modules/format.c

   1 /* GNU m4 -- A simple macro processor
   2    Copyright (C) 1989, 1990, 1991, 1992, 1993, 1994, 2001, 2006, 2007,
   3    2008, 2009, 2010 Free Software Foundation, Inc.
   4
   5    This file is part of GNU M4.
   6
   7    GNU M4 is free software: you can redistribute it and/or modify
   8    it under the terms of the GNU General Public License as published by
   9    the Free Software Foundation, either version 3 of the License, or
  10    (at your option) any later version.
  11
  12    GNU M4 is distributed in the hope that it will be useful,
  13    but WITHOUT ANY WARRANTY; without even the implied warranty of
  14    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15    GNU General Public License for more details.
  16
  17    You should have received a copy of the GNU General Public License
  18    along with this program.  If not, see <http://www.gnu.org/licenses/>.
  19 */
  20
  21 /* printf like formatting for m4.  */
  22
  23 #include "vasnprintf.h"
  24
  25 /* Simple varargs substitute.  We assume int and unsigned int are the
  26    same size; likewise for long and unsigned long.  We do not yet
  27    handle long double or long long.  */
  28
  29 /* Parse STR of length LEN as an integer, reporting warnings on behalf
  30    of ME.  */
  31 static int
  32 arg_int (struct m4 *context, const m4_call_info *me, const char *str,
  33          size_t len)
  34 {
  35   char *endp;
  36   long value;
  37
  38   /* TODO - also allow parsing `'a' or `"a' which results in the
  39      numeric value of 'a', as in printf(1).  */
  40   if (!len)
  41     {
  42       m4_warn (context, 0, me, _("empty string treated as 0"));
  43       return 0;
  44     }
  45   errno = 0;
  46   value = strtol (str, &endp, 10);
  47   if (endp - str != len)
  48     m4_warn (context, 0, me, _("non-numeric argument %s"),
  49              quotearg_style_mem (locale_quoting_style, str, len));
  50   else if (isspace (to_uchar (*str)))
  51     m4_warn (context, 0, me, _("leading whitespace ignored"));
  52   else if (errno == ERANGE || (int) value != value)
  53     m4_warn (context, 0, me, _("numeric overflow detected"));
  54   return value;
  55 }
  56
  57 /* Parse STR of length LEN as a long, reporting warnings on behalf of
  58    ME.  */
  59 static long
  60 arg_long (struct m4 *context, const m4_call_info *me, const char *str,
  61           size_t len)
  62 {
  63   char *endp;
  64   long value;
  65
  66   /* TODO - also allow parsing `'a' or `"a' which results in the
  67      numeric value of 'a', as in printf(1).  */
  68   if (!len)
  69     {
  70       m4_warn (context, 0, me, _("empty string treated as 0"));
  71       return 0L;
  72     }
  73   errno = 0;
  74   value = strtol (str, &endp, 10);
  75   if (endp - str != len)
  76     m4_warn (context, 0, me, _("non-numeric argument %s"),
  77              quotearg_style_mem (locale_quoting_style, str, len));
  78   else if (isspace (to_uchar (*str)))
  79     m4_warn (context, 0, me, _("leading whitespace ignored"));
  80   else if (errno == ERANGE)
  81     m4_warn (context, 0, me, _("numeric overflow detected"));
  82   return value;
  83 }
  84
  85 /* Check STR of length LEN for embedded NUL, reporting warnings on
  86    behalf of ME.  */
  87 static const char *
  88 arg_string (struct m4 *context, const m4_call_info *me, const char *str,
  89             size_t len)
  90 {
  91   if (strlen (str) < len)
  92     m4_warn (context, 0, me, _("argument %s truncated"),
  93              quotearg_style_mem (locale_quoting_style, str, len));
  94   return str;
  95 }
  96
  97 /* Parse STR of length LEN as a double, reporting warnings on behalf
  98    of ME.  */
  99 static double
 100 arg_double (struct m4 *context, const m4_call_info *me, const char *str,
 101             size_t len)
 102 {
 103   char *endp;
 104   double value;
 105
 106   if (!len)
 107     {
 108       m4_warn (context, 0, me, _("empty string treated as 0"));
 109       return 0.0;
 110     }
 111   errno = 0;
 112   value = strtod (str, &endp);
 113   if (endp - str != len)
 114     m4_warn (context, 0, me, _("non-numeric argument %s"),
 115              quotearg_style_mem (locale_quoting_style, str, len));
 116   else if (isspace (to_uchar (*str)))
 117     m4_warn (context, 0, me, _("leading whitespace ignored"));
 118   else if (errno == ERANGE)
 119     m4_warn (context, 0, me, _("numeric overflow detected"));
 120   return value;
 121 }
 122
 123 #define ARG_INT(i, argc, argv)                                  \
 124   ((argc <= ++i) ? 0 : arg_int (context, me, M4ARG (i), M4ARGLEN (i)))
 125
 126 #define ARG_LONG(i, argc, argv)                                 \
 127   ((argc <= ++i) ? 0L : arg_long (context, me, M4ARG (i), M4ARGLEN (i)))
 128
 129 #define ARG_STR(i, argc, argv)                                  \
 130   ((argc <= ++i) ? "" : arg_string (context, me, M4ARG (i), M4ARGLEN (i)))
 131
 132 #define ARG_DOUBLE(i, argc, argv)                               \
 133   ((argc <= ++i) ? 0.0 : arg_double (context, me, M4ARG (i), M4ARGLEN (i)))
 134 \f
 135
 136 /* The main formatting function.  Output is placed on the obstack OBS,
 137    the first argument in ARGV is the formatting string, and the rest
 138    is arguments for the string.  Warn rather than invoke unspecified
