diplodocus.org Git - nmh/blob - sbr/dtimep.l

   1 /* dtimep.l -- parse dates and times.
   2  *
   3  * This exceeds the default table capacities for some old versions
   4  * of lex (and the minimum defaults as specified by POSIX).  The following
   5  * choices meet or exceed the lex defaults for older SunOS4.x, Solaris,
   6  * HPUX, and AIX.
   7  */
   8 %e4000
   9 %p7000
  10 %n2500
  11 %a5000
  12 %{
  13 #include <h/nmh.h>
  14 #include <h/tws.h>
  15
  16   /* Since we're looking at a string at a time, don't worry about
  17    *  wrapping to the next buffer.
  18    */
  19 #if YY_FLEX_MAJOR_VERSION == 2 && \
  20     YY_FLEX_MINOR_VERSION == 6 && \
  21     YY_FLEX_SUBMINOR_VERSION == 3
  22 /* https://github.com/westes/flex/issues/162 */
  23 #undef yywrap
  24 #endif
  25 #define yywrap() 1
  26 #define YY_SKIP_YYWRAP
  27
  28 #define YY_NO_INPUT
  29
  30   /* This is the tricky thing that makes this function cool.  We
  31    *  replace the traditional int yylex(void) declaration with our
  32    *  dparsetime() declaration, essentially piggy-backing off the
  33    *  utility of the yylex() function and adding what we need to make
  34    *  the parsing function useful to us.
  35    */
  36 #define YY_DECL struct tws *dparsetime(char *lexstr)
  37
  38   /* yyerminate() is called after the input string is matched to
  39    * completion (actually, when the lexer reaches an EOF).  The only
  40    * thing that really needs to be in this macro function is the
  41    * return call, which must be substituted inline into dparsetime.
  42    */
  43
  44 #define yyterminate() (void)yy_delete_buffer(lexhandle); \
  45   if(!(tw.tw_flags & TW_SUCC)) { \
  46     return (struct tws *)NULL; \
  47   } \
  48   if(tw.tw_year < 1970) \
  49     tw.tw_year += 1900; \
  50   if(tw.tw_year < 1970) \
  51     tw.tw_year += 100; \
  52   return(&tw)
  53
  54 /*
  55  * Table to convert month names to numeric month.  We use the
  56  * fact that the low order 5 bits of the sum of the 2nd & 3rd
  57  * characters of the name is a hash with no collisions for the 12
  58  * valid month names.  (The mask to 5 bits maps any combination of
  59  * upper and lower case into the same hash value).
  60  */
  61 static int month_map[] = {
  62         0,
  63         6,      /* 1 - Jul */
  64         3,      /* 2 - Apr */
  65         5,      /* 3 - Jun */
  66         0,
  67         10,     /* 5 - Nov */
  68         0,
  69         1,      /* 7 - Feb */
  70         11,     /* 8 - Dec */
  71         0,
  72         0,
  73         0,
  74         0,
  75         0,
  76         0,
  77         0,      /*15 - Jan */
  78         0,
  79         0,
  80         0,
  81         2,      /*19 - Mar */
  82         0,
  83         8,      /*21 - Sep */
  84         0,
  85         9,      /*23 - Oct */
  86         0,
  87         0,
  88         4,      /*26 - May */
  89         0,
  90         7       /*28 - Aug */
  91 };
  92
  93 /*
  94  * Lookup table for day-of-week using the same hash trick as for above name-of-
  95  * month table, but using the first and second character, not second and third.
  96  *
  97  * Compute index into table using: (day_name[0] & 7) + (day_name[1] & 4)
  98  */
  99 static int day_map[] = {
 100         0,
 101         0,
 102         0,
 103         6,      /* 3 - Sat */
 104         4,      /* 4 - Thu */
 105         0,
 106         5,      /* 6 - Fri */
 107         0,      /* 7 - Sun */
 108         2,      /* 8 - Tue */
 109         1       /* 9 - Mon */,
 110         0,
 111         3       /*11 - Wed */
 112 };
 113
 114 /* The SET* macros will parse for the appropriate field, and leave the
 115  * cp pointer at the first character after the desired field. Be
 116  * careful with variable-length fields or alpha-num mixes.
