mpdm_s.c

   1 /*
   2
   3     MPDM - Minimum Profit Data Manager
   4     Copyright (C) 2003/2010 Angel Ortega <angel@triptico.com>
   5
   6     mpdm_s.c - String management
   7
   8     This program is free software; you can redistribute it and/or
   9     modify it under the terms of the GNU General Public License
  10     as published by the Free Software Foundation; either version 2
  11     of the License, or (at your option) any later version.
  12
  13     This program is distributed in the hope that it will be useful,
  14     but WITHOUT ANY WARRANTY; without even the implied warranty of
  15     MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  16     GNU General Public License for more details.
  17
  18     You should have received a copy of the GNU General Public License
  19     along with this program; if not, write to the Free Software
  20     Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA  02111-1307, USA.
  21
  22     http://www.triptico.com
  23
  24 */
  25
  26 #include "config.h"
  27
  28 #include <stdio.h>
  29 #include <stdlib.h>
  30 #include <string.h>
  31 #include <wchar.h>
  32 #include <malloc.h>
  33 #include <locale.h>
  34 #include <wctype.h>
  35
  36 #ifdef CONFOPT_GETTEXT
  37 #include <libintl.h>
  38 #endif
  39
  40 #ifdef CONFOPT_WIN32
  41 #include <windows.h>
  42 #endif
  43
  44 #include "mpdm.h"
  45
  46
  47 /** code **/
  48
  49 void *mpdm_poke_o(void *dst, int *dsize, int *offset, const void *org,
  50                   int osize, int esize)
  51 {
  52     if (org != NULL && osize) {
  53         /* enough room? */
  54         if (*offset + osize > *dsize) {
  55             /* no; enlarge */
  56             *dsize += osize;
  57
  58             dst = realloc(dst, *dsize * esize);
  59         }
  60
  61         memcpy((char *) dst + (*offset * esize), org, osize * esize);
  62         *offset += osize;
  63     }
  64
  65     return dst;
  66 }
  67
  68
  69 void *mpdm_poke(void *dst, int *dsize, const void *org, int osize,
  70                 int esize)
  71 /* pokes (adds) org into dst, which is a dynamic string, making it grow */
  72 {
  73     int offset = *dsize;
  74
  75     return mpdm_poke_o(dst, dsize, &offset, org, osize, esize);
  76 }
  77
  78
  79 wchar_t *mpdm_pokewsn(wchar_t * dst, int *dsize, const wchar_t * str,
  80                       int slen)
  81 /* adds a wide string to dst using mpdm_poke() with size */
  82 {
  83     if (str)
  84         dst = mpdm_poke(dst, dsize, str, slen, sizeof(wchar_t));
  85
  86     return dst;
  87 }
  88
  89
  90 wchar_t *mpdm_pokews(wchar_t * dst, int *dsize, const wchar_t * str)
  91 /* adds a wide string to dst using mpdm_poke() */
  92 {
  93     if (str)
  94         dst = mpdm_pokewsn(dst, dsize, str, wcslen(str));
  95
  96     return dst;
  97 }
  98
  99
 100 wchar_t *mpdm_pokev(wchar_t * dst, int *dsize, const mpdm_t v)
 101 /* adds the string in v to dst using mpdm_poke() */
 102 {
 103     if (v != NULL) {
 104         const wchar_t *ptr = mpdm_string(v);
 105
 106         mpdm_ref(v);
 107         dst = mpdm_pokews(dst, dsize, ptr);
 108         mpdm_unref(v);
 109     }
 110
 111     return dst;
 112 }
 113
 114
 115 wchar_t *mpdm_mbstowcs(const char *str, int *s, int l)
 116 /* converts an mbs to a wcs, but filling invalid chars
 117    with question marks instead of just failing */
 118 {
 119     wchar_t *ptr = NULL;
 120     char tmp[64];               /* really MB_CUR_MAX + 1 */
 121     wchar_t wc;
 122     int n, i, c, t = 0;
 123     char *cstr;
 124
 125     /* allow NULL values for s */
 126     if (s == NULL)
 127         s = &t;
 128
 129     /* if there is a limit, duplicate and break the string */
 130     if (l >= 0) {
 131         cstr = strdup(str);
 132         cstr[l] = '\0';
 133     }
 134     else
 135         cstr = (char *) str;
 136
 137     /* try first a direct conversion with mbstowcs */
 138     if ((*s = mbstowcs(NULL, cstr, 0)) != -1) {
 139         /* direct conversion is possible; do it */
 140         if ((ptr = malloc((*s + 1) * sizeof(wchar_t))) != NULL) {
 141             mbstowcs(ptr, cstr, *s);
 142             ptr[*s] = L'\0';
 143         }
 144     }
 145     else {
 146         /* zero everything */
 147         *s = n = i = 0;
 148
 149         for (;;) {
 150             /* no more characters to process? */
 151             if ((c = cstr[n + i]) == '\0' && i == 0)
 152                 break;
 153
 154             tmp[i++] = c;
 155             tmp[i] = '\0';
 156
 157             /* try to convert */
 158             if (mbstowcs(&wc, tmp, 1) == (size_t) - 1) {
 159                 /* can still be an incomplete multibyte char? */
 160                 if (c != '\0' && i <= (int) MB_CUR_MAX)
 161                     continue;
 162                 else {
 163                     /* too many failing bytes; skip 1 byte */
 164                     wc = L'?';
 165                     i = 1;
 166                 }
 167             }
 168
 169             /* skip used bytes and back again */
 170             n += i;
 171             i = 0;
 172
 173             /* store new char */