 139    behavior in the underlying printf when we do not recognize a
 140    format.  */
 141
 142 static void
 143 format (m4 *context, m4_obstack *obs, int argc, m4_macro_args *argv)
 144 {
 145   const m4_call_info *me = m4_arg_info (argv);
 146   const char *f;                        /* Format control string.  */
 147   size_t f_len;                         /* Length of f.  */
 148   const char *fmt;                      /* Position within f.  */
 149   char fstart[] = "%'+- 0#*.*hhd";      /* Current format spec.  */
 150   char *p;                              /* Position within fstart.  */
 151   unsigned char c;                      /* A simple character.  */
 152   int i = 1;                            /* Index within argc used so far.  */
 153   bool valid_format = true;             /* True if entire format string ok.  */
 154
 155   /* Flags.  */
 156   char flags;                           /* Flags to use in fstart.  */
 157   enum {
 158     THOUSANDS   = 0x01, /* '\''.  */
 159     PLUS        = 0x02, /* '+'.  */
 160     MINUS       = 0x04, /* '-'.  */
 161     SPACE       = 0x08, /* ' '.  */
 162     ZERO        = 0x10, /* '0'.  */
 163     ALT         = 0x20, /* '#'.  */
 164     DONE        = 0x40  /* No more flags.  */
 165   };
 166
 167   /* Precision specifiers.  */
 168   int width;                    /* Minimum field width.  */
 169   int prec;                     /* Precision.  */
 170   char lflag;                   /* Long flag.  */
 171
 172   /* Specifiers we are willing to accept.  ok['x'] implies %x is ok.
 173      Various modifiers reduce the set, in order to avoid undefined
 174      behavior in printf.  */
 175   char ok[128];
 176
 177   /* Check that formatted text succeeded with correct type.  */
 178   int result = 0;
 179   enum {CHAR, INT, LONG, DOUBLE, STR} datatype;
 180
 181   f = fmt = M4ARG (1);
 182   f_len = M4ARGLEN (1);
 183   assert (!f[f_len]); /* Requiring a terminating NUL makes parsing simpler.  */
 184   memset (ok, 0, sizeof ok);
 185   while (1)
 186     {
 187       const char *percent = (char *) memchr (fmt, '%', f_len);
 188       if (!percent)
 189         {
 190           obstack_grow (obs, fmt, f_len);
 191           break;
 192         }
 193       obstack_grow (obs, fmt, percent - fmt);
 194       f_len -= percent - fmt + 1;
 195       fmt = percent + 1;
 196
 197       if (*fmt == '%')
 198         {
 199           obstack_1grow (obs, '%');
 200           fmt++;
 201           f_len--;
 202           continue;
 203         }
 204
 205       p = fstart + 1; /* % */
 206       lflag = 0;
 207       ok['a'] = ok['A'] = ok['c'] = ok['d'] = ok['e'] = ok['E']
 208         = ok['f'] = ok['F'] = ok['g'] = ok['G'] = ok['i'] = ok['o']
 209         = ok['s'] = ok['u'] = ok['x'] = ok['X'] = 1;
 210
 211       /* Parse flags.  */
 212       flags = 0;
 213       do
 214         {
 215           switch (*fmt)
 216             {
 217             case '\'': /* thousands separator */
 218               ok['a'] = ok['A'] = ok['c'] = ok['e'] = ok['E']
 219                 = ok['o'] = ok['s'] = ok['x'] = ok['X'] = 0;
 220               flags |= THOUSANDS;
 221               break;
 222
 223             case '+': /* mandatory sign */
 224               ok['c'] = ok['o'] = ok['s'] = ok['u'] = ok['x'] = ok['X'] = 0;
 225               flags |= PLUS;
 226               break;
 227
 228             case ' ': /* space instead of positive sign */
 229               ok['c'] = ok['o'] = ok['s'] = ok['u'] = ok['x'] = ok['X'] = 0;
 230               flags |= SPACE;
 231               break;
 232
 233             case '0': /* zero padding */
 234               ok['c'] = ok['s'] = 0;
 235               flags |= ZERO;
 236               break;
 237
 238             case '#': /* alternate output */
 239               ok['c'] = ok['d'] = ok['i'] = ok['s'] = ok['u'] = 0;
 240               flags |= ALT;
 241               break;
 242
 243             case '-': /* left justification */
 244               flags |= MINUS;
 245               break;
 246
 247             default:
 248               flags |= DONE;
 249               break;
 250             }
 251         }
 252       while (!(flags & DONE) && (f_len--, fmt++));
 253       if (flags & THOUSANDS)
 254         *p++ = '\'';
 255       if (flags & PLUS)
 256         *p++ = '+';
 257       if (flags & MINUS)
 258         *p++ = '-';
 259       if (flags & SPACE)
 260         *p++ = ' ';
 261       if (flags & ZERO)
 262         *p++ = '0';
 263       if (flags & ALT)
 264         *p++ = '#';
 265
 266       /* Minimum field width; an explicit 0 is the same as not giving
 267          the width.  */
 268       width = 0;
 269       *p++ = '*';
 270       if (*fmt == '*')
 271         {
 272           width = ARG_INT (i, argc, argv);
 273           fmt++;
 274           f_len--;
 275         }
 276       else
 277         while (isdigit ((unsigned char) *fmt))
 278           {
 279             width = 10 * width + *fmt - '0';
 280             fmt++;
 281             f_len--;
 282           }
 283
 284       /* Maximum precision; an explicit negative precision is the same
 285          as not giving the precision.  A lone '.' is a precision of 0.  */
 286       prec = -1;
 287       *p++ = '.';
 288       *p++ = '*';
 289       if (*fmt == '.')