 117
 118  * The SKIP* macros skip over characters of a particular class and
 119  * leave cp at the position of the first character that doesn't match
 120  * that class. Correspondingly, SKIPTO* skips until it reaches a
 121  * character of a particular class.
 122  */
 123
 124 #define INIT()       { cp = yytext;}
 125 #define SETWDAY()    { tw.tw_wday= day_map[(cp[0] & 7) + (cp[1] & 4)]; \
 126                        tw.tw_flags &= ~TW_SDAY; tw.tw_flags |= TW_SEXP; \
 127                        SKIPA(); }
 128 #define SETMON()     { cp++; \
 129                        tw.tw_mon = month_map[(((unsigned char) cp[0]) + ((unsigned char) cp[1])) & 0x1f]; \
 130                        SKIPA(); }
 131 #define SETMON_NUM() { tw.tw_mon = atoi(cp)-1; \
 132                        SKIPD(); }
 133 #define SETYEAR()    { tw.tw_year = atoi(cp); \
 134                        SKIPD(); }
 135 #define SETDAY()     { tw.tw_mday = atoi(cp); \
 136                        tw.tw_flags |= TW_SUCC; \
 137                        SKIPD(); }
 138 #define SETTIME()    { tw.tw_hour = atoi(cp); \
 139                        cp += 2; \
 140                        SKIPTOD(); \
 141                        tw.tw_min = atoi(cp); \
 142                        cp += 2; \
 143                        if(*cp == ':') { \
 144                           tw.tw_sec = atoi(++cp); SKIPD(); } }
 145 #define SETZONE(x)   { tw.tw_zone = ((x)/100)*60+(x)%100; \
 146                        tw.tw_flags |= TW_SZEXP; \
 147                        SKIPD(); }
 148 #define SETDST()     { tw.tw_flags |= TW_DST; }
 149 #define SKIPD()      { while ( isdigit((unsigned char) *cp++) ) ; \
 150                        --cp; }
 151 #define SKIPTOD()    { while ( !isdigit((unsigned char) *cp++) ) ; \
 152                        --cp; }
 153 #define SKIPA()      { while ( isalpha((unsigned char) *cp++) ) ; \
 154                        --cp; }
 155 #define SKIPTOA()    { while ( !isalpha((unsigned char) *cp++) ) ; \
 156                        --cp; }
 157 #define SKIPSP()     { while ( isspace((unsigned char) *cp++) ) ; \
 158                        --cp; }
 159 #define SKIPTOSP()   { while ( !isspace((unsigned char) *cp++) ) ; \
 160                        --cp; }
 161
 162 #ifdef ADJUST_NUMERIC_ONLY_TZ_OFFSETS_WRT_DST
 163 # ifdef HAVE_SYS_TIME_H
 164 #  include <sys/time.h>
 165 # endif
 166 #include <time.h>
 167
 168 static void
 169 zonehack (struct tws *tw)
 170 {
 171     struct tm *tm;
 172
 173     if (dmktime (tw) == (time_t) -1)
 174         return;
 175
 176     tm = localtime (&tw->tw_clock);
 177     if (tm->tm_isdst) {
 178         tw->tw_flags |= TW_DST;
 179         tw->tw_zone -= 60;
 180     }
 181 }
 182 #endif  /* ADJUST_NUMERIC_ONLY_TZ_OFFSETS_WRT_DST */
 183 %}
 184
 185 sun     ([Ss]un(day)?)
 186 mon     ([Mm]on(day)?)
 187 tue     ([Tt]ue(sday)?)
 188 wed     ([Ww]ed(nesday)?)
 189 thu     ([Tt]hu(rsday)?)
 190 fri     ([Ff]ri(day)?)
 191 sat     ([Ss]at(urday)?)
 192
 193 DAY     ({sun}|{mon}|{tue}|{wed}|{thu}|{fri}|{sat})
 194
 195 jan     ([Jj]an(uary)?)
 196 feb     ([Ff]eb(ruary)?)
 197 mar     ([Mm]ar(ch)?)
 198 apr     ([Aa]pr(il)?)
