src/stdio/vfwprintf.c

   1 #include "stdio_impl.h"
   2 #include <errno.h>
   3 #include <ctype.h>
   4 #include <limits.h>
   5 #include <string.h>
   6 #include <stdarg.h>
   7 #include <stddef.h>
   8 #include <stdlib.h>
   9 #include <wchar.h>
  10 #include <inttypes.h>
  11
  12 /* Convenient bit representation for modifier flags, which all fall
  13  * within 31 codepoints of the space character. */
  14
  15 #define ALT_FORM   (1U<<'#'-' ')
  16 #define ZERO_PAD   (1U<<'0'-' ')
  17 #define LEFT_ADJ   (1U<<'-'-' ')
  18 #define PAD_POS    (1U<<' '-' ')
  19 #define MARK_POS   (1U<<'+'-' ')
  20 #define GROUPED    (1U<<'\''-' ')
  21
  22 #define FLAGMASK (ALT_FORM|ZERO_PAD|LEFT_ADJ|PAD_POS|MARK_POS|GROUPED)
  23
  24 /* State machine to accept length modifiers + conversion specifiers.
  25  * Result is 0 on failure, or an argument type to pop on success. */
  26
  27 enum {
  28         BARE, LPRE, LLPRE, HPRE, HHPRE, BIGLPRE,
  29         ZTPRE, JPRE,
  30         STOP,
  31         PTR, INT, UINT, ULLONG,
  32         LONG, ULONG,
  33         SHORT, USHORT, CHAR, UCHAR,
  34         LLONG, SIZET, IMAX, UMAX, PDIFF, UIPTR,
  35         DBL, LDBL,
  36         NOARG,
  37         MAXSTATE
  38 };
  39
  40 #define S(x) [(x)-'A']
  41
  42 static const unsigned char states[]['z'-'A'+1] = {
  43         { /* 0: bare types */
  44                 S('d') = INT, S('i') = INT,
  45                 S('o') = UINT, S('u') = UINT, S('x') = UINT, S('X') = UINT,
  46                 S('e') = DBL, S('f') = DBL, S('g') = DBL, S('a') = DBL,
  47                 S('E') = DBL, S('F') = DBL, S('G') = DBL, S('A') = DBL,
  48                 S('c') = INT, S('C') = UINT,
  49                 S('s') = PTR, S('S') = PTR, S('p') = UIPTR, S('n') = PTR,
  50                 S('m') = NOARG,
  51                 S('l') = LPRE, S('h') = HPRE, S('L') = BIGLPRE,
  52                 S('z') = ZTPRE, S('j') = JPRE, S('t') = ZTPRE,
  53         }, { /* 1: l-prefixed */
  54                 S('d') = LONG, S('i') = LONG,
  55                 S('o') = ULONG, S('u') = ULONG, S('x') = ULONG, S('X') = ULONG,
  56                 S('e') = DBL, S('f') = DBL, S('g') = DBL, S('a') = DBL,
  57                 S('E') = DBL, S('F') = DBL, S('G') = DBL, S('A') = DBL,
  58                 S('c') = UINT, S('s') = PTR, S('n') = PTR,
  59                 S('l') = LLPRE,
  60         }, { /* 2: ll-prefixed */
  61                 S('d') = LLONG, S('i') = LLONG,
  62                 S('o') = ULLONG, S('u') = ULLONG,
  63                 S('x') = ULLONG, S('X') = ULLONG,
  64                 S('n') = PTR,
  65         }, { /* 3: h-prefixed */
  66                 S('d') = SHORT, S('i') = SHORT,
  67                 S('o') = USHORT, S('u') = USHORT,
  68                 S('x') = USHORT, S('X') = USHORT,
  69                 S('n') = PTR,
  70                 S('h') = HHPRE,
  71         }, { /* 4: hh-prefixed */
  72                 S('d') = CHAR, S('i') = CHAR,
  73                 S('o') = UCHAR, S('u') = UCHAR,
  74                 S('x') = UCHAR, S('X') = UCHAR,
  75                 S('n') = PTR,
  76         }, { /* 5: L-prefixed */
  77                 S('e') = LDBL, S('f') = LDBL, S('g') = LDBL, S('a') = LDBL,
  78                 S('E') = LDBL, S('F') = LDBL, S('G') = LDBL, S('A') = LDBL,
  79                 S('n') = PTR,
  80         }, { /* 6: z- or t-prefixed (assumed to be same size) */
  81                 S('d') = PDIFF, S('i') = PDIFF,
  82                 S('o') = SIZET, S('u') = SIZET,
  83                 S('x') = SIZET, S('X') = SIZET,
  84                 S('n') = PTR,
  85         }, { /* 7: j-prefixed */
  86                 S('d') = IMAX, S('i') = IMAX,
  87                 S('o') = UMAX, S('u') = UMAX,
  88                 S('x') = UMAX, S('X') = UMAX,
  89                 S('n') = PTR,
