ui_str.c

   1 /*@ S-nail - a mail user agent derived from Berkeley Mail.
   2  *@ UserInterface: string related operations.
   3  *
   4  * Copyright (c) 2000-2004 Gunnar Ritter, Freiburg i. Br., Germany.
   5  * Copyright (c) 2012 - 2016 Steffen (Daode) Nurpmeso <steffen@sdaoden.eu>.
   6  */
   7 /*
   8  * Redistribution and use in source and binary forms, with or without
   9  * modification, are permitted provided that the following conditions
  10  * are met:
  11  * 1. Redistributions of source code must retain the above copyright
  12  *    notice, this list of conditions and the following disclaimer.
  13  * 2. Redistributions in binary form must reproduce the above copyright
  14  *    notice, this list of conditions and the following disclaimer in the
  15  *    documentation and/or other materials provided with the distribution.
  16  * 3. Neither the name of the University nor the names of its contributors
  17  *    may be used to endorse or promote products derived from this software
  18  *    without specific prior written permission.
  19  *
  20  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  21  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  22  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  23  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  24  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  25  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  26  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  27  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  28  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  29  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  30  * SUCH DAMAGE.
  31  */
  32 #undef n_FILE
  33 #define n_FILE ui_str
  34
  35 #ifndef HAVE_AMALGAMATION
  36 # include "nail.h"
  37 #endif
  38
  39 #include <ctype.h>
  40
  41 FL bool_t
  42 n_visual_info(struct n_visual_info_ctx *vicp, enum n_visual_info_flags vif){
  43 #ifdef HAVE_C90AMEND1
  44    mbstate_t *mbp;
  45 #endif
  46    size_t il;
  47    char const *ib;
  48    bool_t rv;
  49    NYD2_ENTER;
  50
  51    assert(vicp != NULL);
  52    assert(vicp->vic_inlen == 0 || vicp->vic_indat != NULL);
  53    assert(!(vif & n__VISUAL_INFO_FLAGS) || !(vif & n_VISUAL_INFO_ONE_CHAR));
  54
  55    rv = TRU1;
  56    ib = vicp->vic_indat;
  57    if((il = vicp->vic_inlen) == UIZ_MAX)
  58       il = vicp->vic_inlen = strlen(ib);
  59
  60    if((vif & (n_VISUAL_INFO_WIDTH_QUERY | n_VISUAL_INFO_WOUT_PRINTABLE)) ==
  61          n_VISUAL_INFO_WOUT_PRINTABLE)
  62       vif |= n_VISUAL_INFO_WIDTH_QUERY;
  63
  64    vicp->vic_chars_seen = vicp->vic_bytes_seen = vicp->vic_vi_width = 0;
  65    if(vif & n_VISUAL_INFO_WOUT_CREATE){
  66       if(vif & n_VISUAL_INFO_WOUT_SALLOC)
  67          vicp->vic_woudat = salloc(sizeof(*vicp->vic_woudat) * (il +1));
  68       vicp->vic_woulen = 0;
  69    }
  70 #ifdef HAVE_C90AMEND1
  71    if((mbp = vicp->vic_mbstate) == NULL)
  72       mbp = &vicp->vic_mbs_def;
  73 #endif
  74
  75    if(il > 0){
  76       do/* while(!(vif & n_VISUAL_INFO_ONE_CHAR) && il > 0) */{
  77 #ifdef HAVE_C90AMEND1
  78          size_t i = mbrtowc(&vicp->vic_waccu, ib, il, mbp);
  79
  80          if(i == (size_t)-2){
  81             rv = FAL0;
  82             break;
  83          }else if(i == (size_t)-1){
