 19 This is the coding scheme used by MIME to allow 20 binary data to be transferred by SMTP mail. 21 22 Groups of 3 bytes from a binary stream are coded as 23 groups of 4 bytes in a text stream. 24 25 The input stream is 'padded' with zeros to create 26 an input that is an even multiple of 3. 27 28 A special character ('=') is used to denote padding so 29 that the stream can be decoded back to its exact size. 30 31 Encoded output is formatted in lines which should 32 be a maximum of 72 characters to conform to the 33 specification. This program defaults to 72 characters, 34 but will allow more or less through the use of a 35 switch. The program enforces a minimum line size 36 of 4 characters. 37 38 Example encoding: 39 40 The stream 'ABCD' is 32 bits long. It is mapped as 41 follows: 42 43 ABCD 44 45 A (65) B (66) C (67) D (68) (None) (None) 46 01000001 01000010 01000011 01000100 47 48 16 (Q) 20 (U) 9 (J) 3 (D) 17 (R) 0 (A) NA (=) NA (=) 49 010000 010100 001001 000011 010001 000000 000000 000000 50 51 52 QUJDRA== 53 54 Decoding is the process in reverse. A 'decode' lookup 55 table has been created to avoid string scans. 56 57 DESIGN GOALS: Specifically: 58 Code is a stand-alone utility to perform base64 59 encoding/decoding. It should be genuinely useful 60 when the need arises and it meets a need that is 61 likely to occur for some users. 62 Code acts as sample code to show the author's 63 design and coding style. 64 65 Generally: 66 This program is designed to survive: 67 Everything you need is in a single source file. 68 It compiles cleanly using a vanilla ANSI C compiler. 69 It does its job correctly with a minimum of fuss. 70 The code is not overly clever, not overly simplistic 71 and not overly verbose. 72 Access is 'cut and paste' from a web page. 73 Terms of use are reasonable. 74 75 VALIDATION: Non-trivial code is never without errors. This 76 file likely has some problems, since it has only 77 been tested by the author. It is expected with most 78 source code that there is a period of 'burn-in' when 79 problems are identified and corrected. That being 80 said, it is possible to have 'reasonably correct' 81 code by following a regime of unit test that covers 82 the most likely cases and regression testing prior 83 to release. This has been done with this code and 84 it has a good probability of performing as expected. 85 86 Unit Test Cases: 87 88 case 0:empty file: 89 CASE0.DAT -> -> 90 (Zero length target file created 91 on both encode and decode.) 92 93 case 1:One input character: 94 CASE1.DAT A -> QQ== -> A 95 96 case 2:Two input characters: 97 CASE2.DAT AB -> QUJD -> AB 98 99 case 3:Three input characters: 100 CASE3.DAT ABC -> QUJD -> ABC 101 102 case 4:Four input characters: 103 case4.dat ABCD -> QUJDRA== -> ABCD 104 105 case 5:All chars from 0 to ff, linesize set to 50: 106 107 AAECAwQFBgcICQoLDA0ODxAREhMUFRYXGBkaGxwdHh8gISIj 108 JCUmJygpKissLS4vMDEyMzQ1Njc4OTo7PD0+P0BBQkNERUZH 109 SElKS0xNTk9QUVJTVFVWV1hZWltcXV5fYGFiY2RlZmdoaWpr 110 bG1ub3BxcnN0dXZ3eHl6e3x9fn+AgYKDhIWGh4iJiouMjY6P 111 kJGSk5SVlpeYmZqbnJ2en6ChoqOkpaanqKmqq6ytrq+wsbKz 112 tLW2t7i5uru8vb6/wMHCw8TFxsfIycrLzM3Oz9DR0tPU1dbX 113 2Nna29zd3t/g4eLj5OXm5+jp6uvs7e7v8PHy8/T19vf4+fr7 114 /P3+/w== 115 116 case 6:Mime Block from e-mail: 117 (Data same as test case 5) 118 119 case 7: Large files: 120 Tested 28 MB file in/out. 121 122 case 8: Random Binary Integrity: 123 This binary program (b64.exe) was encoded to base64, 124 back to binary and then executed. 