  90         }
  91 };
  92
  93 #define OOB(x) ((unsigned)(x)-'A' > 'z'-'A')
  94
  95 union arg
  96 {
  97         uintmax_t i;
  98         long double f;
  99         void *p;
 100 };
 101
 102 static void pop_arg(union arg *arg, int type, va_list *ap)
 103 {
 104         switch (type) {
 105                case PTR:        arg->p = va_arg(*ap, void *);
 106         break; case INT:        arg->i = va_arg(*ap, int);
 107         break; case UINT:       arg->i = va_arg(*ap, unsigned int);
 108         break; case LONG:       arg->i = va_arg(*ap, long);
 109         break; case ULONG:      arg->i = va_arg(*ap, unsigned long);
 110         break; case ULLONG:     arg->i = va_arg(*ap, unsigned long long);
 111         break; case SHORT:      arg->i = (short)va_arg(*ap, int);
 112         break; case USHORT:     arg->i = (unsigned short)va_arg(*ap, int);
 113         break; case CHAR:       arg->i = (signed char)va_arg(*ap, int);
 114         break; case UCHAR:      arg->i = (unsigned char)va_arg(*ap, int);
 115         break; case LLONG:      arg->i = va_arg(*ap, long long);
 116         break; case SIZET:      arg->i = va_arg(*ap, size_t);
 117         break; case IMAX:       arg->i = va_arg(*ap, intmax_t);
 118         break; case UMAX:       arg->i = va_arg(*ap, uintmax_t);
 119         break; case PDIFF:      arg->i = va_arg(*ap, ptrdiff_t);
 120         break; case UIPTR:      arg->i = (uintptr_t)va_arg(*ap, void *);
 121         break; case DBL:        arg->f = va_arg(*ap, double);
 122         break; case LDBL:       arg->f = va_arg(*ap, long double);
 123         }
 124 }
 125
 126 static void out(FILE *f, const wchar_t *s, size_t l)
 127 {
 128         while (l-- && !ferror(f)) fputwc(*s++, f);
 129 }
 130
 131 static void pad(FILE *f, int n, int fl)
 132 {
 133         if ((fl & LEFT_ADJ) || !n || ferror(f)) return;
 134         fprintf(f, "%*s", n, "");
 135 }
 136
 137 static int getint(wchar_t **s) {
 138         int i;
 139         for (i=0; iswdigit(**s); (*s)++) {
 140                 if (i > INT_MAX/10U || **s-'0' > INT_MAX-10*i) i = -1;
 141                 else i = 10*i + (**s-'0');
 142         }
 143         return i;
 144 }
 145
 146 static const char sizeprefix['y'-'a'] = {
 147 ['a'-'a']='L', ['e'-'a']='L', ['f'-'a']='L', ['g'-'a']='L',
 148 ['d'-'a']='j', ['i'-'a']='j', ['o'-'a']='j', ['u'-'a']='j', ['x'-'a']='j',
 149 ['p'-'a']='j'
 150 };
 151
 152 static int wprintf_core(FILE *f, const wchar_t *fmt, va_list *ap, union arg *nl_arg, int *nl_type)
 153 {
 154         wchar_t *a, *z, *s=(wchar_t *)fmt;
 155         unsigned l10n=0, fl;
 156         int w, p, xp;
 157         union arg arg;
 158         int argpos;
 159         unsigned st, ps;
 160         int cnt=0, l=0;
 161         int i;
 162         int t;
 163         char *bs;
 164         char charfmt[16];
 165         wchar_t wc;
 166
 167         for (;;) {
 168                 /* This error is only specified for snprintf, but since it's
 169                  * unspecified for other forms, do the same. Stop immediately
 170                  * on overflow; otherwise %n could produce wrong results. */
 171                 if (l > INT_MAX - cnt) goto overflow;
 172
 173                 /* Update output count, end loop when fmt is exhausted */
 174                 cnt += l;
 175                 if (!*s) break;
 176
 177                 /* Handle literal text and %% format specifiers */
 178                 for (a=s; *s && *s!='%'; s++);
 179                 for (z=s; s[0]=='%' && s[1]=='%'; z++, s+=2);
 180                 if (z-a > INT_MAX-cnt) goto overflow;
 181                 l = z-a;
 182                 if (f) out(f, a, l);
 183                 if (l) continue;
 184
 185                 if (iswdigit(s[1]) && s[2]=='$') {
 186                         l10n=1;
 187                         argpos = s[1]-'0';
