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 #include "rcv.h"
  20
  21 #include "extern.h"
  22
  23 /*
  24  * Quotation filter
  25  */
  26
  27 #undef HAVE_QUOTE_FOLD /* TODO disabled for v14.4, not multibyte aware */
  28
  29 #ifdef HAVE_QUOTE_FOLD
  30 CTASSERT(QUOTE_MAX > 3);
  31
  32 enum qf_state {
  33    _QF_NONE,   /* Clean line, nothing read yet */
  34    _QF_PREFIX, /* Still collecting prefix */
  35    _QF_DATA    /* Prefix passed */
  36 };
  37
  38 struct qf_vc {
  39    struct quoteflt * self;
  40    char const *      buf;
  41    size_t            len;
  42 };
  43
  44 static ssize_t _qf_state_prefix(struct qf_vc *vc);
  45 static ssize_t _qf_state_data(struct qf_vc *vc);
  46
  47 #if 0
  48 static size_t
  49 prefixwrite(char const *ptr, size_t size, FILE *f,
  50    char const *prefix, size_t prefixlen)
  51 {
  52    p = ptr;
  53    maxp = p + size;
  54
  55    } else {
  56       /* After writing a real newline followed by our prefix,
  57        * compress the quoted prefixes;
  58        * note that \n is only matched by spacechar(), not by
  59        * blankchar() or blankspacechar() */
  60       /*
  61        * TODO the problem we have is again the odd 4:3 relation of
  62        * TODO base64 -- if we quote a mail that is in base64 then
  63        * TODO prefixwrite() doesn't get invoked with partial multi-
  64        * TODO byte characters (S-nail uses the `rest' mechanism to
  65        * TODO avoid that), but it may of course be invoked with a
  66        * TODO partial line, and even with multiple thereof.
  67        * TODO this has to be addressed in 15.0 with the MIME and send
  68        * TODO layer rewrite.  The solution is that `prefixwrite' has
  69        * TODO to be an object with state -- if a part is to be quoted
  70        * TODO you create it, and simply feed in data; once the part
  71        * TODO is done, you'll release it;  the object itself gobbles
  72        * TODO data unless a *hard* newline is seen.
  73        * TODO in fact we can then even implement paragraph-wise
  74        * TODO quoting! FIXME in fact that is the way: objects
  75        * TODO then, evaluate quote-fold when sending starts, ONCE!
  76        * FIXME NOTE: base64 (yet TODO for qp) may have CRLF line
  77        * FIXME endings, these need to be removed in a LOWER LAYER!!
  78        */
  79 fprintf(stderr, "ENTRY: lastlen=%zu, size=%zu <%.*s>\n",lastlen,size,(int)size,p);
  80       if ((lnlen = lastlen) != 0)
  81          goto jcontb64;
  82 jhardnl:
  83       lastlen = prefixlen;
  84       for (zipl = i = 0; p + i < maxp; ++i) {
  85          c = p[i];
  86          if (blankspacechar(c))
  87             continue;
  88          if (! ISQUOTE(c))
  89             break;
  90          if (zipl == sizeof(zipb) - 1) {
  91             zipb[sizeof(zipb) - 2] = '.';
  92             zipb[sizeof(zipb) - 3] = '.';
  93             zipb[sizeof(zipb) - 4] = '.';
  94             continue;
  95          }
  96          zipb[zipl++] = c;
  97       }
  98       zipb[zipl] = '\0';
  99       p += i;
 100 jsoftnl:
 101       if (zipl > 0) {
 102          wsz += fwrite(zipb, sizeof *zipb, zipl, f);
 103          lnlen += zipl;
 104          lastlen = lnlen;
 105       }
 106 jcontb64:
 107       /* Search forward until either *quote-fold* or NL.
 108        * In the former case try to break at whitespace,
 109        * but only if's located in the 2nd half of the data */
 110       for (c = i = 0; p + i < maxp;) {
 111          c = p[i++];
 112          if (c == '\n')
 113             break;
 114          if (lnlen + i <= qfold_max)
 115             continue;
 116
 117          /* We're excessing bounds -- but don't "continue"
 118           * trailing WS nor a continuation */
 119          if (c == '\\' || spacechar(c)) {
 120             char const *cp;
 121
 122             for (cp = p + i; cp < maxp; ++cp)
 123                if (! spacechar(*cp))
 124                   break;
 125             if (cp == maxp || (*cp == '\\' &&
 126                   cp[1] == '\n')) {
 127                i = (size_t)(maxp - p);
 128                c = 0;
 129                break;
 130             }
 131          } else if (p + i < maxp && p[i] == '\n') {
 132             ++i;
 133             c = 0;
 134             break;
 135          }
 136
 137          /* We have to fold this line */
 138          if (qfold_min < lnlen) {
 139             /* This is because of base64 and odd 4:3.
