src/util/conv.c

   1 /* Conversion functions */
   2
   3 #ifdef HAVE_CONFIG_H
   4 #include "config.h"
   5 #endif
   6
   7 #include <ctype.h>
   8 #include <errno.h>
   9 #ifdef HAVE_LIMITS_H
  10 #include <limits.h>
  11 #endif
  12 #include <stdlib.h>
  13 #include <string.h>
  14
  15 #include "elinks.h"
  16
  17 #include "intl/charsets.h" /* NBSP_CHAR */
  18 #include "util/conv.h"
  19 #include "util/error.h"
  20 #include "util/string.h"
  21 #include "util/time.h"
  22
  23
  24
  25 /* This function takes string @s and stores the @number (of a result width
  26  * @width) in string format there, starting at position [*@slen]. If the number
  27  * would take more space than @width, it is truncated and only the _last_
  28  * digits of it are inserted to the string. If the number takes less space than
  29  * @width, it is padded by @fillchar from left.
  30  * @base defined which base should be used (10, 16, 8, 2, ...)
  31  * @upper selects either hexa uppercased chars or lowercased chars.
  32  *
  33  * A NUL char is always added at the end of the string. @s must point to a
  34  * sufficiently large memory space, at least *@slen + @width + 1.
  35  *
  36  * Examples:
  37  *
  38  * elinks_ulongcat(s, NULL, 12345, 4, 0, 10, 0) : s = "2345"
  39  * elinks_ulongcat(s, NULL, 255, 4, '*', 16, 1) : s = "**FF"
  40  * elinks_ulongcat(s, NULL, 123, 5, '0', 10, 0) : s = "00123"
  41  *
  42  * Note that this function exists to provide a fast and efficient, however
  43  * still quite powerful alternative to sprintf(). It is optimized for speed and
  44  * is *MUCH* faster than sprintf(). If you can use it, use it ;-). But do not
  45  * get too enthusiastic, do not use it in cases where it would break i18n.
  46  */
  47 /* The function returns 0 if OK or width needed for the whole number to fit
  48  * there, if it had to be truncated. A negative value signs an error. */
  49 int inline
  50 elinks_ulongcat(unsigned char *s, unsigned int *slen,
  51                 unsigned long number, unsigned int width,
  52                 unsigned char fillchar, unsigned int base,
  53                 unsigned int upper)
  54 {
  55         static const unsigned char unum[]= "0123456789ABCDEF";
  56         static const unsigned char lnum[]= "0123456789abcdef";
  57         const unsigned char *to_num = (upper ? unum : lnum);
  58         unsigned int start = slen ? *slen : 0;
  59         unsigned int nlen = 1; /* '0' is one char, we can't have less. */
  60         unsigned int pos = start; /* starting position of the number */
  61         unsigned long q = number;
  62         int ret = 0;
  63
  64         if (width < 1 || !s || base < 2 || base > 16) return -1;
  65
  66         /* Count the length of the number in chars. */
  67         while (q > (base - 1)) {
  68                 nlen++;
  69                 q /= base;
  70         }
  71
  72         /* If max. width attained, truncate. */
  73         if (nlen > width) {
  74                 ret = nlen;
  75                 nlen = width;
  76         }
  77
  78         if (slen) *slen += nlen;
  79
  80         /* Fill left space with fillchar. */
  81         if (fillchar) {
  82                 /* ie. width = 4 nlen = 2 -> pad = 2 */
  83                 unsigned int pad = width - nlen;
  84
  85                 if (pad > 0) {
  86                         /* Relocate the start of number. */
  87                         if (slen) *slen += pad;
  88                         pos += pad;
  89
  90                         /* Pad. */
  91                         while (pad > 0) s[--pad + start] = fillchar;
  92                 }
  93         }
  94
  95         s[pos + nlen] = '\0';
  96
  97         /* Now write number starting from end. */
  98         while (nlen > 0) {
  99                 s[--nlen + pos] = to_num[(number % base)];
 100                 number /= base;
 101         }
 102
 103         return ret;
 104 }
 105
 106 /* Similar to elinks_ulongcat() but for long number. */
 107 int inline
 108 elinks_longcat(unsigned char *s, unsigned int *slen,
 109                long number, unsigned int width,