  84             if(!(vif & n_VISUAL_INFO_SKIP_ERRORS)){
  85                rv = FAL0;
  86                break;
  87             }
  88             memset(mbp, 0, sizeof *mbp);
  89             vicp->vic_waccu = (options & OPT_UNICODE) ? 0xFFFD : '?';
  90             i = 1;
  91          }else if(i == 0){
  92             il = 0;
  93             break;
  94          }
  95
  96          ++vicp->vic_chars_seen;
  97          vicp->vic_bytes_seen += i;
  98          ib += i;
  99          il -= i;
 100
 101          if(vif & n_VISUAL_INFO_WIDTH_QUERY){
 102             int w;
 103             wchar_t wc = vicp->vic_waccu;
 104
 105 # ifdef HAVE_WCWIDTH
 106             w = (wc == '\t' ? 1 : wcwidth(wc));
 107 # else
 108             if(wc == '\t' || iswprint(wc))
 109                w = 1 + (wc >= 0x1100u); /* S-CText isfullwidth() */
 110             else
 111                w = -1;
 112 # endif
 113             if(w > 0)
 114                vicp->vic_vi_width += w;
 115             else if(vif & n_VISUAL_INFO_WOUT_PRINTABLE)
 116                continue;
 117          }
 118 #else /* HAVE_C90AMEND1 */
 119          char c = *ib;
 120
 121          if(c == '\0'){
 122             il = 0;
 123             break;
 124          }
 125
 126          ++vicp->vic_chars_seen;
 127          ++vicp->vic_bytes_seen;
 128          vicp->vic_waccu = c;
 129          if(vif & n_VISUAL_INFO_WIDTH_QUERY)
 130             vicp->vic_vi_width += (c == '\t' || isprint(c)); /* XXX */
 131
 132          ++ib;
 133          --il;
 134 #endif
 135
 136          if(vif & n_VISUAL_INFO_WOUT_CREATE)
 137             vicp->vic_woudat[vicp->vic_woulen++] = vicp->vic_waccu;
 138       }while(!(vif & n_VISUAL_INFO_ONE_CHAR) && il > 0);
 139    }
 140
 141    if(vif & n_VISUAL_INFO_WOUT_CREATE)
 142       vicp->vic_woudat[vicp->vic_woulen] = L'\0';
 143    vicp->vic_oudat = ib;
 144    vicp->vic_oulen = il;
 145    vicp->vic_flags = vif;
 146    NYD2_LEAVE;
 147    return rv;
 148 }
 149
 150 FL size_t
 151 field_detect_clip(size_t maxlen, char const *buf, size_t blen)/*TODO mbrtowc()*/
 152 {
 153    size_t rv;
 154    NYD_ENTER;
 155
 156 #ifdef HAVE_NATCH_CHAR
 157    maxlen = n_MIN(maxlen, blen);
 158    for (rv = 0; maxlen > 0;) {
 159       int ml = mblen(buf, maxlen);
 160       if (ml <= 0) {
 161          mblen(NULL, 0);
 162          break;
 163       }
 164       buf += ml;
 165       rv += ml;
 166       maxlen -= ml;
 167    }
 168 #else
 169    rv = n_MIN(blen, maxlen);
 170 #endif
 171    NYD_LEAVE;
 172    return rv;
 173 }
 174
 175 FL char *
 176 colalign(char const *cp, int col, int fill, int *cols_decr_used_or_null)
 177 {
 178    n_NATCH_CHAR( struct bidi_info bi; )
 179    int col_orig = col, n, sz;
 180    bool_t isbidi, isuni, istab, isrepl;
 181    char *nb, *np;
 182    NYD_ENTER;
 183
 184    /* Bidi only on request and when there is 8-bit data */
 185    isbidi = isuni = FAL0;
 186 #ifdef HAVE_NATCH_CHAR
 187    isuni = ((options & OPT_UNICODE) != 0);
 188    bidi_info_create(&bi);
 189    if (bi.bi_start.l == 0)
 190       goto jnobidi;
 191    if (!(isbidi = bidi_info_needed(cp, strlen(cp))))
 192       goto jnobidi;
 193
 194    if ((size_t)col >= bi.bi_pad)
 195       col -= bi.bi_pad;
 196    else
 197       col = 0;
 198 jnobidi:
