1 /* Formatted output to strings.
2    Copyright (C) 1999-2000, 2002-2003 Free Software Foundation, Inc.
3 
4    This program is free software; you can redistribute it and/or modify
5    it under the terms of the GNU General Public License as published by
6    the Free Software Foundation; either version 2, or (at your option)
7    any later version.
8 
9    This program is distributed in the hope that it will be useful,
10    but WITHOUT ANY WARRANTY; without even the implied warranty of
11    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
12    GNU General Public License for more details.
13 
14    You should have received a copy of the GNU General Public License along
15    with this program; if not, write to the Free Software Foundation,
16    Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.  */
17 
18 #ifdef HAVE_CONFIG_H
19 # include <config.h>
20 #endif
21 
22 /* Specification.  */
23 #if WIDE_CHAR_VERSION
24 # include "wprintf-parse.h"
25 #else
26 # include "printf-parse.h"
27 #endif
28 
29 /* Get size_t, NULL.  */
30 #include <stddef.h>
31 
32 /* Get intmax_t.  */
33 #if HAVE_STDINT_H_WITH_UINTMAX
34 # include <stdint.h>
35 #endif
36 #if HAVE_INTTYPES_H_WITH_UINTMAX
37 # include <inttypes.h>
38 #endif
39 
40 /* malloc(), realloc(), free().  */
41 #include <stdlib.h>
42 
43 /* Checked size_t computations.  */
44 #include "xsize.h"
45 
46 #if WIDE_CHAR_VERSION
47 # define PRINTF_PARSE wprintf_parse
48 # define CHAR_T wchar_t
49 # define DIRECTIVE wchar_t_directive
50 # define DIRECTIVES wchar_t_directives
51 #else
52 # define PRINTF_PARSE printf_parse
53 # define CHAR_T char
54 # define DIRECTIVE char_directive
55 # define DIRECTIVES char_directives
56 #endif
57 
58 #ifdef STATIC
59 STATIC
60 #endif
61 int
62 PRINTF_PARSE (const CHAR_T *format, DIRECTIVES *d, arguments *a)
63 {
64   const CHAR_T *cp = format;		/* pointer into format */
65   size_t arg_posn = 0;		/* number of regular arguments consumed */
66   size_t d_allocated;			/* allocated elements of d->dir */
67   size_t a_allocated;			/* allocated elements of a->arg */
68   size_t max_width_length = 0;
69   size_t max_precision_length = 0;
70 
71   d->count = 0;
72   d_allocated = 1;
73   d->dir = malloc (d_allocated * sizeof (DIRECTIVE));
74   if (d->dir == NULL)
75     /* Out of memory.  */
76     return -1;
77 
78   a->count = 0;
79   a_allocated = 0;
80   a->arg = NULL;
81 
82 #define REGISTER_ARG(_index_,_type_) \
83   {									\
84     size_t n = (_index_);						\
85     if (n >= a_allocated)						\
86       {									\
87 	size_t memory_size;						\
88 	argument *memory;						\
89 									\
90 	a_allocated = xtimes (a_allocated, 2);				\
91 	if (a_allocated <= n)						\
92 	  a_allocated = xsum (n, 1);					\
93 	memory_size = xtimes (a_allocated, sizeof (argument));		\
94 	if (size_overflow_p (memory_size))				\
95 	  /* Overflow, would lead to out of memory.  */			\
96 	  goto error;							\
97 	memory = (a->arg						\
98 		  ? realloc (a->arg, memory_size)			\
99 		  : malloc (memory_size));				\
100 	if (memory == NULL)						\
101 	  /* Out of memory.  */						\
102 	  goto error;							\
103 	a->arg = memory;						\
104       }									\
105     while (a->count <= n)						\
106       a->arg[a->count++].type = TYPE_NONE;				\
107     if (a->arg[n].type == TYPE_NONE)					\
108       a->arg[n].type = (_type_);					\
109     else if (a->arg[n].type != (_type_))				\
110       /* Ambiguous type for positional argument.  */			\