 174             if ((ptr = mpdm_poke(ptr, s, &wc, 1, sizeof(wchar_t))) == NULL)
 175                 break;
 176         }
 177
 178         /* null terminate and count one less */
 179         if (ptr != NULL) {
 180             ptr = mpdm_poke(ptr, s, L"", 1, sizeof(wchar_t));
 181             (*s)--;
 182         }
 183     }
 184
 185     /* free the duplicate */
 186     if (cstr != str)
 187         free(cstr);
 188
 189     return ptr;
 190 }
 191
 192
 193 char *mpdm_wcstombs(const wchar_t * str, int *s)
 194 /* converts a wcs to an mbs, but filling invalid chars
 195    with question marks instead of just failing */
 196 {
 197     char *ptr = NULL;
 198     char tmp[64];               /* really MB_CUR_MAX + 1 */
 199     int l, t = 0;
 200
 201     /* allow NULL values for s */
 202     if (s == NULL)
 203         s = &t;
 204
 205     /* try first a direct conversion with wcstombs */
 206     if ((*s = wcstombs(NULL, str, 0)) != -1) {
 207         /* direct conversion is possible; do it and return */
 208         if ((ptr = malloc(*s + 1)) != NULL) {
 209             wcstombs(ptr, str, *s);
 210             ptr[*s] = '\0';
 211         }
 212
 213         return ptr;
 214     }
 215
 216     /* invalid encoding? convert characters one by one */
 217     *s = 0;
 218
 219     while (*str) {
 220         if ((l = wctomb(tmp, *str)) <= 0) {
 221             /* if char couldn't be converted,
 222                write a question mark instead */
 223             l = wctomb(tmp, L'?');
 224         }
 225
 226         tmp[l] = '\0';
 227         if ((ptr = mpdm_poke(ptr, s, tmp, l, 1)) == NULL)
 228             break;
 229
 230         str++;
 231     }
 232
 233     /* null terminate and count one less */
 234     if (ptr != NULL) {
 235         ptr = mpdm_poke(ptr, s, "", 1, 1);
 236         (*s)--;
 237     }
 238
 239     return ptr;
 240 }
 241
 242
 243 mpdm_t mpdm_new_wcs(int flags, const wchar_t * str, int size, int cpy)
 244 /* creates a new string value from a wcs */
 245 {
 246     wchar_t *ptr;
 247
 248     /* a size of -1 means 'calculate it' */
 249     if (size == -1 && str != NULL)
 250         size = wcslen(str);
 251
 252     /* create a copy? */
 253     if (cpy) {
 254         /* free() on destruction */
 255         flags |= MPDM_FREE;
 256
 257         /* allocs */
 258         if ((ptr = malloc((size + 1) * sizeof(wchar_t))) == NULL)
 259             return NULL;
 260
 261         /* if no source, reset to zeroes; otherwise, copy */
 262         if (str == NULL)
 263             memset(ptr, '\0', size * sizeof(wchar_t));
 264         else {
 265             wcsncpy(ptr, str, size);
 266             ptr[size] = L'\0';
 267         }
 268     }
 269     else
 270         ptr = (wchar_t *) str;
 271
 272     /* it's a string */
 273     flags |= MPDM_STRING;
 274
 275     return mpdm_new(flags, ptr, size);
 276 }
 277
 278
 279 mpdm_t mpdm_new_mbstowcs(int flags, const char *str, int l)
 280 /* creates a new string value from an mbs */
 281 {
 282     wchar_t *ptr;
 283     int size;
 284
 285     if ((ptr = mpdm_mbstowcs(str, &size, l)) == NULL)
 286         return NULL;
 287
 288     /* it's a string */
 289     flags |= (MPDM_STRING | MPDM_FREE);
 290
 291     return mpdm_new(flags, ptr, size);
 292 }
 293
 294
 295 mpdm_t mpdm_new_wcstombs(int flags, const wchar_t * str)
 296 /* creates a new mbs value from a wbs */
 297 {
 298     char *ptr;
 299     int size;
 300
 301     ptr = mpdm_wcstombs(str, &size);
 302
 303     flags |= MPDM_FREE;
 304
 305     /* unset the string flag; mbs,s are not 'strings' */
 306     flags &= ~MPDM_STRING;
 307
 308     return mpdm_new(flags, ptr, size);
 309 }
 310
 311
 312 mpdm_t mpdm_new_i(int ival)
 313 /* creates a new string value from an integer */
 314 {
 315     mpdm_t v;
 316     char tmp[32];
 317
 318     /* creates the visual representation */
 319     snprintf(tmp, sizeof(tmp), "%d", ival);
 320
 321     v = MPDM_MBS(tmp);
 322
 323     return mpdm_set_ival(v, ival);
 324 }
 325
 326
 327 mpdm_t mpdm_new_r(double rval)
 328 /* creates a new string value from a real number */
 329 {
 330     mpdm_t v;
 331     char tmp[128];
 332
 333     /* creates the visual representation */
 334     snprintf(tmp, sizeof(tmp), "%lf", rval);
 335
 336     /* manually strip useless zeroes */
 337     if (strchr(tmp, '.') != NULL) {
 338         char *ptr;
 339
 340         for (ptr = tmp + strlen(tmp) - 1; *ptr == '0'; ptr--);
 341
 342         /* if it's over the ., strip it also */
 343         if (*ptr != '.')
 344             ptr++;
 345
 346         *ptr = '\0';
 347     }
 348
 349     v = MPDM_MBS(tmp);
 350
 351     return mpdm_set_rval(v, rval);
 352 }
 353
 354
 355 /* interface */
 356
 357 /**
 358  * mpdm_string2 - Returns a printable representation of a value (with buffer).
 359  * @v: the value
 360  * @wtmp: the external buffer
 361  *
 362  * Returns a printable representation of a value. For strings, it's
 363  * the value data itself; for any other type, a conversion to string
 364  * is returned instead. If @v is not a string, the @wtmp buffer
 365  * can be used as a placeholder for the string representation.
