File: get.c 1 /* 2 The MIT License (MIT) 3 4 Copyright © 2020-2025 pacman64 5 6 Permission is hereby granted, free of charge, to any person obtaining a copy of 7 this software and associated documentation files (the “Software”), to deal 8 in the Software without restriction, including without limitation the rights to 9 use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies 10 of the Software, and to permit persons to whom the Software is furnished to do 11 so, subject to the following conditions: 12 13 The above copyright notice and this permission notice shall be included in all 14 copies or substantial portions of the Software. 15 16 THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR 17 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, 18 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE 19 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER 20 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, 21 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE 22 SOFTWARE. 23 */ 24 25 /* 26 You can build this command-line app by running 27 28 sudo apt install libcurl4-openssl-dev 29 cc -Wall -s -O3 -march=native -mtune=native -flto -o ./get ./get.c -l curl 30 */ 31 32 #include <stdbool.h> 33 #include <stdint.h> 34 #include <stdio.h> 35 #include <stdlib.h> 36 #include <string.h> 37 #include <sys/stat.h> 38 39 #ifdef _WIN32 40 #include <fcntl.h> 41 #include <windows.h> 42 #endif 43 44 #include <curl/curl.h> 45 46 #ifdef RED_ERRORS 47 #define ERROR_STYLE "\x1b[38;2;204;0;0m" 48 #ifdef __APPLE__ 49 #define ERROR_STYLE "\x1b[31m" 50 #endif 51 #define RESET_STYLE "\x1b[0m" 52 #else 53 #define ERROR_STYLE 54 #define RESET_STYLE 55 #endif 56 57 #define ERROR_LINE(MSG) (ERROR_STYLE MSG RESET_STYLE "\n") 58 59 const char* info = "" 60 "get [options...] [filenames/data-URIs/URIs...]\n" 61 "\n" 62 "\n" 63 "Load bytes from all the named sources given, be them files, base64-encoded\n" 64 "URIs, or HTTP/HTTPS URIs. The name `-` stands for the standard input. When\n" 65 "no names are given, the standard input is used by default.\n" 66 "\n" 67 "The help option is `-h`, `--h`, `-help`, or `--help`.\n" 68 ""; 69 70 void handle_reader(FILE* w, FILE* r) { 71 const size_t bufsize = 32 * 1024; 72 unsigned char buf[bufsize]; 73 74 while (!feof(w)) { 75 size_t len = fread(&buf, sizeof(buf[0]), sizeof(buf), r); 76 if (len < 1) { 77 break; 78 } 79 fwrite(&buf, 1, len, w); 80 } 81 82 fflush(w); 83 } 84 85 // handle_file handles data from the filename given; returns false only when 86 // the file can't be opened 87 bool handle_file(FILE* w, const char* path) { 88 FILE* f = fopen(path, "rb"); 89 if (f == NULL) { 90 fprintf(stderr, ERROR_LINE("can't open file named '%s'"), path); 91 return false; 92 } 93 94 handle_reader(w, f); 95 fclose(f); 96 return true; 97 } 98 99 bool fetch(FILE* w, CURL* curl, const char* uri) { 100 curl_easy_setopt(curl, CURLOPT_URL, uri); 101 102 CURLcode code = curl_easy_perform(curl); 103 if (code != CURLE_OK) { 104 fputc('\n', w); 105 const char* msg = curl_easy_strerror(code); 106 fprintf(stderr, ERROR_LINE("%s: %s"), uri, msg); 107 return false; 108 } 109 110 fflush(w); 111 return true; 112 } 113 114 // INVALID signals an input byte isn't allowed in a base64 stream 115 #define INVALID 0xff 116 117 const unsigned char base64_rev_lookup[256] = { 118 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 119 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 120 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 121 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 122 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 123 0xff, 0xff, 0xff, 0x3e, 0xff, 0xff, 0xff, 0x3f, 124 0x34, 0x35, 0x36, 0x37, 0x38, 0x39, 0x3a, 0x3b, 125 0x3c, 0x3d, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 126 0xff, 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 127 0x07, 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 128 0x0f, 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 