111       goto error;							\
112   }
113 
114   while (*cp != '\0')
115     {
116       CHAR_T c = *cp++;
117       if (c == '%')
118 	{
119 	  size_t arg_index = ARG_NONE;
120 	  DIRECTIVE *dp = &d->dir[d->count];/* pointer to next directive */
121 
122 	  /* Initialize the next directive.  */
123 	  dp->dir_start = cp - 1;
124 	  dp->flags = 0;
125 	  dp->width_start = NULL;
126 	  dp->width_end = NULL;
127 	  dp->width_arg_index = ARG_NONE;
128 	  dp->precision_start = NULL;
129 	  dp->precision_end = NULL;
130 	  dp->precision_arg_index = ARG_NONE;
131 	  dp->arg_index = ARG_NONE;
132 
133 	  /* Test for positional argument.  */
134 	  if (*cp >= '0' && *cp <= '9')
135 	    {
136 	      const CHAR_T *np;
137 
138 	      for (np = cp; *np >= '0' && *np <= '9'; np++)
139 		;
140 	      if (*np == '$')
141 		{
142 		  size_t n = 0;
143 
144 		  for (np = cp; *np >= '0' && *np <= '9'; np++)
145 		    n = xsum (xtimes (n, 10), *np - '0');
146 		  if (n == 0)
147 		    /* Positional argument 0.  */
148 		    goto error;
149 		  if (size_overflow_p (n))
150 		    /* n too large, would lead to out of memory later.  */
151 		    goto error;
152 		  arg_index = n - 1;
153 		  cp = np + 1;
154 		}
155 	    }
156 
157 	  /* Read the flags.  */
158 	  for (;;)
159 	    {
160 	      if (*cp == '\'')
161 		{
162 		  dp->flags |= FLAG_GROUP;
163 		  cp++;
164 		}
165 	      else if (*cp == '-')
166 		{
167 		  dp->flags |= FLAG_LEFT;
168 		  cp++;
169 		}
170 	      else if (*cp == '+')
171 		{
172 		  dp->flags |= FLAG_SHOWSIGN;
173 		  cp++;
174 		}
175 	      else if (*cp == ' ')
176 		{
177 		  dp->flags |= FLAG_SPACE;
178 		  cp++;
179 		}
180 	      else if (*cp == '#')
181 		{
182 		  dp->flags |= FLAG_ALT;
183 		  cp++;
184 		}
185 	      else if (*cp == '0')
186 		{
187 		  dp->flags |= FLAG_ZERO;
188 		  cp++;
189 		}
190 	      else
191 		break;
192 	    }
193 
194 	  /* Parse the field width.  */
195 	  if (*cp == '*')
196 	    {
197 	      dp->width_start = cp;
198 	      cp++;
199 	      dp->width_end = cp;
200 	      if (max_width_length < 1)
201 		max_width_length = 1;
202 
203 	      /* Test for positional argument.  */
204 	      if (*cp >= '0' && *cp <= '9')
205 		{
206 		  const CHAR_T *np;
207 
208 		  for (np = cp; *np >= '0' && *np <= '9'; np++)
209 		    ;
210 		  if (*np == '$')
211 		    {
212 		      size_t n = 0;
213 
214 		      for (np = cp; *np >= '0' && *np <= '9'; np++)
215 			n = xsum (xtimes (n, 10), *np - '0');
216 		      if (n == 0)
217 			/* Positional argument 0.  */
218 			goto error;
219 		      if (size_overflow_p (n))
220 			/* n too large, would lead to out of memory later.  */
221 			goto error;
222 		      dp->width_arg_index = n - 1;
223 		      cp = np + 1;
224 		    }
225 		}
226 	      if (dp->width_arg_index == ARG_NONE)
227 		{
228 		  dp->width_arg_index = arg_posn++;
229 		  if (dp->width_arg_index == ARG_NONE)
230 		    /* arg_posn wrapped around.  */
231 		    goto error;
232 		}
233 	      REGISTER_ARG (dp->width_arg_index, TYPE_INT);
234 	    }
235 	  else if (*cp >= '0' && *cp <= '9')
236 	    {
237 	      size_t width_length;
238 
239 	      dp->width_start = cp;
240 	      for (; *cp >= '0' && *cp <= '9'; cp++)
241 		;
242 	      dp->width_end = cp;
243 	      width_length = dp->width_end - dp->width_start;
244 	      if (max_width_length < width_length)
245 		max_width_length = width_length;
246 	    }
247 
248 	  /* Parse the precision.  */
249 	  if (*cp == '.')