 199 may     ([Mm]ay)
 200 jun     ([Jj]un(e)?)
 201 jul     ([Jj]ul(y)?)
 202 aug     ([Aa]ug(ust)?)
 203 sep     ([Ss]ep(tember)?)
 204 oct     ([Oo]ct(ober)?)
 205 nov     ([Nn]ov(ember)?)
 206 dec     ([Dd]ec(ember)?)
 207
 208 MONTH   ({jan}|{feb}|{mar}|{apr}|{may}|{jun}|{jul}|{aug}|{sep}|{oct}|{nov}|{dec})
 209
 210 TIME    ({D}:{d}{d}(:{d}{d})?)
 211
 212      /* The year can either be 2 digits, or 4. However, after
 213         Y2K, we found that some MUA were reporting the year 100, hence
 214         the middle term here. yyterminate() resolves the actual
 215         issues with 2-digit years.
 216      */
 217
 218 YEAR    (({d}{d})|(1{d}{d})|({d}{4}))
 219
 220 w       ([ \t]*)
 221 W       ([ \t]+)
 222 D       ([0-9]?[0-9])
 223 d       [0-9]
 224 nl      [ \t\n()]
 225
 226 %%
 227 %{
 228   /* This section begins the definition of dparsetime().
 229      Put here any local variable definitions and initializations */
 230
 231   YY_BUFFER_STATE lexhandle;
 232
 233   char *cp;
 234   static struct tws tw;
 235
 236   memset(&tw,0,sizeof(struct tws));
 237
 238   lexhandle = yy_scan_string(lexstr);
 239 %}
 240
 241 {DAY}","?{W}{MONTH}{W}{D}{W}{TIME}{W}{YEAR}   {
 242                                      INIT();
 243                                      SETWDAY();
 244                                      SKIPTOA();
 245                                      SETMON();
 246                                      SKIPTOD();
 247                                      SETDAY();
 248                                      SKIPTOD();
 249                                      SETTIME();
 250                                      SKIPTOD();
 251                                      SETYEAR();
 252                                      }
 253
 254 {DAY}","?{W}{D}{W}{MONTH}{W}{YEAR}{W}{TIME}   {
 255                                      INIT();
 256                                      SETWDAY();
 257                                      SKIPTOD();
 258                                      SETDAY();
 259                                      SKIPTOA();
 260                                      SETMON();
 261                                      SKIPTOD();
 262                                      SETYEAR();
 263                                      SKIPTOD();
 264                                      SETTIME();
 265                                      }
 266 {D}{W}{MONTH}{W}{YEAR}{W}{TIME}               {
 267                                      INIT();
 268                                      SETDAY();
 269                                      SKIPTOA();
 270                                      SETMON();
 271                                      SKIPTOD();
 272                                      SETYEAR();
 273                                      SKIPTOD();
 274                                      SETTIME();
 275                                      }
 276 {DAY}","?{W}{MONTH}{W}{D}","?{W}{YEAR}","?{W}{TIME} {
 277                                      INIT();
 278                                      SETWDAY();
 279                                      SKIPTOA();
 280                                      SETMON();
 281                                      SKIPTOD();
 282                                      SETDAY();
 283                                      SKIPTOD();
 284                                      SETYEAR();
 285                                      SKIPTOD();
 286                                      SETTIME();
 287                                      }
 288 {DAY}","?{W}{MONTH}{W}{D}","?{W}{YEAR}        {
 289                                      INIT();
 290                                      SETWDAY();
 291                                      SKIPTOA();
 292                                      SETMON();
 293                                      SKIPTOD();
 294                                      SETDAY();
 295                                      SKIPTOD();
 296                                      SETYEAR();
 297                                      }
 298 {MONTH}{W}{D}","?{W}{YEAR}","?{W}{DAY}        {
 299                                      INIT();
 300                                      SETMON();
 301                                      SKIPTOD();
 302                                      SETDAY();
 303                                      SKIPTOD();
 304                                      SETYEAR();
 305                                      SKIPTOA();
 306                                      SETWDAY();
