libs/rfc822/imapsubj.c

   1 /*
   2 ** Copyright 2000 Double Precision, Inc.
   3 ** See COPYING for distribution information.
   4 */
   5
   6 /*
   7 */
   8 #include        "config.h"
   9 #include        <stdio.h>
  10 #include        <ctype.h>
  11 #include        <stdlib.h>
  12 #include        <string.h>
  13 #include        "rfc822.h"
  14
  15 #if     HAVE_STRCASECMP
  16
  17 #else
  18 #define strcasecmp      stricmp
  19 #endif
  20
  21 #if     HAVE_STRNCASECMP
  22
  23 #else
  24 #define strncasecmp     strnicmp
  25 #endif
  26
  27 /* Skip over blobs */
  28
  29 static char *skipblob(char *p, char **save_blob_ptr)
  30 {
  31         char *q;
  32         char *orig_p=p;
  33         int isalldigits=1;
  34
  35         if (*p == '[')
  36         {
  37                 for (q= p+1; *q; q++)
  38                         if (*q == '[' || *q == ']')
  39                                 break;
  40                         else if (strchr("0123456789", *q) == NULL)
  41                                 isalldigits=0;
  42
  43                 if (*q == ']')
  44                 {
  45                         p=q+1;
  46
  47                         while (isspace((int)(unsigned char)*p))
  48                         {
  49                                 ++p;
  50                         }
  51
  52                         if (save_blob_ptr && *save_blob_ptr && !isalldigits)
  53                         {
  54                                 while (orig_p != p)
  55                                         *(*save_blob_ptr)++=*orig_p++;
  56                         }
  57
  58                         return (p);
  59                 }
  60         }
  61         return (p);
  62 }
  63
  64 static char *skipblobs(char *p, char **save_blob_ptr)
  65 {
  66         char *q=p;
  67
  68         do
  69         {
  70                 p=q;
  71                 q=skipblob(p, save_blob_ptr);
  72         } while (q != p);
  73         return (q);
  74 }
  75
  76 /* Remove artifacts from the subject header */
  77
  78 static void stripsubj(char *s, int *hasrefwd, char *save_blob_buf)
  79 {
  80         char    *p;
  81         char    *q;
  82         int doit;
  83
  84         for (p=q=s; *p; p++)
  85         {
  86                 if (!isspace((int)(unsigned char)*p))
  87                 {
  88                         *q++=*p;
  89                         continue;
  90                 }
  91                 while (p[1] && isspace((int)(unsigned char)p[1]))
  92                 {
  93                         ++p;
  94                 }
  95                 *q++=' ';
  96         }
  97         *q=0;
  98
  99         do
 100         {
 101                 doit=0;
 102                 /*
 103                 **
 104                 ** (2) Remove all trailing text of the subject that matches
 105                 ** the subj-trailer ABNF, repeat until no more matches are
 106                 ** possible.
 107                 **
 108                 **  subj-trailer    = "(fwd)" / WSP
 109                 */
 110
 111                 for (p=s; *p; p++)
 112                         ;
 113                 while (p > s)
 114                 {
 115                         if ( isspace((int)(unsigned char)p[-1]))
 116                         {
 117                                 --p;
 118                                 continue;
 119                         }
 120                         if (p-s >= 5 && strncasecmp(p-5, "(FWD)", 5) == 0)
 121                         {
 122                                 p -= 5;
 123                                 *hasrefwd |= CORESUBJ_FWD;
 124                                 continue;
 125                         }
 126                         break;
 127                 }
 128                 *p=0;
 129
 130                 for (p=s; *p; )
 131                 {
 132                         for (;;)
 133                         {
 134                                 char *orig_blob_ptr;
 135                                 int flag=CORESUBJ_FWD;
 136
 137                                 /*
 138                                 **
 139                                 ** (3) Remove all prefix text of the subject
 140                                 ** that matches the subj-leader ABNF.