 199 #endif
 200
 201    np = nb = salloc(mb_cur_max * strlen(cp) +
 202          ((fill ? col : 0)
 203          n_NATCH_CHAR( + (isbidi ? bi.bi_start.l + bi.bi_end.l : 0) )
 204          +1));
 205
 206 #ifdef HAVE_NATCH_CHAR
 207    if (isbidi) {
 208       memcpy(np, bi.bi_start.s, bi.bi_start.l);
 209       np += bi.bi_start.l;
 210    }
 211 #endif
 212
 213    while (*cp != '\0') {
 214       istab = FAL0;
 215 #ifdef HAVE_C90AMEND1
 216       if (mb_cur_max > 1) {
 217          wchar_t  wc;
 218
 219          n = 1;
 220          isrepl = TRU1;
 221          if ((sz = mbtowc(&wc, cp, mb_cur_max)) == -1)
 222             sz = 1;
 223          else if (wc == L'\t') {
 224             cp += sz - 1; /* Silly, no such charset known (.. until S-Ctext) */
 225             isrepl = FAL0;
 226             istab = TRU1;
 227          } else if (iswprint(wc)) {
 228 # ifndef HAVE_WCWIDTH
 229             n = 1 + (wc >= 0x1100u); /* TODO use S-CText isfullwidth() */
 230 # else
 231             if ((n = wcwidth(wc)) == -1)
 232                n = 1;
 233             else
 234 # endif
 235                isrepl = FAL0;
 236          }
 237       } else
 238 #endif
 239       {
 240          n = sz = 1;
 241          istab = (*cp == '\t');
 242          isrepl = !(istab || isprint((uc_i)*cp));
 243       }
 244
 245       if (n > col)
 246          break;
 247       col -= n;
 248
 249       if (isrepl) {
 250          if (isuni) {
 251             /* Contained in mb_cur_max, then */
 252             memcpy(np, n_unirepl, sizeof(n_unirepl) -1);
 253             np += sizeof(n_unirepl) -1;
 254          } else
 255             *np++ = '?';
 256          cp += sz;
 257       } else if (istab || (sz == 1 && spacechar(*cp))) {
 258          *np++ = ' ';
 259          ++cp;
 260       } else
 261          while (sz--)
 262             *np++ = *cp++;
 263    }
 264
 265    if (fill && col != 0) {
 266       if (fill > 0) {
 267          memmove(nb + col, nb, PTR2SIZE(np - nb));
 268          memset(nb, ' ', col);
 269       } else
 270          memset(np, ' ', col);
 271       np += col;
 272       col = 0;
 273    }
 274
 275 #ifdef HAVE_NATCH_CHAR
 276    if (isbidi) {
 277       memcpy(np, bi.bi_end.s, bi.bi_end.l);
 278       np += bi.bi_end.l;
 279    }
 280 #endif
 281
 282    *np = '\0';
 283    if (cols_decr_used_or_null != NULL)
 284       *cols_decr_used_or_null -= col_orig - col;
 285    NYD_LEAVE;
 286    return nb;
 287 }
 288
 289 FL void
 290 makeprint(struct str const *in, struct str *out)
 291 {
 292    char const *inp, *maxp;
 293    char *outp;
 294    DBG( size_t msz; )
 295    NYD_ENTER;
 296
 297    out->s = outp = smalloc(DBG( msz = ) in->l*mb_cur_max + 2u*mb_cur_max +1);
 298    inp = in->s;
 299    maxp = inp + in->l;
 300
 301 #ifdef HAVE_NATCH_CHAR
 302    if (mb_cur_max > 1) {
 303       char mbb[MB_LEN_MAX + 1];
 304       wchar_t wc;
 305       int i, n;
 306       bool_t isuni = ((options & OPT_UNICODE) != 0);
 307
 308       out->l = 0;
 309       while (inp < maxp) {
 310          if (*inp & 0200)
 311             n = mbtowc(&wc, inp, PTR2SIZE(maxp - inp));
 312          else {
 313             wc = *inp;
 314             n = 1;
 315          }
 316          if (n == -1) {
 317             /* FIXME Why mbtowc() resetting here?