 110                unsigned char fillchar, unsigned int base,
 111                unsigned int upper)
 112 {
 113         unsigned char *p = s;
 114
 115         if (number < 0 && width > 0) {
 116                 if (slen) p[(*slen)++] = '-';
 117                 else *(p++) = '-';
 118                 number = -number;
 119                 width--;
 120         }
 121
 122         return elinks_ulongcat(p, slen, number, width, fillchar, base, upper);
 123 }
 124
 125
 126 struct string *
 127 add_long_to_string(struct string *string, long number)
 128 {
 129         unsigned char buffer[32];
 130         int length = 0;
 131         int width;
 132
 133         assert(string);
 134         if_assert_failed { return NULL; }
 135
 136         width = longcat(buffer, &length, number, sizeof(buffer) - 1, 0);
 137         if (width < 0 || !length) return NULL;
 138
 139         return add_bytes_to_string(string, buffer, length);
 140 }
 141
 142 struct string *
 143 add_knum_to_string(struct string *string, long num)
 144 {
 145         int ret;
 146         unsigned char t[32];
 147         int tlen = 0;
 148
 149         if (num && (num / (1024 * 1024)) * (1024 * 1024) == num) {
 150                 ret = longcat(&t, &tlen, num / (1024 * 1024), sizeof(t) - 2, 0);
 151                 t[tlen++] = 'M';
 152                 t[tlen] = '\0';
 153         } else if (num && (num / 1024) * 1024 == num) {
 154                 ret = longcat(&t, &tlen, num / 1024, sizeof(t) - 2, 0);
 155                 t[tlen++] = 'k';
 156                 t[tlen] = '\0';
 157         } else {
 158                 ret = longcat(&t, &tlen, num, sizeof(t) - 1, 0);
 159         }
 160
 161         if (ret < 0 || !tlen) return NULL;
 162
 163         add_bytes_to_string(string, t, tlen);
 164
 165         return string;
 166 }
 167
 168 struct string *
 169 add_xnum_to_string(struct string *string, off_t xnum)
 170 {
 171         unsigned char suff[3] = "\0i";
 172         off_t d = -1;
 173
 174         /* XXX: I don't completely like the computation of d here. --pasky */
 175         /* Mebi (Mi), 2^20 */
 176         if (xnum >= 1024 * 1024) {
 177                 suff[0] = 'M';
 178                 d = (xnum * (int) 10 / (int) ((int) (1024 * 1024))) % 10;
 179                 xnum /= 1024*1024;
 180         /* Kibi (Ki), 2^10 */
 181         } else if (xnum >= 1024) {
 182                 suff[0] = 'K';
 183                 d = (xnum * (int) 10 / (int) 1024) % 10;
 184                 xnum /= 1024;
 185         }
 186
 187         assert(xnum == (long) xnum);
 188         add_long_to_string(string, xnum);
 189
 190         if (xnum < 10 && d != -1) {
 191                 add_char_to_string(string, '.');
 192                 add_long_to_string(string, d);
 193         }
 194         add_char_to_string(string, ' ');
 195
 196         if (suff[0]) add_to_string(string, suff);
 197         add_char_to_string(string, 'B');
 198         return string;
 199 }
 200
 201 struct string *
 202 add_duration_to_string(struct string *string, long seconds)
 203 {
 204         unsigned char q[64];
 205         int qlen = 0;
 206
 207         if (seconds < 0) seconds = 0;
 208
 209         /* Days */
 210         if (seconds >= (24 * 3600)) {
 211                 ulongcat(q, &qlen, (seconds / (24 * 3600)), 5, 0);
 212                 q[qlen++] = 'd';
 213                 q[qlen++] = ' ';
 214         }
 215
 216         /* Hours and minutes */
 217         if (seconds >= 3600) {
 218                 seconds %= (24 * 3600);
 219                 ulongcat(q, &qlen, (seconds / 3600), 4, 0);
 220                 q[qlen++] = ':';
 221                 ulongcat(q, &qlen, ((seconds / 60) % 60), 2, '0');
 222         } else {
 223                 /* Only minutes */
 224                 ulongcat(q, &qlen, (seconds / 60), 2, 0);
 225         }
 226
 227         /* Seconds */
 228         q[qlen++] = ':';
 229         ulongcat(q, &qlen, (seconds % 60), 2, '0');
 230
 231         add_to_string(string, q);
 232         return string;
 233 }
 234
 235 struct string *
 236 add_timeval_to_string(struct string *string, timeval_T *timeval)