 366  *
 367  * The reference count value in @v is not touched.
 368  * [Strings]
 369  */
 370 wchar_t *mpdm_string2(const mpdm_t v, wchar_t *wtmp)
 371 {
 372     char tmp[32];
 373     wchar_t *ret;
 374
 375     /* if it's NULL, return a constant */
 376     if (v == NULL)
 377         ret = L"[NULL]";
 378     else
 379         /* if it's a string, return it */
 380     if (v->flags & MPDM_STRING)
 381         ret = (wchar_t *) v->data;
 382     else {
 383         /* otherwise, return a visual representation */
 384         snprintf(tmp, sizeof(tmp), "%p", v);
 385         mbstowcs(wtmp, tmp, sizeof(tmp) * sizeof(wchar_t));
 386
 387         ret = wtmp;
 388     }
 389
 390     return ret;
 391 }
 392
 393
 394 /**
 395  * mpdm_string - Returns a printable representation of a value.
 396  * @v: the value
 397  *
 398  * Returns a printable representation of a value. For strings, it's
 399  * the value data itself; for any other type, a conversion to string
 400  * is returned instead. This value should be used immediately, as it
 401  * can be a pointer to a static buffer.
 402  *
 403  * The reference count value in @v is not touched.
 404  * [Strings]
 405  */
 406 wchar_t *mpdm_string(const mpdm_t v)
 407 {
 408     static wchar_t tmp[32];
 409
 410     return mpdm_string2(v, tmp);
 411 }
 412
 413
 414 /**
 415  * mpdm_cmp - Compares two values.
 416  * @v1: the first value
 417  * @v2: the second value
 418  *
 419  * Compares two values. If both has the MPDM_STRING flag set,
 420  * a comparison using wcscoll() is returned; if both are arrays,
 421  * the size is compared first and, if they have the same number
 422  * elements, each one is compared; otherwise, a simple visual
 423  * representation comparison is done.
 424  * [Strings]
 425  */
 426 int mpdm_cmp(const mpdm_t v1, const mpdm_t v2)
 427 {
 428     int r;
 429
 430     mpdm_ref(v1);
 431     mpdm_ref(v2);
 432
 433     /* same values? */
 434     if (v1 == v2)
 435         r = 0;
 436     else
 437         /* is any value NULL? */
 438     if (v1 == NULL)
 439         r = -1;
 440     else
 441     if (v2 == NULL)
 442         r = 1;
 443     else
 444         /* different values, but same content? (unlikely) */
 445     if (v1->data == v2->data)
 446         r = 0;
 447     else
 448     if (MPDM_IS_ARRAY(v1) && MPDM_IS_ARRAY(v2)) {
 449         /* compare first the sizes */
 450         if ((r = mpdm_size(v1) - mpdm_size(v2)) == 0) {
 451             int n;
 452
 453             /* they have the same size;
 454                compare each pair of elements */
 455             for (n = 0; n < mpdm_size(v1); n++) {
 456                 if ((r = mpdm_cmp(mpdm_aget(v1, n),
 457                                   mpdm_aget(v2, n))) != 0)
 458                     break;
 459             }
 460         }
 461     }
 462     else {
 463         wchar_t tmp[32];
 464
 465         r = wcscoll(mpdm_string(v1), mpdm_string2(v2, tmp));
 466     }
 467
 468     mpdm_unref(v2);
 469     mpdm_unref(v1);
 470
 471     return r;
 472 }
 473
 474
 475 /**
 476  * mpdm_cmp_s - Compares two values (string version).
 477  * @v1: the first value
 478  * @v2: the second value
 479  *
 480  * Compares two values. If both has the MPDM_STRING flag set,
 481  * a comparison using wcscoll() is returned; if both are arrays,
 482  * the size is compared first and, if they have the same number
 483  * elements, each one is compared; otherwise, a simple visual
 484  * representation comparison is done.
 485  */
 486 int mpdm_cmp_s(const mpdm_t v1, const wchar_t * v2)
 487 {
 488     return mpdm_cmp(v1, MPDM_S(v2));
 489 }
 490
 491
 492 /**
 493  * mpdm_splice - Creates a new string value from another.
 494  * @v: the original value
 495  * @i: the value to be inserted
 496  * @offset: offset where the substring is to be inserted
 497  * @del: number of characters to delete
 498  *
 499  * Creates a new string value from @v, deleting @del chars at @offset
 500  * and substituting them by @i. If @del is 0, no deletion is done.
 501  * both @offset and @del can be negative; if this is the case, it's
 502  * assumed as counting from the end of @v. If @v is NULL, @i will become
 503  * the new string, and both @offset and @del will be ignored. If @v is
 504  * not NULL and @i is, no insertion process is done (only deletion, if
 505  * applicable).
 506  *
 507  * Returns a two element array, with the new string in the first
 508  * element and the deleted string in the second (with a NULL value
 509  * if @del is 0).
