filter.c

   1 /*@ S-nail - a mail user agent derived from Berkeley Mail.
   2  *@ Filter objects.
   3  *
   4  * Copyright (c) 2013 Steffen "Daode" Nurpmeso <sdaoden@users.sf.net>.
   5  *
   6  * Permission to use, copy, modify, and/or distribute this software for any
   7  * purpose with or without fee is hereby granted, provided that the above
   8  * copyright notice and this permission notice appear in all copies.
   9  *
  10  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
  11  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  12  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
  13  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  14  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  15  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  16  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  17  */
  18
  19 #ifndef HAVE_AMALGAMATION
  20 # include "nail.h"
  21 #endif
  22
  23 /*
  24  * Quotation filter
  25  */
  26
  27 /*
  28  * TODO quotation filter: anticipate in future data: don't break if only WS
  29  * TODO or a LF escaping \ follows on the line (simply reuse the latter).
  30  */
  31
  32 #ifdef HAVE_QUOTE_FOLD
  33 CTA(QUOTE_MAX > 3);
  34
  35 enum qf_state {
  36    _QF_CLEAN,
  37    _QF_PREFIX,
  38    _QF_DATA
  39 };
  40
  41 struct qf_vc {
  42    struct quoteflt *self;
  43    char const *buf;
  44    size_t len;
  45 };
  46
  47 /* Print out prefix and current quote */
  48 static ssize_t _qf_dump_prefix(struct quoteflt *self);
  49
  50 /* Add one data character */
  51 static ssize_t _qf_add_data(struct quoteflt *self, wchar_t wc);
  52
  53 /* State machine handlers */
  54 static ssize_t _qf_state_prefix(struct qf_vc *vc);
  55 static ssize_t _qf_state_data(struct qf_vc *vc);
  56
  57 static ssize_t
  58 _qf_dump_prefix(struct quoteflt *self)
  59 {
  60    ssize_t rv;
  61    size_t i;
  62
  63    if ((i = self->qf_pfix_len) > 0 && i != fwrite(self->qf_pfix, 1, i,
  64          self->qf_os))
  65       goto jerr;
  66    rv = i;
  67
  68    if ((i = self->qf_currq.l) > 0 && i != fwrite(self->qf_currq.s, 1, i,
  69          self->qf_os))
  70       goto jerr;
  71    rv += i;
  72 jleave:
  73    return rv;
  74 jerr:
  75    rv = -1;
  76    goto jleave;
  77 }
  78
  79 static ssize_t
  80 _qf_add_data(struct quoteflt *self, wchar_t wc)
  81 {
  82    char *save_b;
  83    ui_it save_l, save_w;
  84    ssize_t rv = 0;
  85    int w, l;
  86
  87    save_l = save_w = 0; /* silence cc */
  88    save_b = NULL;
  89    /* <newline> ends state */
  90    if (wc == L'\n')
  91       goto jflush;
  92    if (wc == L'\r') /* TODO CR should be stripped in lower level!! */
  93       goto jleave;
  94
  95    /* Unroll <tab> to spaces */
  96    if (wc == L'\t') {
  97       save_l = self->qf_datw;
  98       save_w = (save_l + QUOTE_TAB_SPACES) & ~(QUOTE_TAB_SPACES - 1);
  99       save_w -= save_l;
 100       while (save_w-- > 0) {
 101          ssize_t j = _qf_add_data(self, L' ');
 102          if (j < 0) {
 103             rv = j;
 104             break;
 105          }
 106          rv += j;
 107       }
 108       goto jleave;
 109    }
 110
 111    w = wcwidth(wc);
 112    if (w == -1) {
 113 jbad:
 114       ++self->qf_datw;
 115       self->qf_dat.s[self->qf_dat.l++] = '?';
 116    } else {
 117       l = wctomb(self->qf_dat.s + self->qf_dat.l, wc);
 118       if (l < 0)
 119          goto jbad;
 120       self->qf_datw += (ui_it)w;
 121       self->qf_dat.l += (size_t)l;
 122    }
 123
 124    /* TODO The last visual may excess *qfold-max* if it's a wide one;
 125     * TODO place it on the next line, break before */
 126    if (self->qf_datw >= self->qf_qfold_max) {