 188                         s+=3;
 189                 } else {
 190                         argpos = -1;
 191                         s++;
 192                 }
 193
 194                 /* Read modifier flags */
 195                 for (fl=0; (unsigned)*s-' '<32 && (FLAGMASK&(1U<<*s-' ')); s++)
 196                         fl |= 1U<<*s-' ';
 197
 198                 /* Read field width */
 199                 if (*s=='*') {
 200                         if (iswdigit(s[1]) && s[2]=='$') {
 201                                 l10n=1;
 202                                 nl_type[s[1]-'0'] = INT;
 203                                 w = nl_arg[s[1]-'0'].i;
 204                                 s+=3;
 205                         } else if (!l10n) {
 206                                 w = f ? va_arg(*ap, int) : 0;
 207                                 s++;
 208                         } else goto inval;
 209                         if (w<0) fl|=LEFT_ADJ, w=-w;
 210                 } else if ((w=getint(&s))<0) goto overflow;
 211
 212                 /* Read precision */
 213                 if (*s=='.' && s[1]=='*') {
 214                         if (isdigit(s[2]) && s[3]=='$') {
 215                                 nl_type[s[2]-'0'] = INT;
 216                                 p = nl_arg[s[2]-'0'].i;
 217                                 s+=4;
 218                         } else if (!l10n) {
 219                                 p = f ? va_arg(*ap, int) : 0;
 220                                 s+=2;
 221                         } else goto inval;
 222                         xp = (p>=0);
 223                 } else if (*s=='.') {
 224                         s++;
 225                         p = getint(&s);
 226                         xp = 1;
 227                 } else {
 228                         p = -1;
 229                         xp = 0;
 230                 }
 231
 232                 /* Format specifier state machine */
 233                 st=0;
 234                 do {
 235                         if (OOB(*s)) goto inval;
 236                         ps=st;
 237                         st=states[st]S(*s++);
 238                 } while (st-1<STOP);
 239                 if (!st) goto inval;
 240
 241                 /* Check validity of argument type (nl/normal) */
 242                 if (st==NOARG) {
 243                         if (argpos>=0) goto inval;
 244                 } else {
 245                         if (argpos>=0) nl_type[argpos]=st, arg=nl_arg[argpos];
 246                         else if (f) pop_arg(&arg, st, ap);
 247                         else return 0;
 248                 }
 249
 250                 if (!f) continue;
 251
 252                 /* Do not process any new directives once in error state. */
 253                 if (ferror(f)) return -1;
 254
 255                 t = s[-1];
 256                 if (ps && (t&15)==3) t&=~32;
 257
 258                 switch (t) {
 259                 case 'n':
 260                         switch(ps) {
 261                         case BARE: *(int *)arg.p = cnt; break;
 262                         case LPRE: *(long *)arg.p = cnt; break;
 263                         case LLPRE: *(long long *)arg.p = cnt; break;
 264                         case HPRE: *(unsigned short *)arg.p = cnt; break;
 265                         case HHPRE: *(unsigned char *)arg.p = cnt; break;
 266                         case ZTPRE: *(size_t *)arg.p = cnt; break;
 267                         case JPRE: *(uintmax_t *)arg.p = cnt; break;
 268                         }
 269                         continue;
 270                 case 'c':
 271                 case 'C':
 272                         if (w<1) w=1;
 273                         pad(f, w-1, fl);
 274                         out(f, &(wchar_t){t=='C' ? arg.i : btowc(arg.i)}, 1);