 237 {
 238         return add_duration_to_string(string, timeval_to_seconds(timeval));
 239 }
 240
 241 #ifdef HAVE_STRFTIME
 242 struct string *
 243 add_date_to_string(struct string *string, unsigned char *fmt, time_t *date)
 244 {
 245         unsigned char buffer[MAX_STR_LEN];
 246         time_t when_time = date ? *date : time(NULL);
 247         struct tm *when_local = localtime(&when_time);
 248
 249         if (strftime(buffer, sizeof(buffer), fmt, when_local) <= 0)
 250                 return NULL;
 251
 252         return add_to_string(string, buffer);
 253 }
 254 #endif
 255
 256 /* Encoders and string changers */
 257
 258 struct string *
 259 add_string_replace(struct string *string, unsigned char *src, int len,
 260                    unsigned char replaceable, unsigned char replacement)
 261 {
 262         int oldlength = string->length;
 263
 264         if (!add_bytes_to_string(string, src, len))
 265                 return NULL;
 266
 267         for (src = string->source + oldlength; len; len--, src++)
 268                 if (*src == replaceable)
 269                         *src = replacement;
 270
 271         return string;
 272 }
 273
 274 struct string *
 275 add_html_to_string(struct string *string, unsigned char *src, int len)
 276 {
 277
 278 #define isalphanum(q) (isalnum(q) || (q) == '-' || (q) == '_')
 279
 280         for (; len; len--, src++) {
 281                 if (isalphanum(*src) || *src == ' '
 282                     || *src == '.' || *src == ':' || *src == ';') {
 283                         add_bytes_to_string(string, src, 1);
 284                 } else {
 285                         add_bytes_to_string(string, "&#", 2);
 286                         add_long_to_string(string, (long) *src);
 287                         add_char_to_string(string, ';');
 288                 }
 289         }
 290
 291 #undef isalphanum
 292
 293         return string;
 294 }
 295
 296 /* TODO Optimize later --pasky */
 297 struct string *
 298 add_quoted_to_string(struct string *string, const unsigned char *src, int len)
 299 {
 300         for (; len; len--, src++) {
 301                 if (isquote(*src) || *src == '\\')
 302                         add_char_to_string(string, '\\');
 303                 add_char_to_string(string, *src);
 304         }
 305
 306         return string;
 307 }
 308
 309 struct string *
 310 add_shell_quoted_to_string(struct string *string, unsigned char *src, int len)
 311 {
 312         add_char_to_string(string, '\'');
 313         for (; len; len--, ++src)
 314                 if (*src == '\'')
 315                         add_to_string(string, "'\\''");
 316                 else
 317                         add_char_to_string(string, *src);
 318         add_char_to_string(string, '\'');
 319
 320         return string;
 321 }
 322
 323 struct string *
 324 add_shell_safe_to_string(struct string *string, unsigned char *cmd, int cmdlen)
 325 {
 326         int prev_safe = 0;
 327
 328         for (; cmdlen; cmdlen--, cmd++) {
 329                 if ((*cmd == '-' && prev_safe) ||
 330                     (prev_safe = is_safe_in_shell(*cmd))) {
 331                         add_char_to_string(string, *cmd);
 332                 } else {
 333                         /* XXX: Not all programs we might exec are capable of
 334                          * decoding these.  For some, we should just report
 335                          * an error rather than exec with an encoded string. */
 336                         add_char_to_string(string, '%');
 337                         add_char_to_string(string, hx((*cmd & 0xf0) >> 4));
 338                         add_char_to_string(string, hx(*cmd & 0x0f));
 339                 }
 340         }
 341
 342         return string;
 343 }
 344
 345
 346 long
 347 strtolx(unsigned char *str, unsigned char **end)
 348 {
 349         long num;
 350         unsigned char postfix;
 351
 352         errno = 0;
 353         num = strtol(str, (char **) end, 10);
 354         if (errno) return 0;
 355         if (!*end) return num;
 356
 357         postfix = toupper(**end);