 127       /* If we have seen a nice breakpoint during traversal, shuffle data
 128        * around a bit so as to restore the trailing part after flushing */
 129       if (self->qf_brkl > 0) {
 130          save_w = self->qf_datw - self->qf_brkw;
 131          save_l = self->qf_dat.l - self->qf_brkl;
 132          save_b = self->qf_dat.s + self->qf_brkl + 2;
 133          memmove(save_b, save_b - 2, save_l);
 134          self->qf_dat.l = self->qf_brkl;
 135       }
 136
 137       self->qf_dat.s[self->qf_dat.l++] = '\\';
 138 jflush:
 139       self->qf_dat.s[self->qf_dat.l++] = '\n';
 140       rv = quoteflt_flush(self);
 141
 142       /* Restore takeovers, if any */
 143       if (save_b != NULL) {
 144          self->qf_brk_isws = FAL0;
 145          self->qf_datw += save_w;
 146          self->qf_dat.l = save_l;
 147          memmove(self->qf_dat.s, save_b, save_l);
 148       }
 149
 150    } else if (self->qf_datw >= self->qf_qfold_min && ! self->qf_brk_isws) {
 151       bool_t isws = iswspace(wc);
 152
 153       if ((isws && ! self->qf_brk_isws) || self->qf_brkl == 0) {
 154          self->qf_brkl = self->qf_dat.l;
 155          self->qf_brkw = self->qf_datw;
 156          self->qf_brk_isws = isws;
 157       }
 158    }
 159
 160    /* If state changed to prefix, perform full reset (note this implies that
 161     * quoteflt_flush() performs too much work..) */
 162    if (wc == '\n') {
 163       self->qf_state = _QF_PREFIX;
 164       self->qf_wscnt = self->qf_datw = 0;
 165       self->qf_currq.l = 0;
 166    }
 167 jleave:
 168    return rv;
 169 }
 170
 171 static ssize_t
 172 _qf_state_prefix(struct qf_vc *vc)
 173 {
 174    struct quoteflt *self = vc->self;
 175    ssize_t rv = 0;
 176    char const *buf;
 177    size_t len, i;
 178    wchar_t wc;
 179
 180    for (buf = vc->buf, len = vc->len; len > 0;) {
 181       /* TODO NULL BYTE! */
 182       i = mbrtowc(&wc, buf, len, self->qf_mbps);
 183       if (i == (size_t)-1) {
 184          /* On hard error, don't modify mbstate_t and step one byte */
 185          self->qf_mbps[0] = self->qf_mbps[1];
 186          ++buf;
 187          --len;
 188          self->qf_wscnt = 0;
 189          continue;
 190       }
 191       self->qf_mbps[1] = self->qf_mbps[0];
 192       if (i == (size_t)-2) {
 193          /* Redundant shift sequence, out of buffer */
 194          len = 0;
 195          break;
 196       }
 197       buf += i;
 198       len -= i;
 199
 200       if (wc == L'\n')
 201          goto jfin;
 202       if (iswspace(wc)) {
 203          ++self->qf_wscnt;
 204          continue;
 205       }
 206       if (i == 1 && ISQUOTE(wc)) {
 207          self->qf_wscnt = 0;
 208          if (self->qf_currq.l >= QUOTE_MAX - 3) {
 209             self->qf_currq.s[QUOTE_MAX - 3] = '.';
 210             self->qf_currq.s[QUOTE_MAX - 2] = '.';
 211             self->qf_currq.s[QUOTE_MAX - 1] = '.';
 212             self->qf_currq.l = QUOTE_MAX;
 213          } else
 214             self->qf_currq.s[self->qf_currq.l++] = buf[-1];
 215          continue;
 216       }
 217
 218       /* The quote is parsed and compressed; dump it */
 219 jfin:
 220       self->qf_datw = self->qf_pfix_len + self->qf_currq.l;
 221       self->qf_state = _QF_DATA;
 222       /* Overtake WS (xxx but we de-facto "normalize" to ASCII SP here) */
 223       while (self->qf_wscnt-- > 0 && self->qf_currq.l < QUOTE_MAX)
 224          self->qf_currq.s[self->qf_currq.l++] = ' ';
 225       self->qf_wscnt = 0;
 226       rv = _qf_add_data(self, wc);
 227       break;
 228    }
 229
 230    vc->buf = buf;
 231    vc->len = len;
 232    return rv;
 233 }
 234
 235 static ssize_t
 236 _qf_state_data(struct qf_vc *vc)
 237 {