125 126 case 9 Stress: 127 All files in a working directory encoded/decoded 128 and compared with file comparison utility to 129 ensure that multiple runs do not cause problems 130 such as exhausting file handles, tmp storage, etc. 131 132 ------------- 133 134 Syntax, operation and failure: 135 All options/switches tested. Performs as 136 expected. 137 138 case 10: 139 No Args -- Shows Usage Screen 140 Return Code 1 (Invalid Syntax) 141 case 11: 142 One Arg (invalid) -- Shows Usage Screen 143 Return Code 1 (Invalid Syntax) 144 case 12: 145 One Arg Help (-?) -- Shows detailed Usage Screen. 146 Return Code 0 (Success -- help request is valid). 147 case 13: 148 One Arg Help (-h) -- Shows detailed Usage Screen. 149 Return Code 0 (Success -- help request is valid). 150 case 14: 151 One Arg (valid) -- Uses stdin/stdout (filter) 152 Return Code 0 (Sucess) 153 case 15: 154 Two Args (invalid file) -- shows system error. 155 Return Code 2 (File Error) 156 case 16: 157 Encode non-existent file -- shows system error. 158 Return Code 2 (File Error) 159 case 17: 160 Out of disk space -- shows system error. 161 Return Code 3 (File I/O Error) 162 163 ------------- 164 165 Compile/Regression test: 166 gcc compiled binary under Cygwin 167 Microsoft Visual Studio under Windows 2000 168 Microsoft Version 6.0 C under Windows 2000 169 170 DEPENDENCIES: None 171 172 LICENCE: Copyright (c) 2001 Bob Trower, Trantor Standard Systems Inc. 173 174 Permission is hereby granted, free of charge, to any person 175 obtaining a copy of this software and associated 176 documentation files (the "Software"), to deal in the 177 Software without restriction, including without limitation 178 the rights to use, copy, modify, merge, publish, distribute, 179 sublicense, and/or sell copies of the Software, and to 180 permit persons to whom the Software is furnished to do so, 181 subject to the following conditions: 182 183 The above copyright notice and this permission notice shall 184 be included in all copies or substantial portions of the 185 Software. 186 187 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY 188 KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE 189 WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR 190 PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS 191 OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR 192 OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR 193 OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE 194 SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. 195 196 VERSION HISTORY: 197 Bob Trower 08/04/01 -- Create Version 0.00.00B 198 199 \******************************************************************* */ 200 201 #include <inttypes.h> 202 #include <stdio.h> 203 #include <stdlib.h> 204 205 #include "b64.h" 206 207 /* 208 ** Translation Table as described in RFC1113 209 */ 210 static const char cb64[] = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/"; 211 212 /* 213 ** Translation Table to decode (created by author) 214 */ 215 static const char cd64[] = "|$$$}rstuvwxyz{$$$$$$$>?@ABCDEFGHIJKLMNOPQRSTUVW$$$$$$XYZ[\\]^_`abcdefghijklmnopq"; 216 217 /* 218 ** encodeblock 219 ** 220 ** encode 3 8-bit binary bytes as 4 '6-bit' characters 221 */ 222 static void 223 encodeblock(uint8_t *wordin, uint8_t *wordout, int wordlen) 224 { 225 wordout[0] = cb64[(unsigned)wordin[0] >> 2]; 226 wordout[1] = cb64[((unsigned)(wordin[0] & 0x03) << 4) | ((unsigned)(wordin[1] & 0xf0) >> 4)]; 227 wordout[2] = (uint8_t)(wordlen > 1) ? 