129 0x17, 0x18, 0x19, 0xff, 0xff, 0xff, 0xff, 0xff, 130 0xff, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, 0x20, 131 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, 0x28, 132 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f, 0x30, 133 0x31, 0x32, 0x33, 0xff, 0xff, 0xff, 0xff, 0xff, 134 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 135 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 136 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 137 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 138 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 139 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 140 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 141 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 142 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 143 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 144 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 145 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 146 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 147 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 148 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 149 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 150 }; 151 152 // unsigned char rev_lookup_byte(unsigned char b) { 153 // if ('A' <= b && b <= 'Z') { 154 // return b - 'A'; 155 // } 156 // if ('a' <= b && b <= 'z') { 157 // return (b - 'a') + 26; 158 // } 159 // if ('0' <= b && b <= '9') { 160 // return (b - '0') + 52; 161 // } 162 // 163 // switch (b) { 164 // case '+': 165 // return 62; 166 // case '/': 167 // return 63; 168 // default: 169 // return INVALID; 170 // } 171 // } 172 173 static inline unsigned char rev_lookup_byte(unsigned char b) { 174 return base64_rev_lookup[b]; 175 } 176 177 void show_invalid_byte(unsigned char b, size_t line, size_t pos) { 178 const char* msg = "invalid base64 data"; 179 const char* fmt = ERROR_LINE("%s (byte %d, line: %ld, pos: %ld)"); 180 fprintf(stderr, fmt, msg, b, (long)line, (long)pos); 181 } 182 183 bool match_lead(unsigned char* buf, size_t n, char* to) { 184 for (; n > 0 && *to != 0; buf++, to++, n--) { 185 if (*buf != *to) { 186 return false; 187 } 188 } 189 return true; 190 } 191 192 size_t skip_data_uri(unsigned char* buf, size_t n) { 193 for (size_t i = 0; i < n; i++) { 194 if (match_lead(buf + i, n - i, ";base64,")) { 195 return i + (sizeof(";base64,") - 1); 196 } 197 } 198 return 0; 199 } 200 201 bool handle_data_uri(FILE* w, char* uri) { 202 uint64_t line = 1; 203 uint64_t pos = 1; 204 uint64_t payload = 0; 205 uint64_t padding = 0; 206 207 unsigned char quad[4]; 208 quad[0] = 0; 209 quad[1] = 0; 210 quad[2] = 0; 211 quad[3] = 0; 212 213 size_t n = strlen(uri); 214 unsigned char* chunk = (unsigned char*)uri; 215 216 // skip leading utf-8 byte-order-mark bytes, if present 217 if (n >= 3 && match_lead(chunk, n, "\xef\xbb\xbf")) { 218 chunk += 3; 219 n -= 3; 220 } 221 222 // skip leading data-URI prelude, if present 223 if (match_lead(chunk, n, "data:")) { 224 const int skip = skip_data_uri(chunk, n); 225 chunk += skip; 226 n -= skip; 227 } 228 229 for (size_t i = 0; i < n; i++) { 230 const unsigned char v = chunk[i]; 231 232 // ignore carriage-returns to support CRLF lines 233 if (v == '\r') { 234 continue; 235 } 236 237 // base64 streams can span multiple lines 238 if (v == '\n') { 239 line++; 240 pos = 1; 241 continue; 242 } 243 244 pos++; 245 246 if (v == '=') { 247 padding++; 248 continue; 249 } 250 251 if (padding > 0 && v != '=') { 252 fputc('\n', w); 253 fflush(w); 254 255 const char* msg = "equal signs are only valid at the end"; 256 const char* fmt = ERROR_LINE("%s (line %ld, pos %ld)"); 257 fprintf(stderr, fmt, msg, (long)line, (long)pos); 258 return false; 259 } 260 261 unsigned char b = rev_lookup_byte(v); 262 263 if (b == INVALID) { 264 show_invalid_byte(v, line, pos); 265 return false; 266 } 267 268 const size_t step = payload % 4; 269 quad[step] = b; 270 