 307                                      }
 308 {MONTH}{W}{D}","?{W}{YEAR}                    {
 309                                      INIT();
 310                                      SETMON();
 311                                      SKIPTOD();
 312                                      SETDAY();
 313                                      SKIPTOD();
 314                                      SETYEAR();
 315                                      }
 316 {D}("-"|"/"){D}("-"|"/"){YEAR}{W}{TIME}       {
 317                                      INIT();
 318                                      /* MM/DD/YY */
 319                                      SETMON_NUM();
 320                                      SKIPTOD();
 321                                      SETDAY();
 322                                      SKIPTOD();
 323                                      SETYEAR();
 324                                      SKIPTOD();
 325                                      SETTIME();
 326                                      }
 327 {D}("-"|"/"){D}("-"|"/"){YEAR}                {
 328                                      INIT();
 329                                      /* MM/DD/YY */
 330                                      SETMON_NUM();
 331                                      SKIPTOD();
 332                                      SETDAY();
 333                                      SKIPTOD();
 334                                      SETYEAR();
 335                                      }
 336
 337 "[Aa][Mm]"
 338 "[Pp][Mm]"                           tw.tw_hour += 12;
 339
 340 "+"{D}{d}{d}                                  {
 341                                     INIT();
 342                                     SKIPTOD();
 343                                     SETZONE(atoi(cp));
 344 #ifdef  ADJUST_NUMERIC_ONLY_TZ_OFFSETS_WRT_DST
 345                                     zonehack (&tw);
 346 #endif  /* ADJUST_NUMERIC_ONLY_TZ_OFFSETS_WRT_DST */
 347                                     yyterminate();
 348                                     }
 349 "-"{D}{d}{d}                                  {
 350                                     INIT();
 351                                     SKIPTOD();
 352                                     SETZONE(-atoi(cp));
 353 #ifdef  ADJUST_NUMERIC_ONLY_TZ_OFFSETS_WRT_DST
 354                                     zonehack (&tw);
 355 #endif  /* ADJUST_NUMERIC_ONLY_TZ_OFFSETS_WRT_DST */
 356                                     yyterminate();
 357
 358                                     }
 359 {nl}("ut"|"UT")                     INIT(); SETZONE(0); yyterminate();
 360 {nl}("gmt"|"GMT")                   INIT(); SETZONE(0); yyterminate();
 361 {nl}("est"|"EST")                   INIT(); SETZONE(-500); yyterminate();
 362 {nl}("edt"|"EDT")                   { INIT(); SETDST(); SETZONE(-500);
 363                                       yyterminate(); }
 364 {nl}("cst"|"CST")                   INIT(); SETZONE(-600); yyterminate();
 365 {nl}("cdt"|"CDT")                   { INIT(); SETDST(); SETZONE(-600);
 366                                       yyterminate(); }
 367 {nl}("mst"|"MST")                   INIT(); SETZONE(-700); yyterminate();
 368 {nl}("mdt"|"MDT")                   { INIT(); SETDST(); SETZONE(-700);
 369                                       yyterminate(); }
 370 {nl}("pst"|"PST")                   INIT(); SETZONE(-800); yyterminate();
 371 {nl}("pdt"|"PDT")                   { INIT(); SETDST(); SETZONE(-800);
 372                                       yyterminate(); }
 373 {nl}("nst"|"NST")                   INIT(); SETZONE(-330); yyterminate();
 374 {nl}("ast"|"AST")                   INIT(); SETZONE(-400); yyterminate();
 375 {nl}("adt"|"ADT")                   { INIT(); SETDST(); SETZONE(-400);
 376                                       yyterminate(); }
 377 {nl}("hst"|"HST")                   INIT(); SETZONE(-1000); yyterminate();
 378 {nl}("hdt"|"HDT")                   { INIT(); SETDST(); SETZONE(-1000);
 379                                       yyterminate(); }
 380 .|\n
 381
 382 %%
 383 /* This is a portable way to squash a warning about the yyunput()
 384  * function being static but never used. It costs us a tiny amount
 385  * of extra code in the binary but the other options are:
 386  *  "%option nounput" which is flex-specific
 387  *  makefile hackery just to compile dtimep.c with different flags
 388  */
 389 void dtimep_yyunput(int c)
 390 {
 391     unput(c);
 392 }