 140              * i.e., entered with some partial line yet
 141              * written..  This is weird, as we may have
 142              * written out `qfold_max' already.. */
 143             j = 0;
 144             size = --i;
 145 fprintf(stderr, "-- size=%zu p<%.*s>\n",size,(int)size,p);
 146          } else {
 147             j = qfold_min - lnlen;
 148             i =
 149             size = j + ((qfold_max - qfold_min) >> 1);
 150             assert(p + i < maxp);
 151          }
 152          while (i > j && ! spacechar(p[i - 1]))
 153             --i;
 154          if (i == j)
 155             i = size;
 156          c = 0;
 157          break;
 158       }
 159
 160       if (i > 0) {
 161          wsz += fwrite(p, sizeof *p, i, f);
 162          p += i;
 163          lastlen += i;
 164       }
 165
 166       if (p < maxp) {
 167          if (c != '\n') {
 168             putc('\\', f);
 169             putc('\n', f);
 170             wsz += 2;
 171          }
 172          wsz += fwrite(prefix, sizeof *prefix, prefixlen, f);
 173          lnlen = prefixlen;
 174          if (c != '\n')
 175             goto jsoftnl;
 176          goto jhardnl;
 177       }
 178    }
 179
 180    lastc = p[-1];
 181    return wsz;
 182 }
 183 #endif
 184
 185
 186 static ssize_t
 187 _qf_state_prefix(struct qf_vc *vc)
 188 {
 189    struct quoteflt *self = vc->self;
 190    char const *buf = vc->buf;
 191    size_t len = vc->len, i;
 192    ssize_t rv = 0;
 193    wchar_t wc;
 194
 195
 196
 197    for (;;) {
 198       if (len < mb_cur_max) {
 199          len = vc->len - len;
 200          memcpy(self->dat.s + self->dat.l, vc->buf, len);
 201          self->dat.l += len;
 202          break;
 203       }
 204
 205       i = mbrtowc(&wc, buf, len, &self->mbps);
 206       if (i == (size_t)-1 || i == (size_t)-2)
 207          ;
 208
 209       buf += i;
 210       len -= i;
 211
 212       if (iswspace(wc))
 213          continue;
 214
 215       if (i == 1 && ISQUOTE(wc)) {
 216          if (self->currq.l == QUOTE_MAX - 3) {
 217             self->currq.s[QUOTE_MAX - 3] = '.';
 218             self->currq.s[QUOTE_MAX - 2] = '.';
 219             self->currq.s[QUOTE_MAX - 1] = '.';
 220          } else
 221             self->currq.s[self->currq.l++] = buf[-1];
 222          continue;
 223       }
 224
 225       self->qf_state = _QF_DATA;
 226       break;
 227
 228
 229    }
 230
 231
 232 jleave:
 233
 234    return rv;
 235 }
 236
 237 static ssize_t
 238 _qf_state_data(struct qf_vc *vc)
 239 {
 240
 241
 242
 243
 244    return ;
 245 }
 246 #endif /* HAVE_QUOTE_FOLD */
 247
 248 struct quoteflt *
 249 quoteflt_dummy(void) /* TODO LEGACY */
 250 {
 251    static struct quoteflt qf_i;
 252
 253    return &qf_i;
 254 }
 255
 256 void
 257 quoteflt_init(struct quoteflt *self, char const *prefix)
 258 {
 259 #ifdef HAVE_QUOTE_FOLD
 260    char *xcp, *cp;
 261 #endif
 262
 263    memset(self, 0, sizeof *self);
 264
 265    if ((self->qf_pfix = prefix) != NULL)
 266       self->qf_pfix_len = (ui_it)strlen(prefix);
 267
 268    /* Check wether the user wants the more fancy quoting algorithm */
 269 #ifdef HAVE_QUOTE_FOLD
 270    if ((cp = voption("quote-fold")) != NULL) {
 271       ui_it qmin, qmax = (ui_it)strtol(cp, (char**)&xcp, 10);
 272       if (qmax < self->qf_pfix_len + 6)
 273          qmax = self->qf_pfix_len + 6;
 274       --qmax; /* The newline escape */
 275       if (cp == xcp || *xcp == '\0')
 276          qmin = (qmax >> 1) + (qmax >> 2) + (qmax >> 5);
 277       else {
 278          qmin = (ui_it)strtol(xcp + 1, NULL, 10);
 279          if (qmin < qmax >> 1)
 280             qmin = qmax >> 1;
 281          else if (qmin > qmax - 2)
 282             qmin = qmax - 2;
 283       }
 284       self->qf_qfold_min = qmin;
 285       self->qf_qfold_max = qmax;
 286
 287       self->qf_dat.s = salloc((qmax + 2) * mb_cur_len);
 288       self->qf_currq.s = salloc((QUOTE_MAX + 1) * mb_cur_len);
 289    }
 290 #endif
 291 }
 292
 293 void
 294 quoteflt_destroy(struct quoteflt *self)
 295 {
 296 #ifndef HAVE_QUOTE_FOLD