250 	    {
251 	      cp++;
252 	      if (*cp == '*')
253 		{
254 		  dp->precision_start = cp - 1;
255 		  cp++;
256 		  dp->precision_end = cp;
257 		  if (max_precision_length < 2)
258 		    max_precision_length = 2;
259 
260 		  /* Test for positional argument.  */
261 		  if (*cp >= '0' && *cp <= '9')
262 		    {
263 		      const CHAR_T *np;
264 
265 		      for (np = cp; *np >= '0' && *np <= '9'; np++)
266 			;
267 		      if (*np == '$')
268 			{
269 			  size_t n = 0;
270 
271 			  for (np = cp; *np >= '0' && *np <= '9'; np++)
272 			    n = xsum (xtimes (n, 10), *np - '0');
273 			  if (n == 0)
274 			    /* Positional argument 0.  */
275 			    goto error;
276 			  if (size_overflow_p (n))
277 			    /* n too large, would lead to out of memory
278 			       later.  */
279 			    goto error;
280 			  dp->precision_arg_index = n - 1;
281 			  cp = np + 1;
282 			}
283 		    }
284 		  if (dp->precision_arg_index == ARG_NONE)
285 		    {
286 		      dp->precision_arg_index = arg_posn++;
287 		      if (dp->precision_arg_index == ARG_NONE)
288 			/* arg_posn wrapped around.  */
289 			goto error;
290 		    }
291 		  REGISTER_ARG (dp->precision_arg_index, TYPE_INT);
292 		}
293 	      else
294 		{
295 		  size_t precision_length;
296 
297 		  dp->precision_start = cp - 1;
298 		  for (; *cp >= '0' && *cp <= '9'; cp++)
299 		    ;
300 		  dp->precision_end = cp;
301 		  precision_length = dp->precision_end - dp->precision_start;
302 		  if (max_precision_length < precision_length)
303 		    max_precision_length = precision_length;
304 		}
305 	    }
306 
307 	  {
308 	    arg_type type;
309 
310 	    /* Parse argument type/size specifiers.  */
311 	    {
312 	      int flags = 0;
313 
314 	      for (;;)
315 		{
316 		  if (*cp == 'h')
317 		    {
318 		      flags |= (1 << (flags & 1));
319 		      cp++;
320 		    }
321 		  else if (*cp == 'L')
322 		    {
323 		      flags |= 4;
324 		      cp++;
325 		    }
326 		  else if (*cp == 'l')
327 		    {
328 		      flags += 8;
329 		      cp++;
330 		    }
331 #ifdef HAVE_INTMAX_T
332 		  else if (*cp == 'j')
333 		    {
334 		      if (sizeof (intmax_t) > sizeof (long))
335 			{
336 			  /* intmax_t = long long */
337 			  flags += 16;
338 			}
339 		      else if (sizeof (intmax_t) > sizeof (int))
340 			{
341 			  /* intmax_t = long */
342 			  flags += 8;
343 			}
344 		      cp++;
345 		    }
346 #endif
347 		  else if (*cp == 'z' || *cp == 'Z')
348 		    {
349 		      /* 'z' is standardized in ISO C 99, but glibc uses 'Z'
350 			 because the warning facility in gcc-2.95.2 understands
351 			 only 'Z' (see gcc-2.95.2/gcc/c-common.c:1784).  */
352 		      if (sizeof (size_t) > sizeof (long))
353 			{
354 			  /* size_t = long long */
355 			  flags += 16;
356 			}
357 		      else if (sizeof (size_t) > sizeof (int))
358 			{
359 			  /* size_t = long */
360 			  flags += 8;
361 			}
362 		      cp++;
363 		    }
364 		  else if (*cp == 't')
365 		    {
366 		      if (sizeof (ptrdiff_t) > sizeof (long))
367 			{
368 			  /* ptrdiff_t = long long */
369 			  flags += 16;
370 			}
371 		      else if (sizeof (ptrdiff_t) > sizeof (int))
372 			{
373 			  /* ptrdiff_t = long */
374 			  flags += 8;
375 			}
376 		      cp++;
377 		    }
378 		  else
379 		    break;
380 		}
381 
382 	      /* Read the conversion character.  */
383 	      c = *cp++;
384 	      switch (c)
385 		{
386 		case 'd': case 'i':
387 #ifdef HAVE_LONG_LONG
388 		  if (flags >= 16 || (flags & 4))
389 		    type = TYPE_LONGLONGINT;