 510  * [Strings]
 511  */
 512 mpdm_t mpdm_splice(const mpdm_t v, const mpdm_t i, int offset, int del)
 513 {
 514     mpdm_t w;
 515     mpdm_t n = NULL;
 516     mpdm_t d = NULL;
 517     int os, ns, r;
 518     int ins = 0;
 519     wchar_t *ptr;
 520
 521     mpdm_ref(v);
 522     mpdm_ref(i);
 523
 524     if (v != NULL) {
 525         os = mpdm_size(v);
 526
 527         /* negative offsets start from the end */
 528         if (offset < 0)
 529             offset = os + 1 - offset;
 530
 531         /* never add further the end */
 532         if (offset > os)
 533             offset = os;
 534
 535         /* negative del counts as 'characters left' */
 536         if (del < 0)
 537             del = os + 1 - offset + del;
 538
 539         /* something to delete? */
 540         if (del > 0) {
 541             /* never delete further the end */
 542             if (offset + del > os)
 543                 del = os - offset;
 544
 545             /* deleted string */
 546             d = MPDM_NS(((wchar_t *) v->data) + offset, del);
 547         }
 548         else
 549             del = 0;
 550
 551         /* something to insert? */
 552         ins = mpdm_size(i);
 553
 554         /* new size and remainder */
 555         ns = os + ins - del;
 556         r = offset + del;
 557
 558         n = MPDM_NS(NULL, ns);
 559
 560         ptr = (wchar_t *) n->data;
 561
 562         /* copy the beginning */
 563         if (offset > 0) {
 564             wcsncpy(ptr, v->data, offset);
 565             ptr += offset;
 566         }
 567
 568         /* copy the text to be inserted */
 569         if (ins > 0) {
 570             wcsncpy(ptr, i->data, ins);
 571             ptr += ins;
 572         }
 573
 574         /* copy the remaining */
 575         os -= r;
 576         if (os > 0) {
 577             wcsncpy(ptr, ((wchar_t *) v->data) + r, os);
 578             ptr += os;
 579         }
 580
 581         /* null terminate */
 582         *ptr = L'\0';
 583     }
 584     else
 585         n = i;
 586
 587     /* creates the output array */
 588     w = MPDM_A(2);
 589
 590     mpdm_ref(w);
 591     mpdm_aset(w, n, 0);
 592     mpdm_aset(w, d, 1);
 593     mpdm_unrefnd(w);
 594
 595     mpdm_unref(i);
 596     mpdm_unref(v);
 597
 598     return w;
 599 }
 600
 601
 602 /**
 603  * mpdm_strcat_sn - Concatenates two strings (string with size version).
 604  * @s1: the first string
 605  * @s2: the second string
 606  * @size: the size of the second string
 607  *
 608  * Returns a new string formed by the concatenation of @s1 and @s2.
 609  * [Strings]
 610  */
 611 mpdm_t mpdm_strcat_sn(const mpdm_t s1, const wchar_t * s2, int size)
 612 {
 613     wchar_t *ptr = NULL;
 614     int s = 0;
 615     mpdm_t r;
 616
 617     if (s1 == NULL && s2 == NULL)
 618         r = NULL;
 619     else {
 620         ptr = mpdm_pokev(ptr, &s, s1);
 621         ptr = mpdm_pokewsn(ptr, &s, s2, size);
 622
 623         ptr = mpdm_poke(ptr, &s, L"", 1, sizeof(wchar_t));
 624         r = MPDM_ENS(ptr, s - 1);
 625     }
 626
 627     return r;
 628 }
 629
 630
 631 /**
 632  * mpdm_strcat_s - Concatenates two strings (string version).
 633  * @s1: the first string
 634  * @s2: the second string
 635  *
 636  * Returns a new string formed by the concatenation of @s1 and @s2.
 637  * [Strings]
 638  */
 639 mpdm_t mpdm_strcat_s(const mpdm_t s1, const wchar_t * s2)
 640 {
 641     return mpdm_strcat_sn(s1, s2, s2 ? wcslen(s2) : 0);
 642 }
 643
 644
 645 /**
 646  * mpdm_strcat - Concatenates two strings.
 647  * @s1: the first string
 648  * @s2: the second string
 649  *
 650  * Returns a new string formed by the concatenation of @s1 and @s2.
 651  * [Strings]
 652  */
 653 mpdm_t mpdm_strcat(const mpdm_t s1, const mpdm_t s2)
 654 {
 655     mpdm_t r;
 656
 657     mpdm_ref(s2);
 658     r = mpdm_strcat_s(s1, s2 ? mpdm_string(s2) : NULL);
 659     mpdm_unref(s2);
 660
 661     return r;
 662 }
 663
 664
 665 /**
 666  * mpdm_ival - Returns a value's data as an integer.
 667  * @v: the value
 668  *
 669  * Returns a value's data as an integer. If the value is a string,
 670  * it's converted via sscanf and returned; non-string values have all
 671  * an ival of 0. The converted integer is cached, so costly string
 672  * conversions are only done once. Values created with the MPDM_IVAL
 673  * flag set have its ival cached from the beginning.
 674  * [Strings]
 675  * [Value Management]
 676  */
 677 int mpdm_ival(mpdm_t v)
 678 {
 679     int i = 0;
 680
 681     mpdm_ref(v);
 682
 683     if (v != NULL) {
 684         /* if there is no cached integer, calculate it */
 685         if (!(v->flags & MPDM_IVAL)) {
 686             /* if it's a string, calculate it; other
 687                values will have an ival of 0 */
 688             if (v->flags & MPDM_STRING) {
 689                 char tmp[32];
 690                 char *fmt = "%i";
 691
 692                 wcstombs(tmp, (wchar_t *) v->data, sizeof(tmp));
 693                 tmp[sizeof(tmp) - 1] = '\0';
 694
 695                 /* workaround for mingw32: as it doesn't
 696                    correctly parse octal and hexadecimal
 697                    numbers, they are tried as special cases */
 698                 if (tmp[0] == '0') {
 699                     if (tmp[1] == 'b' || tmp[1] == 'B') {
 700                         /* binary number */
 701                         fmt = NULL;
 702                         char *ptr = &tmp[2];
 703
 704                         while (*ptr == '0' || *ptr == '1') {
 705                             i <<= 1;
 706
 707                             if (*ptr == '1')
 708                                 i |= 1;
 709
 710                             ptr++;
 711                         }
 712                     }
 713                     else
 714                     if (tmp[1] == 'x' || tmp[1] == 'X')
 715                         fmt = "%x";
 716                     else
 717                         fmt = "%o";
 718                 }
 719
 720                 if (fmt != NULL)
 721                     sscanf(tmp, fmt, &i);
 722             }
 723
 724             mpdm_set_ival(v, i);
 725         }
 726
 727         i = v->ival;
 728     }
 729
 730     mpdm_unref(v);
 731
 732     return i;
 733 }
 734
 735
 736 /**
 737  * mpdm_rval - Returns a value's data as a real number (double).