 297    (void)self;
 298 #else
 299    void *p;
 300
 301    if ((p = self->qf_dat.s) != NULL)
 302       free(p);
 303    if ((p = self->qf_currq.s) != NULL)
 304       free(p);
 305 #endif
 306 }
 307
 308 void
 309 quoteflt_reset(struct quoteflt *self, FILE *f) /* XXX inline */
 310 {
 311    self->qf_os = f;
 312 #ifdef HAVE_QUOTE_FOLD
 313    self->qf_state = _QF_NONE;
 314    self->qf_dat.l =
 315    self->qf_currq.l = 0;
 316    memset(self->qf_mbps, 0, sizeof self->qf_mbps);
 317 #endif
 318 }
 319
 320 ssize_t
 321 quoteflt_push(struct quoteflt *self, char const *dat, size_t len)
 322 {
 323    /* (xxx Ideally the actual push() [and flush()] would be functions on their
 324     * xxx own, via indirect vtbl call ..) */
 325    ssize_t i, rv = 0;
 326
 327    if (len == 0)
 328       goto jleave;
 329
 330    /* Simple bypass? XXX Finally, this filter simply should not be used, then */
 331    if (self->qf_pfix_len == 0)
 332       rv = fwrite(dat, sizeof *dat, len, self->qf_os);
 333    /* The simple algorithm: place *indentprefix* at every BOL */
 334    else
 335 #ifdef HAVE_QUOTE_FOLD
 336       if (self->qf_qfold_max == 0)
 337 #endif
 338    {
 339       void *vp;
 340       size_t ll;
 341       bool_t pxok = (self->qf_qfold_min != 0);
 342
 343       for (;;) {
 344          if (! pxok) {
 345             i = fwrite(self->qf_pfix, sizeof *self->qf_pfix, self->qf_pfix_len,
 346                   self->qf_os);
 347             if (i < 0)
 348                goto jerr;
 349             rv += i;
 350             pxok = TRU1;
 351          }
 352
 353          /* xxx Strictly speaking this is invalid, because only `/' and `.' are
 354           * xxx mandated by POSIX.1-2008 as "invariant across all locales
 355           * xxx supported"; though there is no charset known which uses this
 356           * xxx control char as part of a multibyte character; note that S-nail
 357           * XXX (and the Mail codebase as such) do not support EBCDIC */
 358          if ((vp = memchr(dat, '\n', len)) == NULL)
 359             ll = len;
 360          else {
 361             pxok = FAL0;
 362             ll = (size_t)((char*)vp - dat) + 1;
 363          }
 364
 365          i = fwrite(dat, sizeof *dat, ll, self->qf_os);
 366          if (i < 0)
 367             goto jerr;
 368          rv += i;
 369          if ((len -= ll) == 0)
 370             break;
 371          dat += ll;
 372       }
 373
 374       self->qf_qfold_min = pxok;
 375    }
 376    /* More complicated, though still only line-per-line: *quote-fold*.
 377     * - If .qf_currq.l is 0, then we are in a clean state.  Reset .qf_mbps;
 378     *   note this means we assume that lines start with reset escape seq
 379     * - Lookout for a newline
 380     */
 381 #ifdef HAVE_QUOTE_FOLD
 382    else {
 383       struct qf_vc vc;
 384
 385       vc.self = self;
 386       vc.buf = dat;
 387       vc.len = len;
 388
 389       while (vc.len > 0) {
 390          switch (self->qf_state) {
 391          case _QF_NONE:
 392          case _QF_PREFIX:
 393             i = _qf_state_prefix(&vc);
 394             break;
 395          case _QF_DATA:
 396             i = _qf_state_data(&vc);
 397             break;
 398          }
 399
 400          if (i < 0)
 401             goto jerr;
 402          rv += i;
 403       }
 404    }
 405 #endif /* HAVE_QUOTE_FOLD */
 406
 407 jleave:
 408    return rv;
 409
 410 jerr:
 411    rv = -1;
 412    goto jleave;
 413 }
 414
 415 ssize_t
 416 quoteflt_flush(struct quoteflt *self)
 417 {
 418    ssize_t rv = 0;
 419    (void)self;
 420
 421 #ifdef HAVE_QUOTE_FOLD
 422    if (self->qf_dat.l > 0) {
 423       rv = fwrite(self->qf_pfix, sizeof *self->qf_pfix, self->qf_pfix_len,
 424             self->qf_os);
 425       if (rv > 0) {
 426          ssize_t j = fwrite(self->qf_dat.s, sizeof *self->qf_dat.s,
 427                self->qf_dat.l, self->qf_os);
 428          rv = (j < 0) ? j : rv + j;
 429       }
 430       self->qf_dat.l = 0;
 431    }
 432 #endif
 433    return rv;
 434 }
 435
 436 /* vim:set fenc=utf-8:s-it-mode */