 141                                 **
 142                                 **   subj-leader     = (*subj-blob subj-refwd) / WSP
 143                                 **
 144                                 **   subj-blob       = "[" *BLOBCHAR "]" *WSP
 145                                 **
 146                                 **   subj-refwd      = ("re" / ("fw" ["d"])) *WSP [subj-blob] ":"
 147                                 **
 148                                 **   BLOBCHAR        = %x01-5a / %x5c / %x5e-7f
 149                                 **                   ; any CHAR except '[' and ']'
 150                                 */
 151
 152                                 if (isspace((int)(unsigned char)*p))
 153                                 {
 154                                         ++p;
 155                                         continue;
 156                                 }
 157
 158                                 q=skipblobs(p, NULL);
 159
 160                                 if (strncasecmp(q, "RE", 2) == 0)
 161                                 {
 162                                         flag=CORESUBJ_RE;
 163                                         q += 2;
 164                                 }
 165                                 else if (strncasecmp(q, "FWD", 3) == 0)
 166                                 {
 167                                         q += 3;
 168                                 }
 169                                 else if (strncasecmp(q, "FW", 2) == 0)
 170                                 {
 171                                         q += 2;
 172                                 }
 173                                 else q=0;
 174
 175                                 if (q)
 176                                 {
 177                                         orig_blob_ptr=save_blob_buf;
 178
 179                                         q=skipblob(q, &save_blob_buf);
 180                                         if (*q == ':')
 181                                         {
 182                                                 p=q+1;
 183                                                 *hasrefwd |= flag;
 184                                                 continue;
 185                                         }
 186
 187                                         save_blob_buf=orig_blob_ptr;
 188                                 }
 189
 190
 191                                 /*
 192                                 ** (4) If there is prefix text of the subject
 193                                 ** that matches the subj-blob ABNF, and
 194                                 ** removing that prefix leaves a non-empty
 195                                 ** subj-base, then remove the prefix text.
 196                                 **
 197                                 **   subj-base       = NONWSP *([*WSP] NONWSP)
 198                                 **                   ; can be a subj-blob
 199                                 */
 200
 201                                 orig_blob_ptr=save_blob_buf;
 202
 203                                 q=skipblob(p, &save_blob_buf);
 204
 205                                 if (q != p && *q)
 206                                 {
 207                                         p=q;
 208                                         continue;
 209                                 }
 210                                 save_blob_buf=orig_blob_ptr;
 211                                 break;
 212                         }
 213
 214                         /*
 215                         **
 216                         ** (6) If the resulting text begins with the
 217                         ** subj-fwd-hdr ABNF and ends with the subj-fwd-trl
 218                         ** ABNF, remove the subj-fwd-hdr and subj-fwd-trl and
 219                         ** repeat from step (2).
 220                         **
 221                         **   subj-fwd-hdr    = "[fwd:"
 222                         **
 223                         **   subj-fwd-trl    = "]"
 224                         */
 225
 226                         if (strncasecmp(p, "[FWD:", 5) == 0)
 227                         {
 228                                 q=strrchr(p, ']');
 229                                 if (q && q[1] == 0)
 230                                 {
 231                                         *q=0;
 232                                         p += 5;
 233                                         *hasrefwd |= CORESUBJ_FWD;
 234
 235                                         for (q=s; (*q++=*p++) != 0; )
 236                                                 ;
 237                                         doit=1;
 238                                 }
 239                         }
 240                         break;
 241                 }
 242         } while (doit);
 243
 244         q=s;
 245         while ( (*q++ = *p++) != 0)
 246                 ;
 247         if (save_blob_buf)
 248                 *save_blob_buf=0;
 249 }
 250
 251 char *rfc822_coresubj(const char *s, int *hasrefwd)
 252 {
 253         char *q=strdup(s), *r;
 254         int dummy;
 255
 256         if (!hasrefwd)
 257                 hasrefwd= &dummy;
 258
 259         *hasrefwd=0;
 260         if (!q) return (0);
 261
 262         for (r=q; *r; r++)
 263                 if ((*r & 0x80) == 0)   /* Just US-ASCII casing, thanks */
 264                 {
 265                         if (*r >= 'a' && *r <= 'z')
 266                                 *r += 'A'-'a';
 267                 }
 268         stripsubj(q, hasrefwd, 0);
 269         return (q);
 270 }
 271
 272 char *rfc822_coresubj_nouc(const char *s, int *hasrefwd)
 273 {
 274         char *q=strdup(s);
 275         int dummy;
 276
 277         if (!hasrefwd)
 278                 hasrefwd= &dummy;
 279
 280         *hasrefwd=0;
 281         if (!q) return (0);
 282
 283         stripsubj(q, hasrefwd, 0);
 284         return (q);
 285 }
 286
 287 char *rfc822_coresubj_keepblobs(const char *s)
 288 {
 289         char *q=strdup(s), *r;
 290         int dummy;
 291
 292         if (!q) return (0);
 293
 294         r=strdup(s);
 295         if (!r)
 296         {
 297                 free(q);
 298                 return (0);
 299         }
 300
 301         stripsubj(q, &dummy, r);
 302         strcat(r, q);
 303         free(q);
 304         return (r);
 305 }