 738  * @v: the value
 739  *
 740  * Returns a value's data as a real number (double float). If the value
 741  * is a string, it's converted via sscanf and returned; non-string values
 742  * have all an rval of 0. The converted double is cached, so costly string
 743  * conversions are only done once. Values created with the MPDM_RVAL
 744  * flag set have its rval cached from the beginning.
 745  * [Strings]
 746  * [Value Management]
 747  */
 748 double mpdm_rval(mpdm_t v)
 749 {
 750     double r = 0.0;
 751
 752     mpdm_ref(v);
 753
 754     if (v != NULL) {
 755         /* if there is no cached double, calculate it */
 756         if (!(v->flags & MPDM_RVAL)) {
 757             /* if it's a string, calculate it; other
 758                values will have an rval of 0.0 */
 759             if (v->flags & MPDM_STRING) {
 760                 char tmp[128];
 761                 char *prev_locale;
 762
 763                 wcstombs(tmp, (wchar_t *) v->data, sizeof(tmp));
 764                 tmp[sizeof(tmp) - 1] = '\0';
 765
 766                 /* if the number starts with 0, it's
 767                    an octal or hexadecimal number; just
 768                    take the integer value and cast it */
 769                 if (tmp[0] == '0' && tmp[1] != '.')
 770                     r = (double) mpdm_ival(v);
 771                 else {
 772                     /* set locale to C for non locale-dependent
 773                        floating point conversion */
 774                     prev_locale = setlocale(LC_NUMERIC, "C");
 775
 776                     /* read */
 777                     sscanf(tmp, "%lf", &r);
 778
 779                     /* set previous locale */
 780                     setlocale(LC_NUMERIC, prev_locale);
 781                 }
 782             }
 783
 784             mpdm_set_rval(v, r);
 785         }
 786
 787         r = v->rval;
 788     }
 789
 790     mpdm_unref(v);
 791
 792     return r;
 793 }
 794
 795
 796 /**
 797  * mpdm_gettext - Translates a string to the current language.
 798  * @str: the string
 799  *
 800  * Translates the @str string to the current language.
 801  *
 802  * This function can still be used even if there is no real gettext
 803  * support() by manually filling the __I18N__ hash.
 804  *
 805  * If the string is found in the current table, the translation is
 806  * returned; otherwise, the same @str value is returned.
 807  * [Strings]
 808  * [Localization]
 809  */
 810 mpdm_t mpdm_gettext(const mpdm_t str)
 811 {
 812     mpdm_t v;
 813     mpdm_t i18n = NULL;
 814
 815     /* gets the cache */
 816     if ((i18n = mpdm_hget_s(mpdm_root(), L"__I18N__")) == NULL)
 817         i18n = mpdm_hset_s(mpdm_root(), L"__I18N__", MPDM_H(0));
 818
 819     mpdm_ref(str);
 820
 821     /* try first the cache */
 822     if ((v = mpdm_hget(i18n, str)) == NULL) {
 823 #ifdef CONFOPT_GETTEXT
 824         char *s;
 825         mpdm_t t;
 826
 827         /* convert to mbs */
 828         t = mpdm_ref(MPDM_2MBS(str->data));
 829
 830         /* ask gettext for it */
 831         s = gettext((char *) t->data);
 832
 833         if (s != t->data)
 834             v = MPDM_MBS(s);
 835         else
 836             v = str;
 837
 838         mpdm_unref(t);
 839
 840 #else                           /* CONFOPT_GETTEXT */
 841
 842         v = str;
 843
 844 #endif                          /* CONFOPT_GETTEXT */
 845
 846         /* store in the cache */
 847         mpdm_hset(i18n, str, v);
 848     }
 849
 850     mpdm_unref(str);
 851
 852     return v;
 853 }
 854
 855
 856 /**
 857  * mpdm_gettext_domain - Sets domain and data directory for translations.
 858  * @dom: the domain (application name)
 859  * @data: directory contaning the .mo files
 860  *
 861  * Sets the domain (application name) and translation data for translating
 862  * strings that will be returned by mpdm_gettext().@data must point to a
 863  * directory containing the .mo (compiled .po) files.
 864  *
 865  * If there is no gettext support, returns 0, or 1 otherwise.