 238    struct quoteflt *self = vc->self;
 239    ssize_t rv = 0;
 240    char const *buf;
 241    size_t len, i;
 242    wchar_t wc;
 243
 244    for (buf = vc->buf, len = vc->len; len > 0;) {
 245       /* TODO NULL BYTE! */
 246       i = mbrtowc(&wc, buf, len, self->qf_mbps);
 247       if (i == (size_t)-1) {
 248          /* On hard error, don't modify mbstate_t and step one byte */
 249          self->qf_mbps[0] = self->qf_mbps[1];
 250          ++buf;
 251          --len;
 252          continue;
 253       }
 254       self->qf_mbps[1] = self->qf_mbps[0];
 255       if (i == (size_t)-2) {
 256          /* Redundant shift sequence, out of buffer */
 257          len = 0;
 258          break;
 259       }
 260       buf += i;
 261       len -= i;
 262
 263       {  ssize_t j = _qf_add_data(self, wc);
 264          if (j < 0) {
 265             rv = j;
 266             break;
 267          }
 268          rv += j;
 269       }
 270
 271       if (self->qf_state != _QF_DATA)
 272          break;
 273    }
 274
 275    vc->buf = buf;
 276    vc->len = len;
 277    return rv;
 278 }
 279 #endif /* HAVE_QUOTE_FOLD */
 280
 281 FL struct quoteflt *
 282 quoteflt_dummy(void) /* TODO LEGACY */
 283 {
 284    static struct quoteflt qf_i;
 285
 286    return &qf_i;
 287 }
 288
 289 FL void
 290 quoteflt_init(struct quoteflt *self, char const *prefix)
 291 {
 292 #ifdef HAVE_QUOTE_FOLD
 293    char *xcp, *cp;
 294 #endif
 295
 296    memset(self, 0, sizeof *self);
 297
 298    if ((self->qf_pfix = prefix) != NULL)
 299       self->qf_pfix_len = (ui_it)strlen(prefix);
 300
 301    /* Check wether the user wants the more fancy quoting algorithm */
 302    /* TODO *quote-fold*: QUOTE_MAX may excess it! */
 303 #ifdef HAVE_QUOTE_FOLD
 304    if (self->qf_pfix_len > 0 && (cp = ok_vlook(quote_fold)) != NULL) {
 305       ui_it qmin, qmax = (ui_it)strtol(cp, (char**)&xcp, 10);
 306       /* These magic values ensure we don't bail :) */
 307       if (qmax < self->qf_pfix_len + 6)
 308          qmax = self->qf_pfix_len + 6;
 309       --qmax; /* The newline escape */
 310       if (cp == xcp || *xcp == '\0')
 311          qmin = (qmax >> 1) + (qmax >> 2) + (qmax >> 5);
 312       else {
 313          qmin = (ui_it)strtol(xcp + 1, NULL, 10);
 314          if (qmin < qmax >> 1)
 315             qmin = qmax >> 1;
 316          else if (qmin > qmax - 2)
 317             qmin = qmax - 2;
 318       }
 319       self->qf_qfold_min = qmin;
 320       self->qf_qfold_max = qmax;
 321
 322       /* Add pad for takeover copies, backslash and newline */
 323       self->qf_dat.s = salloc((qmax + 3) * mb_cur_max);
 324       self->qf_currq.s = salloc((QUOTE_MAX + 1) * mb_cur_max);
 325    }
 326 #endif
 327 }
 328
 329 FL void
 330 quoteflt_destroy(struct quoteflt *self) /* xxx inline */
 331 {
 332    (void)self;
 333 }
 334
 335 FL void
 336 quoteflt_reset(struct quoteflt *self, FILE *f) /* xxx inline */
 337 {
 338    self->qf_os = f;
 339 #ifdef HAVE_QUOTE_FOLD
 340    self->qf_state = _QF_CLEAN;
 341    self->qf_dat.l =
 342    self->qf_currq.l = 0;
 343    memset(self->qf_mbps, 0, sizeof self->qf_mbps);
 344 #endif
 345 }
 346
 347 FL ssize_t
 348 quoteflt_push(struct quoteflt *self, char const *dat, size_t len)
 349 {
 350    /* (xxx Ideally the actual push() [and flush()] would be functions on their
 351     * xxx own, via indirect vtbl call ..) */
 352    ssize_t rv = 0;
 353
 354    if (len == 0)
 355       goto jleave;
 356
 357    /* Bypass? XXX Finally, this filter simply should not be used, then */
 358    if (self->qf_pfix_len == 0) {
 359       if (len != fwrite(dat, 1, len, self->qf_os))
 360          goto jerr;
 361       rv = len;
 362    }
 363    /* Normal: place *indentprefix* at every BOL */