payload++; 271 272 if (step == 3) { 273 // 01234567 01234567 01234567 01234567 274 // 00000000 11111111 22222222 33333333 275 // xx000000 xx001111 xx111122 xx222222 276 fputc((quad[0] << 2) | (quad[1] >> 4), w); 277 fputc((quad[1] << 4) | (quad[2] >> 2), w); 278 fputc((quad[2] << 6) | (quad[3] >> 0), w); 279 } 280 } 281 282 // try to be resilient to missing trailing/padding equals 283 // if (padding == 0 && payload > 0) { 284 // padding = 4 - (payload % 4); 285 // } 286 287 // don't forget unemitted trailing bytes, if any 288 switch (padding) { 289 case 1: 290 fputc((quad[0] << 2) | (quad[1] >> 4), w); 291 fputc((quad[1] << 4) | (quad[2] >> 2), w); 292 break; 293 case 2: 294 fputc((quad[0] << 2) | (quad[1] >> 4), w); 295 break; 296 } 297 298 fflush(w); 299 return true; 300 } 301 302 bool starts_with(const char* s, const char* prefix) { 303 for (size_t i = 0; prefix[i] != 0; i++) { 304 if (s[i] == 0) { 305 return prefix[i] == 0; 306 } 307 if (s[i] != prefix[i]) { 308 return false; 309 } 310 } 311 312 return true; 313 } 314 315 bool is_file(const char* path) { 316 struct stat meta; 317 return stat(path, &meta) == 0; 318 } 319 320 bool seems_curlable(const char* s) { 321 return false || 322 starts_with(s, "https://") || 323 starts_with(s, "http://") || 324 starts_with(s, "ftp://") || 325 starts_with(s, "ftps://") || 326 starts_with(s, "gopher://") || 327 starts_with(s, "gophers://") || 328 starts_with(s, "rtmp://") || 329 starts_with(s, "rtsp://") || 330 starts_with(s, "scp://") || 331 starts_with(s, "sftp://") || 332 starts_with(s, "smb://") || 333 starts_with(s, "smbs://") || 334 starts_with(s, "telnet://") || 335 starts_with(s, "tftp://") || 336 false; 337 } 338 339 // run returns the number of errors 340 int run(int argc, char** argv, FILE* w) { 341 size_t dashes = 0; 342 CURL* curl = NULL; 343 344 for (int i = 1; i < argc; i++) { 345 if (argv[i][0] == '-' && argv[i][1] == 0) { 346 dashes++; 347 continue; 348 } 349 350 if (curl != NULL || seems_curlable(argv[i])) { 351 curl = curl_easy_init(); 352 if (curl == NULL) { 353 fprintf(stderr, ERROR_LINE("can't initialize libcurl")); 354 return 1; 355 } 356 357 curl_easy_setopt(curl, CURLOPT_FOLLOWLOCATION, 1L); 358 curl_easy_setopt(curl, CURLOPT_WRITEDATA, w); 359 } 360 } 361 362 if (dashes > 1) { 363 const char* msg = "can't use the standard input (dash) more than once"; 364 fprintf(stderr, ERROR_LINE("%s"), msg); 365 return 1; 366 } 367 368 // use stdin when not given any filepaths 369 if (argc <= 1) { 370 handle_reader(w, stdin); 371 return 0; 372 } 373 374 for (int i = 1; i < argc && !feof(w); i++) { 375 if (is_file(argv[i])) { 376 if (!handle_file(w, argv[i])) { 377 return 1; 378 } 379 continue; 380 } 381 382 if (argv[i][0] == '-' && argv[i][1] == 0) { 383 handle_reader(w, stdin); 384 continue; 385 } 386 387 if (starts_with(argv[i], "data:")) { 388 if (!handle_data_uri(w, argv[i])) { 389 return 1; 390 } 391 continue; 392 } 393 394 if (starts_with(argv[i], "file://")) { 395 if (!handle_file(w, argv[i] + sizeof("file://") - 1)) { 396 return 1; 397 } 398 continue; 399 } 400 401 if (seems_curlable(argv[i])) { 402 if (!fetch(w, curl, argv[i])) { 403 return 1; 404 } 405 continue; 406 } 407 408 if (!handle_file(w, argv[i])) { 409 return 1; 410 } 411 } 412 413 if (curl != NULL) { 414 curl_easy_cleanup(curl); 415 } 416 return 0; 417 } 418 419 int main(int argc, char** argv) { 420 #ifdef _WIN32 421 setmode(fileno(stdin), O_BINARY); 422 // ensure output lines end in LF instead of CRLF on windows 423 setmode(fileno(stdout), O_BINARY); 424 setmode(fileno(stderr), O_BINARY); 425 #endif 426 427 if (argc > 1) { 428 if ( 429 strcmp(argv[1], "-h") == 0 || 430 strcmp(argv[1], "-help") == 0 || 431 strcmp(argv[1], "--h") == 0 || 432 strcmp(argv[1], "--help") == 0 433 ) { 434 fprintf(stdout, "%s", info); 435 return 0; 436 } 437 } 438 439 // setvbuf(stdin, NULL, _IOFBF, 0); 440 setvbuf(stdout, NULL, _IOFBF, 0); 441 return run(argc, argv, stdout) == 0 ? 0 : 1; 442 }