 866  * [Strings]
 867  * [Localization]
 868  */
 869 int mpdm_gettext_domain(const mpdm_t dom, const mpdm_t data)
 870 {
 871     int ret = 0;
 872
 873     mpdm_ref(dom);
 874     mpdm_ref(data);
 875
 876 #ifdef CONFOPT_GETTEXT
 877
 878     mpdm_t dm;
 879     mpdm_t dt;
 880
 881     /* convert both to mbs,s */
 882     dm = mpdm_ref(MPDM_2MBS(dom->data));
 883     dt = mpdm_ref(MPDM_2MBS(data->data));
 884
 885     /* bind and set domain */
 886     bindtextdomain((char *) dm->data, (char *) dt->data);
 887     textdomain((char *) dm->data);
 888
 889     mpdm_hset_s(mpdm_root(), L"__I18N__", MPDM_H(0));
 890
 891     mpdm_unref(dt);
 892     mpdm_unref(dm);
 893
 894     ret = 1;
 895
 896 #endif                          /* CONFOPT_GETTEXT */
 897
 898 #ifdef CONFOPT_WIN32
 899
 900     mpdm_t v;
 901
 902     if ((v = mpdm_hget_s(mpdm_root(), L"ENV")) != NULL &&
 903         mpdm_hget_s(v, L"LANG") == NULL) {
 904         wchar_t *wptr = L"en";
 905
 906         /* MS Windows crappy language constants... */
 907
 908         switch ((GetSystemDefaultLangID() & 0x00ff)) {
 909         case 0x01:
 910             wptr = L"ar";
 911             break;              /* arabic */
 912         case 0x02:
 913             wptr = L"bg";
 914             break;              /* bulgarian */
 915         case 0x03:
 916             wptr = L"ca";
 917             break;              /* catalan */
 918         case 0x04:
 919             wptr = L"zh";
 920             break;              /* chinese */
 921         case 0x05:
 922             wptr = L"cz";
 923             break;              /* czech */
 924         case 0x06:
 925             wptr = L"da";
 926             break;              /* danish */
 927         case 0x07:
 928             wptr = L"de";
 929             break;              /* german */
 930         case 0x08:
 931             wptr = L"el";
 932             break;              /* greek */
 933         case 0x09:
 934             wptr = L"en";
 935             break;              /* english */
 936         case 0x0a:
 937             wptr = L"es";
 938             break;              /* spanish */
 939         case 0x0b:
 940             wptr = L"fi";
 941             break;              /* finnish */
 942         case 0x0c:
 943             wptr = L"fr";
 944             break;              /* french */
 945         case 0x0d:
 946             wptr = L"he";
 947             break;              /* hebrew */
 948         case 0x0e:
 949             wptr = L"hu";
 950             break;              /* hungarian */
 951         case 0x0f:
 952             wptr = L"is";
 953             break;              /* icelandic */
 954         case 0x10:
 955             wptr = L"it";
 956             break;              /* italian */
 957         case 0x11:
 958             wptr = L"jp";
 959             break;              /* japanese */
 960         case 0x12:
 961             wptr = L"ko";
 962             break;              /* korean */
 963         case 0x13:
 964             wptr = L"nl";
 965             break;              /* dutch */
 966         case 0x14:
 967             wptr = L"no";
 968             break;              /* norwegian */
 969         case 0x15:
 970             wptr = L"po";
 971             break;              /* polish */
 972         case 0x16:
 973             wptr = L"pt";
 974             break;              /* portuguese */
 975         case 0x17:
 976             wptr = L"rm";
 977             break;              /* romansh (switzerland) */
 978         case 0x18:
 979             wptr = L"ro";
 980             break;              /* romanian */
 981         case 0x19:
 982             wptr = L"ru";
 983             break;              /* russian */
 984         case 0x1a:
 985             wptr = L"sr";
 986             break;              /* serbian */
 987         case 0x1b:
 988             wptr = L"sk";
 989             break;              /* slovak */
 990         case 0x1c:
 991             wptr = L"sq";
 992             break;              /* albanian */
 993         case 0x1d:
 994             wptr = L"sv";
 995             break;              /* swedish */
 996         }
 997
 998         mpdm_hset_s(v, L"LANG", MPDM_S(wptr));
 999     }
1000
1001 #endif                          /* CONFOPT_WIN32 */
1002
1003     mpdm_unref(data);
1004     mpdm_unref(dom);
1005
1006     return ret;
1007 }
1008
1009
1010 #ifdef CONFOPT_WCWIDTH
1011
1012 int wcwidth(wchar_t);
1013
1014 int mpdm_wcwidth(wchar_t c)
1015 {
1016     return wcwidth(c);
1017 }
1018
1019 #else                           /* CONFOPT_WCWIDTH */
1020
1021 #include "wcwidth.c"
1022
1023 int mpdm_wcwidth(wchar_t c)
1024 {
1025     return mk_wcwidth(c);
1026 }
1027
1028 #endif                          /* CONFOPT_WCWIDTH */
1029
1030
1031 /**
1032  * mpdm_sprintf - Formats a sprintf()-like string.
1033  * @fmt: the string format
1034  * @args: an array of values
1035  *
1036  * Formats a string using the sprintf() format taking the values from @args.
1037  * [Strings]
1038  */
1039 mpdm_t mpdm_sprintf(const mpdm_t fmt, const mpdm_t args)
1040 {
1041     const wchar_t *i = fmt->data;
1042     wchar_t *o = NULL;
1043     int l = 0, n = 0;
1044     wchar_t c;
1045
1046     mpdm_ref(fmt);
1047     mpdm_ref(args);
1048
1049     /* loop all characters */
1050     while ((c = *i++) != L'\0') {
1051         int m = 0;
1052         wchar_t *tptr = NULL;
1053         wchar_t *wptr = NULL;
1054
1055         if (c == L'%') {
1056             /* format directive */
1057             char t_fmt[128];
1058             char tmp[1024];
1059             mpdm_t v;
1060             char *ptr = NULL;