 364    else
 365 #ifdef HAVE_QUOTE_FOLD
 366       if (self->qf_qfold_max == 0)
 367 #endif
 368    {
 369       void *vp;
 370       size_t ll;
 371       bool_t pxok = (self->qf_qfold_min != 0);
 372
 373       for (;;) {
 374          if (!pxok) {
 375             ll = self->qf_pfix_len;
 376             if (ll != fwrite(self->qf_pfix, 1, ll, self->qf_os))
 377                goto jerr;
 378             rv += ll;
 379             pxok = TRU1;
 380          }
 381
 382          /* xxx Strictly speaking this is invalid, because only `/' and `.' are
 383           * xxx mandated by POSIX.1-2008 as "invariant across all locales
 384           * xxx supported"; though there is no charset known which uses this
 385           * xxx control char as part of a multibyte character; note that S-nail
 386           * XXX (and the Mail codebase as such) do not support EBCDIC */
 387          if ((vp = memchr(dat, '\n', len)) == NULL)
 388             ll = len;
 389          else {
 390             pxok = FAL0;
 391             ll = (size_t)((char*)vp - dat) + 1;
 392          }
 393
 394          if (ll != fwrite(dat, sizeof *dat, ll, self->qf_os))
 395             goto jerr;
 396          rv += ll;
 397          if ((len -= ll) == 0)
 398             break;
 399          dat += ll;
 400       }
 401
 402       self->qf_qfold_min = pxok;
 403    }
 404    /* Overly complicated, though still only line-per-line: *quote-fold*.
 405     * - If .qf_currq.l is 0, then we are in a clean state.  Reset .qf_mbps;
 406     *   TODO note this means we assume that lines start with reset escape seq,
 407     *   TODO but i don't think this is any worse than what we currently do;
 408     *   TODO in 15.0, with the value carrier, we should carry conversion states
 409     *   TODO all along, only resetting on error (or at words for header =???=);
 410     *   TODO this still is weird for error handling, but we need to act more
 411     *   TODO stream-alike (though in practice i don't think cross-line states
 412     *   TODO can be found, because of compatibility reasons; however, being
 413     *   TODO a problem rather than a solution is not a good thing (tm))
 414     * - Lookout for a newline */
 415 #ifdef HAVE_QUOTE_FOLD
 416    else {
 417       struct qf_vc vc;
 418       ssize_t i;
 419
 420       vc.self = self;
 421       vc.buf = dat;
 422       vc.len = len;
 423       while (vc.len > 0) {
 424          switch (self->qf_state) {
 425          case _QF_CLEAN:
 426          case _QF_PREFIX:
 427             i = _qf_state_prefix(&vc);
 428             break;
 429          default: /* silence cc (`i' unused) */
 430          case _QF_DATA:
 431             i = _qf_state_data(&vc);
 432             break;
 433          }
 434          if (i < 0)
 435             goto jerr;
 436          rv += i;
 437       }
 438    }
 439 #endif /* HAVE_QUOTE_FOLD */
 440
 441 jleave:
 442    return rv;
 443
 444 jerr:
 445    rv = -1;
 446    goto jleave;
 447 }
 448
 449 FL ssize_t
 450 quoteflt_flush(struct quoteflt *self)
 451 {
 452    ssize_t rv = 0;
 453    UNUSED(self);
 454
 455 #ifdef HAVE_QUOTE_FOLD
 456    if (self->qf_dat.l > 0) {
 457       rv = _qf_dump_prefix(self);
 458       if (rv >= 0) {
 459          size_t i = self->qf_dat.l;
 460          if (i == fwrite(self->qf_dat.s, 1, i, self->qf_os))
 461             rv += i;
 462          else
 463             rv = -1;
 464          self->qf_dat.l = 0;
 465          self->qf_brk_isws = FAL0;
 466          self->qf_wscnt = self->qf_brkl = self->qf_brkw = 0;
 467          self->qf_datw = self->qf_pfix_len + self->qf_currq.l;
 468       }
 469    }
 470 #endif
 471    return rv;
 472 }
 473
 474 /* vim:set fenc=utf-8:s-it-mode */