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 - 2018 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 = (n_psonce & n_PSO_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 = ((n_psonce & n_PSO_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(n_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 (n_mb_cur_max > 1) {
 217          wchar_t  wc;
 218
 219          n = 1;
 220          isrepl = TRU1;
 221          if ((sz = mbtowc(&wc, cp, n_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 n_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) /* TODO <-> TTYCHARSET!! */
 291 {
 292    /* TODO: makeprint() should honour *ttycharset*.  This of course does not
 293     * TODO work with ISO C / POSIX since mbrtowc() do know about locales, not
 294     * TODO charsets, and ditto iswprint() etc. do work with the locale too.
 295     * TODO I hope S-CText can do something about that, and/or otherwise add
 296     * TODO some special treatment for UTF-8 (take it from S-CText too then) */
 297    char const *inp, *maxp;
 298    char *outp;
 299    DBG( size_t msz; )
 300    NYD_ENTER;
 301
 302    out->s =
 303    outp = smalloc(DBG( msz = ) in->l*n_mb_cur_max + 2u*n_mb_cur_max +1);
 304    inp = in->s;
 305    maxp = inp + in->l;
 306
 307 #ifdef HAVE_NATCH_CHAR
 308    if (n_mb_cur_max > 1) {
 309       char mbb[MB_LEN_MAX + 1];
 310       wchar_t wc;
 311       int i, n;
 312       bool_t isuni = ((n_psonce & n_PSO_UNICODE) != 0);
 313
 314       out->l = 0;
 315       while (inp < maxp) {
 316          if (*inp & 0200)
 317             n = mbtowc(&wc, inp, PTR2SIZE(maxp - inp));
 318          else {
 319             wc = *inp;
 320             n = 1;
 321          }
 322          if (n == -1) {
 323             /* FIXME Why mbtowc() resetting here?
 324              * FIXME what about ISO 2022-JP plus -- those
 325              * FIXME will loose shifts, then!
 326              * FIXME THUS - we'd need special "known points"
 327              * FIXME to do so - say, after a newline!!
 328              * FIXME WE NEED TO CHANGE ALL USES +MBLEN! */
 329             mbtowc(&wc, NULL, n_mb_cur_max);
 330             wc = isuni ? 0xFFFD : '?';
 331             n = 1;
 332          } else if (n == 0)
 333             n = 1;
 334          inp += n;
 335          if (!iswprint(wc) && wc != '\n' /*&& wc != '\r' && wc != '\b'*/ &&
 336                wc != '\t') {
 337             if ((wc & ~(wchar_t)037) == 0)
 338                wc = isuni ? 0x2400 | wc : '?';
 339             else if (wc == 0177)
 340                wc = isuni ? 0x2421 : '?';
 341             else
 342                wc = isuni ? 0x2426 : '?';
 343          }else if(isuni){ /* TODO ctext */
 344             /* Need to filter out L-TO-R and R-TO-R marks TODO ctext */
 345             if(wc == 0x200E || wc == 0x200F || (wc >= 0x202A && wc <= 0x202E))
 346                continue;
 347             /* And some zero-width messes */
 348             if(wc == 0x00AD || (wc >= 0x200B && wc <= 0x200D))
 349                continue;
 350             /* Oh about the ISO C wide character interfaces, baby! */
 351             if(wc == 0xFEFF)
 352                continue;
 353          }
 354          if ((n = wctomb(mbb, wc)) <= 0)
 355             continue;
 356          out->l += n;
 357          assert(out->l < msz);
 358          for (i = 0; i < n; ++i)
 359             *outp++ = mbb[i];
 360       }
 361    } else
 362 #endif /* NATCH_CHAR */
 363    {
 364       int c;
 365       while (inp < maxp) {
 366          c = *inp++ & 0377;
 367          if (!isprint(c) && c != '\n' && c != '\r' && c != '\b' && c != '\t')
 368             c = '?';
 369          *outp++ = c;
 370       }
 371       out->l = in->l;
 372    }
 373    out->s[out->l] = '\0';
 374    NYD_LEAVE;
 375 }
 376
 377 FL size_t
 378 delctrl(char *cp, size_t len)
 379 {
 380    size_t x, y;
 381    NYD_ENTER;
 382