1061
1062             /* transfer the % */
1063             t_fmt[m++] = '%';
1064
1065             /* transform the format to mbs */
1066             while (*i != L'\0' &&
1067                    m < (int) (sizeof(t_fmt) - MB_CUR_MAX - 1) &&
1068                    wcschr(L"-.0123456789", *i) != NULL)
1069                 m += wctomb(&t_fmt[m], *i++);
1070
1071             /* transfer the directive */
1072             m += wctomb(&t_fmt[m], *i++);
1073
1074             t_fmt[m] = '\0';
1075
1076             /* by default, copies the format */
1077             strcpy(tmp, t_fmt);
1078
1079             /* pick next value */
1080             v = mpdm_aget(args, n++);
1081
1082             switch (t_fmt[m - 1]) {
1083             case 'd':
1084             case 'i':
1085             case 'u':
1086             case 'x':
1087             case 'X':
1088             case 'o':
1089
1090                 /* integer value */
1091                 snprintf(tmp, sizeof(tmp) - 1, t_fmt, mpdm_ival(v));
1092                 break;
1093
1094             case 'f':
1095
1096                 /* float (real) value */
1097                 snprintf(tmp, sizeof(tmp) - 1, t_fmt, mpdm_rval(v));
1098                 break;
1099
1100             case 's':
1101
1102                 /* string value */
1103                 ptr = mpdm_wcstombs(mpdm_string(v), NULL);
1104                 snprintf(tmp, sizeof(tmp) - 1, t_fmt, ptr);
1105                 free(ptr);
1106
1107                 break;
1108
1109             case 'c':
1110
1111                 /* char */
1112                 m = 1;
1113                 wptr = &c;
1114                 c = mpdm_ival(v);
1115                 break;
1116
1117             case 'b':
1118
1119                 ptr = tmp;
1120                 unsigned int mask;
1121                 int p = 0;
1122
1123                 mask = 1 << ((sizeof(int) * 8) - 1);
1124                 while (mask) {
1125                     if (mask & (unsigned int) mpdm_ival(v)) {
1126                         *ptr++ = '1';
1127                         p = 1;
1128                     }
1129                     else
1130                     if (p)
1131                         *ptr++ = '0';
1132
1133                     mask >>= 1;
1134                 }
1135
1136                 if (ptr == tmp)
1137                     *ptr++ = '0';
1138
1139                 *ptr = '\0';
1140                 break;
1141
1142             case '%':
1143
1144                 /* percent sign */
1145                 m = 1;
1146                 wptr = &c;
1147                 break;
1148             }
1149
1150             /* transfer */
1151             if (wptr == NULL)
1152                 wptr = tptr = mpdm_mbstowcs(tmp, &m, -1);
1153         }
1154         else {
1155             /* raw character */
1156             m = 1;
1157             wptr = &c;
1158         }
1159
1160         /* transfer */
1161         o = mpdm_poke(o, &l, wptr, m, sizeof(wchar_t));
1162
1163         /* free the temporary buffer, if any */
1164         if (tptr != NULL)
1165             free(tptr);
1166     }
1167
1168     if (o == NULL)
1169         return NULL;
1170
1171     /* null-terminate */
1172     o = mpdm_poke(o, &l, L"", 1, sizeof(wchar_t));
1173
1174     mpdm_unref(args);
1175     mpdm_unref(fmt);
1176
1177     return MPDM_ENS(o, l - 1);
1178 }
1179
1180
1181 /**
1182  * mpdm_ulc - Converts a string to uppercase or lowecase.
1183  * @s: the string
1184  * @u: convert to uppercase (1) or to lowercase (0).
1185  *
1186  * Converts @s to uppercase (for @u == 1) or to lowercase (@u == 0).
1187  * [Strings]
1188  */
1189 mpdm_t mpdm_ulc(const mpdm_t s, int u)
1190 {
1191     mpdm_t r = NULL;
1192     wchar_t *optr;
1193     int i;
1194
1195     mpdm_ref(s);
1196
1197     i = mpdm_size(s);
1198
1199     if ((optr = malloc((i + 1) * sizeof(wchar_t))) != NULL) {
1200         wchar_t *iptr = mpdm_string(s);
1201         int n;
1202
1203         for (n = 0; n < i; n++)
1204             optr[n] = u ? towupper(iptr[n]) : towlower(iptr[n]);
1205
1206         optr[n] = L'\0';
1207         r = MPDM_ENS(optr, i);
1208     }
1209
1210     mpdm_unref(s);
1211
1212     return r;
1213 }
1214
1215
1216 /* scanf working buffers */
1217 #define SCANF_BUF_SIZE 1024
1218 static wchar_t scanf_yset[SCANF_BUF_SIZE];
1219 static wchar_t scanf_nset[SCANF_BUF_SIZE];
1220 static wchar_t scanf_mark[SCANF_BUF_SIZE];
1221
1222 struct {
1223     wchar_t cmd;
1224     wchar_t *yset;
1225     wchar_t *nset;
1226 } scanf_sets[] = {
1227     { L's',  L"",                         L" \t"},
1228     { L'u',  L"0123456789",               L""},
1229     { L'd',  L"-0123456789",              L""},
1230     { L'i',  L"-0123456789",              L""},
1231     { L'f',  L"-0123456789.",             L""},
1232     { L'x',  L"-0123456789xabcdefABCDEF", L""},
1233     { L'\0', NULL,                        NULL},
1234 };
1235
1236 /**
1237  * mpdm_sscanf - Extracts data like sscanf().
1238  * @str: the string to be parsed
1239  * @fmt: the string format
1240  * @offset: the character offset to start scanning
1241  *
1242  * Extracts data from a string using a special format pattern, very
1243  * much like the scanf() series of functions in the C library. Apart
1244  * from the standard percent-sign-commands (s, u, d, i, f, x,
1245  * n, [, with optional size and * to ignore), it implements S,
1246  * to match a string of characters upto what follows in the format
1247  * string. Also, the [ set of characters can include other % formats.
1248  *
1249  * Returns an array with the extracted values. If %n is used, the
1250  * position in the scanned string is returned as the value.