228 cb64[((unsigned)(wordin[1] & 0x0f) << 2) | ((unsigned)(wordin[2] & 0xc0) >> 6)] : '='; 229 wordout[3] = (uint8_t)(wordlen > 2) ? cb64[wordin[2] & 0x3f] : '='; 230 } 231 232 /* 233 ** encode 234 ** 235 ** base64 encode a stream adding padding and line breaks as per spec. 236 */ 237 int 238 b64encode(const char *in, const size_t insize, void *vp, size_t outsize, int linesize) 239 { 240 const char *inp; 241 unsigned i; 242 uint8_t wordout[4]; 243 uint8_t wordin[3]; 244 char *out = vp; 245 char *outp; 246 int blocksout; 247 int wordlen; 248 249 if (in == NULL || vp == NULL) { 250 return 0; 251 } 252 wordlen = 0; 253 for (blocksout = 0, inp = in, outp = out; (size_t)(inp - in) < insize && (size_t)(outp - out) < outsize;) { 254 for (wordlen = 0, i = 0; i < sizeof(wordin); i++) { 255 wordin[i] = (uint8_t) *inp++; 256 if ((size_t)(inp - in) <= insize) { 257 wordlen++; 258 } else { 259 wordin[i] = 0x0; 260 } 261 } 262 if (wordlen > 0) { 263 encodeblock(wordin, wordout, wordlen); 264 for (i = 0; i < sizeof(wordout) ; i++) { 265 *outp++ = wordout[i]; 266 } 267 blocksout++; 268 } 269 if (linesize > 0) { 270 if (blocksout >= (int)(linesize / sizeof(wordout)) || 271 (size_t)(inp - in) >= insize) { 272 if (blocksout) { 273 *outp++ = '\r'; 274 *outp++ = '\n'; 275 } 276 blocksout = 0; 277 } 278 } 279 } 280 return (int)(outp - out); 281 } 282 283 /* 284 ** decodeblock 285 ** 286 ** decode 4 '6-bit' characters into 3 8-bit binary bytes 287 */ 288 static void 289 decodeblock(uint8_t wordin[4], uint8_t wordout[3]) 290 { 291 wordout[0] = (uint8_t) ((unsigned)wordin[0] << 2 | (unsigned)wordin[1] >> 4); 292 wordout[1] = (uint8_t) ((unsigned)wordin[1] << 4 | (unsigned)wordin[2] >> 2); 293 wordout[2] = (uint8_t) (((wordin[2] << 6) & 0xc0) | wordin[3]); 294 } 295 296 /* 297 ** decode 298 ** 299 ** decode a base64 encoded stream discarding padding, line breaks and noise 300 */ 301 int 302 b64decode(const char *in, const size_t insize, void *vp, size_t outsize) 303 { 304 const char *inp; 305 unsigned wordlen; 306 unsigned i; 307 uint8_t wordout[3]; 308 uint8_t wordin[4]; 309 uint8_t v; 310 char *out = vp; 311 char *outp; 312 313 if (in == NULL || vp == NULL) { 314 return 0; 315 } 316 for (inp = in, outp = out ; (size_t)(inp - in) < insize && (size_t)(outp - out) < outsize ; ) { 317 for (wordlen = 0, i = 0 ; i < sizeof(wordin) && (size_t)(inp - in) < insize ; i++) { 318 /* get a single character */ 319 for (v = 0; (size_t)(inp - in) <= insize && v == 0 ; ) { 320 if (*inp == '\r' && *(inp + 1) == '\n') { 321 inp += 2; 322 } else { 323 v = (uint8_t) *inp++; 324 v = (uint8_t) ((v < 43 || v > 122) ? 0 : cd64[v - 43]); 325 if (v) { 326 v = (uint8_t) ((v == '$') ? 0 : v - 61); 327 } 328 } 329 } 330 /* perhaps 0x0 pad */ 331 if ((size_t)(inp - in) <= insize) { 332 wordlen += 1; 333 if (v) { 334 wordin[i] = (uint8_t) (v - 1); 335 } 336 } else { 337 wordin[i] = 0x0; 338 } 339 } 340 if (wordlen > 0) { 341 decodeblock(wordin, wordout); 342 for (i = 0; i < wordlen - 1 ; i++) { 343 *outp++ = wordout[i]; 344 } 345 } 346 } 347 return (int)(outp - out); 348 } 349 350 /* return the encoded size for n bytes input */ 351 int 352 b64_encsize(unsigned n) 353 { 354 return ((4 * n) / 3) + 4; 355 } 356