 383    for (x = y = 0; x < len; ++x)
 384       if (!cntrlchar(cp[x]))
 385          cp[y++] = cp[x];
 386    cp[y] = '\0';
 387    NYD_LEAVE;
 388    return y;
 389 }
 390
 391 FL char *
 392 prstr(char const *s)
 393 {
 394    struct str in, out;
 395    char *rp;
 396    NYD_ENTER;
 397
 398    in.s = n_UNCONST(s);
 399    in.l = strlen(s);
 400    makeprint(&in, &out);
 401    rp = savestrbuf(out.s, out.l);
 402    free(out.s);
 403    NYD_LEAVE;
 404    return rp;
 405 }
 406
 407 FL int
 408 prout(char const *s, size_t sz, FILE *fp)
 409 {
 410    struct str in, out;
 411    int n;
 412    NYD_ENTER;
 413
 414    in.s = n_UNCONST(s);
 415    in.l = sz;
 416    makeprint(&in, &out);
 417    n = fwrite(out.s, 1, out.l, fp);
 418    free(out.s);
 419    NYD_LEAVE;
 420    return n;
 421 }
 422
 423 FL size_t
 424 putuc(int u, int c, FILE *fp)
 425 {
 426    size_t rv;
 427    NYD_ENTER;
 428    n_UNUSED(u);
 429
 430 #ifdef HAVE_NATCH_CHAR
 431    if ((n_psonce & n_PSO_UNICODE) && (u & ~(wchar_t)0177)) {
 432       char mbb[MB_LEN_MAX];
 433       int i, n;
 434
 435       if ((n = wctomb(mbb, u)) > 0) {
 436          rv = wcwidth(u);
 437          for (i = 0; i < n; ++i)
 438             if (putc(mbb[i] & 0377, fp) == EOF) {
 439                rv = 0;
 440                break;
 441             }
 442       } else if (n == 0)
 443          rv = (putc('\0', fp) != EOF);
 444       else
 445          rv = 0;
 446    } else
 447 #endif
 448       rv = (putc(c, fp) != EOF);
 449    NYD_LEAVE;
 450    return rv;
 451 }
 452
 453 FL bool_t
 454 bidi_info_needed(char const *bdat, size_t blen)
 455 {
 456    bool_t rv = FAL0;
 457    NYD_ENTER;
 458
 459 #ifdef HAVE_NATCH_CHAR
 460    if (n_psonce & n_PSO_UNICODE)
 461       while (blen > 0) {
 462          /* TODO Checking for BIDI character: use S-CText fromutf8
 463           * TODO plus isrighttoleft (or whatever there will be)! */
 464          ui32_t c = n_utf8_to_utf32(&bdat, &blen);
 465          if (c == UI32_MAX)
 466             break;
 467
 468          if (c <= 0x05BE)
 469             continue;
 470
 471          /* (Very very fuzzy, awaiting S-CText for good) */
 472          if ((c >= 0x05BE && c <= 0x08E3) ||
 473                (c >= 0xFB1D && c <= 0xFE00) /* No: variation selectors */ ||
 474                (c >= 0xFE70 && c <= 0xFEFC) ||
 475                (c >= 0x10800 && c <= 0x10C48) ||
 476                (c >= 0x1EE00 && c <= 0x1EEF1)) {
 477             rv = TRU1;
 478             break;
 479          }
 480       }
 481 #endif /* HAVE_NATCH_CHAR */
 482    NYD_LEAVE;
 483    return rv;
 484 }
 485
 486 FL void
 487 bidi_info_create(struct bidi_info *bip)
 488 {
 489    /* Unicode: how to isolate RIGHT-TO-LEFT scripts via *headline-bidi*
 490     * 1.1 (Jun 1993): U+200E (E2 80 8E) LEFT-TO-RIGHT MARK
 491     * 6.3 (Sep 2013): U+2068 (E2 81 A8) FIRST STRONG ISOLATE,
 492     *                 U+2069 (E2 81 A9) POP DIRECTIONAL ISOLATE
 493     * Worse results seen for: U+202D "\xE2\x80\xAD" U+202C "\xE2\x80\xAC" */
 494    n_NATCH_CHAR( char const *hb; )
 495    NYD_ENTER;
 496
 497    memset(bip, 0, sizeof *bip);
 498    bip->bi_start.s = bip->bi_end.s = n_UNCONST(n_empty);
 499
 500 #ifdef HAVE_NATCH_CHAR
 501    if ((n_psonce & n_PSO_UNICODE) && (hb = ok_vlook(headline_bidi)) != NULL) {
 502       switch (*hb) {
 503       case '3':
 504          bip->bi_pad = 2;
 505          /* FALLTHRU */
 506       case '2':
 507          bip->bi_start.s = bip->bi_end.s = n_UNCONST("\xE2\x80\x8E");
 508          break;
 509       case '1':
 510          bip->bi_pad = 2;
 511          /* FALLTHRU */
 512       default:
 513          bip->bi_start.s = n_UNCONST("\xE2\x81\xA8");
 514          bip->bi_end.s = n_UNCONST("\xE2\x81\xA9");
 515          break;
 516       }
 517       bip->bi_start.l = bip->bi_end.l = 3;
 518    }
 519 #endif
 520    NYD_LEAVE;
 521 }
 522
 523 /* s-it-mode */