1251  * [Strings]
1252  */
1253 mpdm_t mpdm_sscanf(const mpdm_t str, const mpdm_t fmt, int offset)
1254 {
1255     wchar_t *i = (wchar_t *) str->data;
1256     wchar_t *f = (wchar_t *) fmt->data;
1257     mpdm_t r;
1258
1259     mpdm_ref(fmt);
1260     mpdm_ref(str);
1261
1262     i += offset;
1263     r = MPDM_A(0);
1264     mpdm_ref(r);
1265
1266     while (*f) {
1267         if (*f == L'%') {
1268             wchar_t *ptr = NULL;
1269             int size = 0;
1270             wchar_t cmd;
1271             int vsize = 0;
1272             int ignore = 0;
1273             int msize = 0;
1274
1275             /* empty all buffers */
1276             scanf_yset[0] = scanf_nset[0] = scanf_mark[0] = L'\0';
1277
1278             f++;
1279
1280             /* an asterisk? don't return next value */
1281             if (*f == L'*') {
1282                 ignore = 1;
1283                 f++;
1284             }
1285
1286             /* does it have a size? */
1287             while (wcschr(L"0123456789", *f)) {
1288                 vsize *= 10;
1289                 vsize += *f - L'0';
1290                 f++;
1291             }
1292
1293             /* if no size, set it to an arbitrary big limit */
1294             if (!vsize)
1295                 vsize = 0xfffffff;
1296
1297             /* now *f should contain a command */
1298             cmd = *f;
1299             f++;
1300
1301             /* is it a verbatim percent sign? */
1302             if (cmd == L'%') {
1303                 vsize = 1;
1304                 ignore = 1;
1305                 wcscpy(scanf_yset, L"%");
1306             }
1307             else
1308                 /* a position? */
1309             if (cmd == L'n') {
1310                 vsize = 0;
1311                 ignore = 1;
1312                 mpdm_push(r, MPDM_I(i - (wchar_t *) str->data));
1313             }
1314             else
1315                 /* string upto a mark */
1316             if (cmd == L'S') {
1317                 wchar_t *tmp = f;
1318
1319                 /* fill the mark upto another command */
1320                 while (*tmp) {
1321                     if (*tmp == L'%') {
1322                         tmp++;
1323
1324                         /* is it an 'n'? ignore and go on */
1325                         if (*tmp == L'n') {
1326                             tmp++;
1327                             continue;
1328                         }
1329                         else
1330                         if (*tmp == L'%')
1331                             scanf_mark[msize++] = *tmp;
1332                         else
1333                             break;
1334                     }
1335                     else
1336                         scanf_mark[msize++] = *tmp;
1337
1338                     tmp++;
1339                 }
1340
1341                 scanf_mark[msize] = L'\0';
1342             }
1343             else
1344                 /* raw set */
1345             if (cmd == L'[') {
1346                 int n = 0;
1347                 wchar_t *set = scanf_yset;
1348
1349                 /* is it an inverse set? */
1350                 if (*f == L'^') {
1351                     set = scanf_nset;
1352                     f++;
1353                 }
1354
1355                 /* first one is a ]? add it */
1356                 if (*f == L']') {
1357                     set[n++] = *f;
1358                     f++;
1359                 }
1360
1361                 /* now build the set */
1362                 for (; n < SCANF_BUF_SIZE - 1 && *f && *f != L']'; f++) {
1363                     /* is it a range? */
1364                     if (*f == L'-') {
1365                         f++;
1366
1367                         /* start or end? hyphen itself */
1368                         if (n == 0 || *f == L']')
1369                             set[n++] = L'-';
1370                         else {
1371                             /* pick previous char */
1372                             wchar_t c = set[n - 1];
1373
1374                             /* fill */
1375                             while (n < SCANF_BUF_SIZE - 1 && c < *f)
1376                                 set[n++] = ++c;
1377                         }
1378                     }
1379                     else
1380                         /* is it another command? */
1381                     if (*f == L'%') {
1382                         int i;
1383
1384                         f++;
1385                         for (i = 0; scanf_sets[i].cmd; i++) {
1386                             if (*f == scanf_sets[i].cmd) {
1387                                 set[n] = L'\0';
1388                                 wcscat(set, scanf_sets[i].yset);
1389                                 n += wcslen(scanf_sets[i].yset);
1390                                 break;
1391                             }
1392                         }
1393                     }
1394                     else
1395                         set[n++] = *f;
1396                 }
1397
1398                 /* skip the ] */
1399                 f++;
1400
1401                 set[n] = L'\0';
1402             }
1403             else
1404                 /* a standard set? */
1405             {
1406                 int n;
1407
1408                 for (n = 0; scanf_sets[n].cmd != L'\0'; n++) {
1409                     if (cmd == scanf_sets[n].cmd) {
1410                         wcscpy(scanf_yset, scanf_sets[n].yset);
1411                         wcscpy(scanf_nset, scanf_sets[n].nset);
1412                         break;
1413                     }
1414                 }
1415             }
1416
1417             /* now fill the dynamic string */
1418             while (vsize &&
1419                    !wcschr(scanf_nset, *i) &&
1420                    (scanf_yset[0] == L'\0' || wcschr(scanf_yset, *i)) &&
1421                    (msize == 0 || wcsncmp(i, scanf_mark, msize) != 0)) {
1422
1423                 /* only add if not being ignored */
1424                 if (!ignore)
1425                     ptr = mpdm_poke(ptr, &size, i, 1, sizeof(wchar_t));
1426
1427                 i++;
1428                 vsize--;
1429             }
1430
1431             if (!ignore && size) {
1432                 /* null terminate and push */
1433                 ptr = mpdm_poke(ptr, &size, L"", 1, sizeof(wchar_t));
1434                 mpdm_push(r, MPDM_ENS(ptr, size - 1));
1435             }
1436         }
1437         else
1438         if (*f == L' ' || *f == L'\t') {
1439             /* if it's a blank, sync to next non-blank */
1440             f++;
1441
1442             while (*i == L' ' || *i == L'\t')
1443                 i++;
1444         }
1445         else
1446             /* test for literals in the format string */
1447         if (*i == *f) {
1448             i++;
1449             f++;
1450         }
1451         else
1452             break;
1453     }
1454
1455     mpdm_unref(str);
1456     mpdm_unref(fmt);
1457
1458     mpdm_unrefnd(r);
1459
1460     return r;
1461 }