 358         if (postfix == 'K') {
 359                 (*end)++;
 360                 if (num < -INT_MAX / 1024) return -INT_MAX;
 361                 if (num > INT_MAX / 1024) return INT_MAX;
 362                 return num * 1024;
 363         }
 364
 365         if (postfix == 'M') {
 366                 (*end)++;
 367                 if (num < -INT_MAX / (1024 * 1024)) return -INT_MAX;
 368                 if (num > INT_MAX / (1024 * 1024)) return INT_MAX;
 369                 return num * (1024 * 1024);
 370         }
 371
 372         return num;
 373 }
 374
 375 int
 376 month2num(const unsigned char *str)
 377 {
 378         unsigned char month[3] = { str[0]|32, str[1]|32, str[2]|32 };
 379
 380         switch (month[0]) {
 381         case 'j': /* jan, jun, jul */
 382                 if (month[1] == 'a') {
 383                         if (month[2] == 'n') return 0; /* jan */
 384                         return -1;
 385                 }
 386                 if (month[1] == 'u') {
 387                         if (month[2] == 'n') return 5; /* jun */
 388                         if (month[2] == 'l') return 6; /* jul */
 389                 }
 390                 return -1;
 391         case 'm': /* mar, may */
 392                 if (month[1] == 'a') {
 393                         if (month[2] == 'r') return 2; /* mar */
 394                         if (month[2] == 'y') return 4; /* may */
 395                 }
 396                 return -1;
 397         case 'a': /* apr, aug */
 398                 if (month[1] == 'p') {
 399                         if (month[2] == 'r') return 3; /* apr */
 400                         return -1;
 401                 }
 402                 if (month[1] == 'u' && month[2] == 'g') return 7; /* aug */
 403                 return -1;
 404         case 's':
 405                 if (month[1] == 'e' && month[2] == 'p') return 8; /* sep */
 406                 return -1;
 407         case 'o':
 408                 if (month[1] == 'c' && month[2] == 't') return 9; /* oct */
 409                 return -1;
 410         case 'n':
 411                 if (month[1] == 'o' && month[2] == 'v') return 10; /* nov */
 412                 return -1;
 413         case 'd':
 414                 if (month[1] == 'e' && month[2] == 'c') return 11; /* dec */
 415                 return -1;
 416         case 'f':
 417                 if (month[1] == 'e' && month[2] == 'b') return 1; /* feb */
 418                 return -1;
 419         default:
 420                 return -1;
 421         }
 422 }
 423
 424 /* This function drops control chars, nbsp char and limit the number of consecutive
 425  * space chars to one. It modifies its argument. */
 426 void
 427 clr_spaces(unsigned char *str)
 428 {
 429         unsigned char *s;
 430         unsigned char *dest = str;
 431
 432         assert(str);
 433
 434         for (s = str; *s; s++)
 435                 if (*s < ' ' || *s == NBSP_CHAR) *s = ' ';
 436
 437         for (s = str; *s; s++) {
 438                 if (*s == ' ' && (dest == str || s[1] == ' ' || !s[1]))
 439                         continue;
 440
 441                 *dest++ = *s;
 442         }
 443
 444         *dest = '\0';
 445 }
 446
 447 /* Replace invalid chars in @title with ' ' and trim all starting/ending
 448  * spaces. */
 449 void
 450 sanitize_title(unsigned char *title)
 451 {
 452         int len = strlen(title);
 453
 454         if (!len) return;
 455
 456         while (len--) {
 457                 if (title[len] < ' ' || title[len] == NBSP_CHAR)
 458                         title[len] = ' ';
 459         }
 460         trim_chars(title, ' ', NULL);
 461 }
 462
 463 /* Returns 0 if @url contains invalid chars, 1 if ok.
 464  * It trims starting/ending spaces. */
 465 int
 466 sanitize_url(unsigned char *url)
 467 {
 468         int len = strlen(url);
 469
 470         if (!len) return 1;
 471
 472         while (len--) {
 473                 if (url[len] < ' ')
 474                         return 0;
 475         }
 476         trim_chars(url, ' ', NULL);
 477         return 1;
 478 }
 479