 275                         pad(f, w-1, fl^LEFT_ADJ);
 276                         l = w;
 277                         continue;
 278                 case 'S':
 279                         a = arg.p;
 280                         z = a + wcsnlen(a, p<0 ? INT_MAX : p);
 281                         if (p<0 && *z) goto overflow;
 282                         p = z-a;
 283                         if (w<p) w=p;
 284                         pad(f, w-p, fl);
 285                         out(f, a, p);
 286                         pad(f, w-p, fl^LEFT_ADJ);
 287                         l=w;
 288                         continue;
 289                 case 'm':
 290                         arg.p = strerror(errno);
 291                 case 's':
 292                         if (!arg.p) arg.p = "(null)";
 293                         bs = arg.p;
 294                         for (i=l=0; l<(p<0?INT_MAX:p) && (i=mbtowc(&wc, bs, MB_LEN_MAX))>0; bs+=i, l++);
 295                         if (i<0) return -1;
 296                         if (p<0 && *bs) goto overflow;
 297                         p=l;
 298                         if (w<p) w=p;
 299                         pad(f, w-p, fl);
 300                         bs = arg.p;
 301                         while (l--) {
 302                                 i=mbtowc(&wc, bs, MB_LEN_MAX);
 303                                 bs+=i;
 304                                 out(f, &wc, 1);
 305                         }
 306                         pad(f, w-p, fl^LEFT_ADJ);
 307                         l=w;
 308                         continue;
 309                 }
 310
 311                 if (xp && p<0) goto overflow;
 312                 snprintf(charfmt, sizeof charfmt, "%%%s%s%s%s%s*.*%c%c",
 313                         "#"+!(fl & ALT_FORM),
 314                         "+"+!(fl & MARK_POS),
 315                         "-"+!(fl & LEFT_ADJ),
 316                         " "+!(fl & PAD_POS),
 317                         "0"+!(fl & ZERO_PAD),
 318                         sizeprefix[(t|32)-'a'], t);
 319
 320                 switch (t|32) {
 321                 case 'a': case 'e': case 'f': case 'g':
 322                         l = fprintf(f, charfmt, w, p, arg.f);
 323                         break;
 324                 case 'd': case 'i': case 'o': case 'u': case 'x': case 'p':
 325                         l = fprintf(f, charfmt, w, p, arg.i);
 326                         break;
 327                 }
 328         }
 329
 330         if (f) return cnt;
 331         if (!l10n) return 0;
 332
 333         for (i=1; i<=NL_ARGMAX && nl_type[i]; i++)
 334                 pop_arg(nl_arg+i, nl_type[i], ap);
 335         for (; i<=NL_ARGMAX && !nl_type[i]; i++);
 336         if (i<=NL_ARGMAX) return -1;
 337         return 1;
 338
 339 inval:
 340         errno = EINVAL;
 341         return -1;
 342 overflow:
 343         errno = EOVERFLOW;
 344         return -1;
 345 }
 346
 347 int vfwprintf(FILE *restrict f, const wchar_t *restrict fmt, va_list ap)
 348 {
 349         va_list ap2;
 350         int nl_type[NL_ARGMAX+1] = {0};
 351         union arg nl_arg[NL_ARGMAX+1];
 352         int olderr;
 353         int ret;
 354
 355         /* the copy allows passing va_list* even if va_list is an array */
 356         va_copy(ap2, ap);
 357         if (wprintf_core(0, fmt, &ap2, nl_arg, nl_type) < 0) {
 358                 va_end(ap2);
 359                 return -1;
 360         }
 361
 362         FLOCK(f);
 363         fwide(f, 1);
 364         olderr = f->flags & F_ERR;
 365         f->flags &= ~F_ERR;
 366         ret = wprintf_core(f, fmt, &ap2, nl_arg, nl_type);
 367         if (ferror(f)) ret = -1;
 368         f->flags |= olderr;
 369         FUNLOCK(f);
 370         va_end(ap2);
 371         return ret;
 372 }