 290         {
 291           ok['c'] = 0;
 292           f_len--;
 293           if (*(++fmt) == '*')
 294             {
 295               prec = ARG_INT (i, argc, argv);
 296               ++fmt;
 297               f_len--;
 298             }
 299           else
 300             {
 301               prec = 0;
 302               while (isdigit ((unsigned char) *fmt))
 303                 {
 304                   prec = 10 * prec + *fmt - '0';
 305                   fmt++;
 306                   f_len--;
 307                 }
 308             }
 309         }
 310
 311       /* Length modifiers.  We don't yet recognize ll, j, t, or z.  */
 312       if (*fmt == 'l')
 313         {
 314           *p++ = 'l';
 315           lflag = 1;
 316           fmt++;
 317           f_len--;
 318           ok['c'] = ok['s'] = 0;
 319         }
 320       else if (*fmt == 'h')
 321         {
 322           *p++ = 'h';
 323           fmt++;
 324           f_len--;
 325           if (*fmt == 'h')
 326             {
 327               *p++ = 'h';
 328               fmt++;
 329               f_len--;
 330             }
 331           ok['a'] = ok['A'] = ok['c'] = ok['e'] = ok['E'] = ok['f'] = ok['F']
 332             = ok['g'] = ok['G'] = ok['s'] = 0;
 333         }
 334
 335       c = *fmt;
 336       if (c > sizeof ok || !ok[c] || !f_len)
 337         {
 338           m4_warn (context, 0, me, _("unrecognized specifier in %s"),
 339                    quotearg_style_mem (locale_quoting_style, f, M4ARGLEN (1)));
 340           valid_format = false;
 341           continue;
 342         }
 343       fmt++;
 344       f_len--;
 345
 346       /* Specifiers.  We don't yet recognize C, S, n, or p.  */
 347       switch (c)
 348         {
 349         case 'c':
 350           datatype = CHAR;
 351           p -= 2; /* %.*c is undefined, so undo the '.*'.  */
 352           break;
 353
 354         case 's':
 355           datatype = STR;
 356           break;
 357
 358         case 'd':
 359         case 'i':
 360         case 'o':
 361         case 'x':
 362         case 'X':
 363         case 'u':
 364           datatype = lflag ? LONG : INT;
 365           break;
 366
 367         case 'a':
 368         case 'A':
 369         case 'e':
 370         case 'E':
 371         case 'f':
 372         case 'F':
 373         case 'g':
 374         case 'G':
 375           datatype = DOUBLE;
 376           break;
 377
 378         default:
 379           abort ();
 380         }
 381       *p++ = c;
 382       *p = '\0';
 383
 384       switch (datatype)
 385         {
 386         case CHAR:
 387           result = obstack_printf (obs, fstart, width,
 388                                    ARG_INT (i, argc, argv));
 389           break;
 390
 391         case INT:
 392           result = obstack_printf (obs, fstart, width, prec,
 393                                    ARG_INT (i, argc, argv));
 394           break;
 395
 396         case LONG:
 397           result = obstack_printf (obs, fstart, width, prec,
 398                                    ARG_LONG (i, argc, argv));
 399           break;
 400
 401         case DOUBLE:
 402           result = obstack_printf (obs, fstart, width, prec,
 403                                    ARG_DOUBLE (i, argc, argv));
 404           break;
 405
 406         case STR:
 407           result = obstack_printf (obs, fstart, width, prec,
 408                                    ARG_STR (i, argc, argv));
 409           break;
 410
 411         default:
 412           abort ();
 413         }
 414       /* Since obstack_printf can only fail with EILSEQ or EINVAL, but
 415          we constructed fstart, the result should not be negative.  */
 416       assert (0 <= result);
 417     }
 418   if (valid_format)
 419     m4_bad_argc (context, argc, me, i, i, true);
 420 }