 318              * FIXME what about ISO 2022-JP plus -- those
 319              * FIXME will loose shifts, then!
 320              * FIXME THUS - we'd need special "known points"
 321              * FIXME to do so - say, after a newline!!
 322              * FIXME WE NEED TO CHANGE ALL USES +MBLEN! */
 323             mbtowc(&wc, NULL, mb_cur_max);
 324             wc = isuni ? 0xFFFD : '?';
 325             n = 1;
 326          } else if (n == 0)
 327             n = 1;
 328          inp += n;
 329          if (!iswprint(wc) && wc != '\n' && wc != '\r' && wc != '\b' &&
 330                wc != '\t') {
 331             if ((wc & ~(wchar_t)037) == 0)
 332                wc = isuni ? 0x2400 | wc : '?';
 333             else if (wc == 0177)
 334                wc = isuni ? 0x2421 : '?';
 335             else
 336                wc = isuni ? 0x2426 : '?';
 337          }else if(isuni){ /* TODO ctext */
 338             /* We need to actively filter out L-TO-R and R-TO-R marks TODO ctext */
 339             if(wc == 0x200E || wc == 0x200F || (wc >= 0x202A && wc <= 0x202E))
 340                continue;
 341             /* And some zero-width messes */
 342             if(wc == 0x00AD || (wc >= 0x200B && wc <= 0x200D))
 343                continue;
 344             /* Oh about the ISO C wide character interfaces, baby! */
 345             if(wc == 0xFEFF)
 346                continue;
 347          }
 348          if ((n = wctomb(mbb, wc)) <= 0)
 349             continue;
 350          out->l += n;
 351          assert(out->l < msz);
 352          for (i = 0; i < n; ++i)
 353             *outp++ = mbb[i];
 354       }
 355    } else
 356 #endif /* NATCH_CHAR */
 357    {
 358       int c;
 359       while (inp < maxp) {
 360          c = *inp++ & 0377;
 361          if (!isprint(c) && c != '\n' && c != '\r' && c != '\b' && c != '\t')
 362             c = '?';
 363          *outp++ = c;
 364       }
 365       out->l = in->l;
 366    }
 367    out->s[out->l] = '\0';
 368    NYD_LEAVE;
 369 }
 370
 371 FL size_t
 372 delctrl(char *cp, size_t len)
 373 {
 374    size_t x, y;
 375    NYD_ENTER;
 376
 377    for (x = y = 0; x < len; ++x)
 378       if (!cntrlchar(cp[x]))
 379          cp[y++] = cp[x];
 380    cp[y] = '\0';
 381    NYD_LEAVE;
 382    return y;
 383 }
 384
 385 FL char *
 386 prstr(char const *s)
 387 {
 388    struct str in, out;
 389    char *rp;
 390    NYD_ENTER;
 391
 392    in.s = n_UNCONST(s);
 393    in.l = strlen(s);
 394    makeprint(&in, &out);
 395    rp = savestrbuf(out.s, out.l);
 396    free(out.s);
 397    NYD_LEAVE;
 398    return rp;
 399 }
 400
 401 FL int
 402 prout(char const *s, size_t sz, FILE *fp)
 403 {
 404    struct str in, out;
 405    int n;
 406    NYD_ENTER;
 407
 408    in.s = n_UNCONST(s);
 409    in.l = sz;
 410    makeprint(&in, &out);
 411    n = fwrite(out.s, 1, out.l, fp);
 412    free(out.s);
 413    NYD_LEAVE;
 414    return n;
 415 }
 416
 417 FL size_t
 418 putuc(int u, int c, FILE *fp)
 419 {
 420    size_t rv;