390 		  else
391 #endif
392 		  if (flags >= 8)
393 		    type = TYPE_LONGINT;
394 		  else if (flags & 2)
395 		    type = TYPE_SCHAR;
396 		  else if (flags & 1)
397 		    type = TYPE_SHORT;
398 		  else
399 		    type = TYPE_INT;
400 		  break;
401 		case 'o': case 'u': case 'x': case 'X':
402 #ifdef HAVE_LONG_LONG
403 		  if (flags >= 16 || (flags & 4))
404 		    type = TYPE_ULONGLONGINT;
405 		  else
406 #endif
407 		  if (flags >= 8)
408 		    type = TYPE_ULONGINT;
409 		  else if (flags & 2)
410 		    type = TYPE_UCHAR;
411 		  else if (flags & 1)
412 		    type = TYPE_USHORT;
413 		  else
414 		    type = TYPE_UINT;
415 		  break;
416 		case 'f': case 'F': case 'e': case 'E': case 'g': case 'G':
417 		case 'a': case 'A':
418 #ifdef HAVE_LONG_DOUBLE
419 		  if (flags >= 16 || (flags & 4))
420 		    type = TYPE_LONGDOUBLE;
421 		  else
422 #endif
423 		  type = TYPE_DOUBLE;
424 		  break;
425 		case 'c':
426 		  if (flags >= 8)
427 #ifdef HAVE_WINT_T
428 		    type = TYPE_WIDE_CHAR;
429 #else
430 		    goto error;
431 #endif
432 		  else
433 		    type = TYPE_CHAR;
434 		  break;
435 #ifdef HAVE_WINT_T
436 		case 'C':
437 		  type = TYPE_WIDE_CHAR;
438 		  c = 'c';
439 		  break;
440 #endif
441 		case 's':
442 		  if (flags >= 8)
443 #ifdef HAVE_WCHAR_T
444 		    type = TYPE_WIDE_STRING;
445 #else
446 		    goto error;
447 #endif
448 		  else
449 		    type = TYPE_STRING;
450 		  break;
451 #ifdef HAVE_WCHAR_T
452 		case 'S':
453 		  type = TYPE_WIDE_STRING;
454 		  c = 's';
455 		  break;
456 #endif
457 		case 'p':
458 		  type = TYPE_POINTER;
459 		  break;
460 		case 'n':
461 #ifdef HAVE_LONG_LONG
462 		  if (flags >= 16 || (flags & 4))
463 		    type = TYPE_COUNT_LONGLONGINT_POINTER;
464 		  else
465 #endif
466 		  if (flags >= 8)
467 		    type = TYPE_COUNT_LONGINT_POINTER;
468 		  else if (flags & 2)
469 		    type = TYPE_COUNT_SCHAR_POINTER;
470 		  else if (flags & 1)
471 		    type = TYPE_COUNT_SHORT_POINTER;
472 		  else
473 		    type = TYPE_COUNT_INT_POINTER;
474 		  break;
475 		case '%':
476 		  type = TYPE_NONE;
477 		  break;
478 		default:
479 		  /* Unknown conversion character.  */
480 		  goto error;
481 		}
482 	    }
483 
484 	    if (type != TYPE_NONE)
485 	      {
486 		dp->arg_index = arg_index;
487 		if (dp->arg_index == ARG_NONE)
488 		  {
489 		    dp->arg_index = arg_posn++;
490 		    if (dp->arg_index == ARG_NONE)
491 		      /* arg_posn wrapped around.  */
492 		      goto error;
493 		  }
494 		REGISTER_ARG (dp->arg_index, type);
495 	      }
496 	    dp->conversion = c;
497 	    dp->dir_end = cp;
498 	  }
499 
500 	  d->count++;
501 	  if (d->count >= d_allocated)
502 	    {
503 	      size_t memory_size;
504 	      DIRECTIVE *memory;
505 
506 	      d_allocated = xtimes (d_allocated, 2);
507 	      memory_size = xtimes (d_allocated, sizeof (DIRECTIVE));
508 	      if (size_overflow_p (memory_size))
509 		/* Overflow, would lead to out of memory.  */
510 		goto error;
511 	      memory = realloc (d->dir, memory_size);
512 	      if (memory == NULL)
513 		/* Out of memory.  */
514 		goto error;
515 	      d->dir = memory;
516 	    }
517 	}
518     }
519   d->dir[d->count].dir_start = cp;
520 
521   d->max_width_length = max_width_length;
522   d->max_precision_length = max_precision_length;
523   return 0;
524 
525 error:
526   if (a->arg)
527     free (a->arg);
528   if (d->dir)
529     free (d->dir);
530   return -1;
531 }
532 
533 #undef DIRECTIVES
534 #undef DIRECTIVE
535 #undef CHAR_T
536 #undef PRINTF_PARSE
537