 421    NYD_ENTER;
 422    n_UNUSED(u);
 423
 424 #ifdef HAVE_NATCH_CHAR
 425    if ((options & OPT_UNICODE) && (u & ~(wchar_t)0177)) {
 426       char mbb[MB_LEN_MAX];
 427       int i, n;
 428
 429       if ((n = wctomb(mbb, u)) > 0) {
 430          rv = wcwidth(u);
 431          for (i = 0; i < n; ++i)
 432             if (putc(mbb[i] & 0377, fp) == EOF) {
 433                rv = 0;
 434                break;
 435             }
 436       } else if (n == 0)
 437          rv = (putc('\0', fp) != EOF);
 438       else
 439          rv = 0;
 440    } else
 441 #endif
 442       rv = (putc(c, fp) != EOF);
 443    NYD_LEAVE;
 444    return rv;
 445 }
 446
 447 FL bool_t
 448 bidi_info_needed(char const *bdat, size_t blen)
 449 {
 450    bool_t rv = FAL0;
 451    NYD_ENTER;
 452
 453 #ifdef HAVE_NATCH_CHAR
 454    if (options & OPT_UNICODE)
 455       while (blen > 0) {
 456          /* TODO Checking for BIDI character: use S-CText fromutf8
 457           * TODO plus isrighttoleft (or whatever there will be)! */
 458          ui32_t c = n_utf8_to_utf32(&bdat, &blen);
 459          if (c == UI32_MAX)
 460             break;
 461
 462          if (c <= 0x05BE)
 463             continue;
 464
 465          /* (Very very fuzzy, awaiting S-CText for good) */
 466          if ((c >= 0x05BE && c <= 0x08E3) ||
 467                (c >= 0xFB1D && c <= 0xFE00) /* No: variation selectors */ ||
 468                (c >= 0xFE70 && c <= 0xFEFC) ||
 469                (c >= 0x10800 && c <= 0x10C48) ||
 470                (c >= 0x1EE00 && c <= 0x1EEF1)) {
 471             rv = TRU1;
 472             break;
 473          }
 474       }
 475 #endif /* HAVE_NATCH_CHAR */
 476    NYD_LEAVE;
 477    return rv;
 478 }
 479
 480 FL void
 481 bidi_info_create(struct bidi_info *bip)
 482 {
 483    /* Unicode: how to isolate RIGHT-TO-LEFT scripts via *headline-bidi*
 484     * 1.1 (Jun 1993): U+200E (E2 80 8E) LEFT-TO-RIGHT MARK
 485     * 6.3 (Sep 2013): U+2068 (E2 81 A8) FIRST STRONG ISOLATE,
 486     *                 U+2069 (E2 81 A9) POP DIRECTIONAL ISOLATE
 487     * Worse results seen for: U+202D "\xE2\x80\xAD" U+202C "\xE2\x80\xAC" */
 488    n_NATCH_CHAR( char const *hb; )
 489    NYD_ENTER;
 490
 491    memset(bip, 0, sizeof *bip);
 492    bip->bi_start.s = bip->bi_end.s = n_UNCONST(n_empty);
 493
 494 #ifdef HAVE_NATCH_CHAR
 495    if ((options & OPT_UNICODE) && (hb = ok_vlook(headline_bidi)) != NULL) {
 496       switch (*hb) {
 497       case '3':
 498          bip->bi_pad = 2;
 499          /* FALLTHRU */
 500       case '2':
 501          bip->bi_start.s = bip->bi_end.s = n_UNCONST("\xE2\x80\x8E");
 502          break;
 503       case '1':
 504          bip->bi_pad = 2;
 505          /* FALLTHRU */
 506       default:
 507          bip->bi_start.s = n_UNCONST("\xE2\x81\xA8");
 508          bip->bi_end.s = n_UNCONST("\xE2\x81\xA9");
 509          break;
 510       }
 511       bip->bi_start.l = bip->bi_end.l = 3;
 512    }
 513 #endif
 514    NYD_LEAVE;
 515 }
 516
 517 /* s-it-mode */