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 #ifdef __wasilibc_unmodified_upstream // Changes to optimize printf/scanf when long double isn't needed
12 #else
13 #include "printscan.h"
14 #endif
15
16 /* Convenient bit representation for modifier flags, which all fall
17 * within 31 codepoints of the space character. */
18
19 #define ALT_FORM (1U<<'#'-' ')
20 #define ZERO_PAD (1U<<'0'-' ')
21 #define LEFT_ADJ (1U<<'-'-' ')
22 #define PAD_POS (1U<<' '-' ')
23 #define MARK_POS (1U<<'+'-' ')
24 #define GROUPED (1U<<'\''-' ')
25
26 #define FLAGMASK (ALT_FORM|ZERO_PAD|LEFT_ADJ|PAD_POS|MARK_POS|GROUPED)
27
28 /* State machine to accept length modifiers + conversion specifiers.
29 * Result is 0 on failure, or an argument type to pop on success. */
30
31 enum {
32 BARE, LPRE, LLPRE, HPRE, HHPRE, BIGLPRE,
33 ZTPRE, JPRE,
34 STOP,
35 PTR, INT, UINT, ULLONG,
36 LONG, ULONG,
37 SHORT, USHORT, CHAR, UCHAR,
38 LLONG, SIZET, IMAX, UMAX, PDIFF, UIPTR,
39 DBL, LDBL,
40 NOARG,
41 MAXSTATE
42 };
43
44 #define S(x) [(x)-'A']
45
46 static const unsigned char states[]['z'-'A'+1] = {
47 { /* 0: bare types */
48 S('d') = INT, S('i') = INT,
49 S('o') = UINT, S('u') = UINT, S('x') = UINT, S('X') = UINT,
50 S('e') = DBL, S('f') = DBL, S('g') = DBL, S('a') = DBL,
51 S('E') = DBL, S('F') = DBL, S('G') = DBL, S('A') = DBL,
52 S('c') = CHAR, S('C') = INT,
53 S('s') = PTR, S('S') = PTR, S('p') = UIPTR, S('n') = PTR,
54 S('m') = NOARG,
55 S('l') = LPRE, S('h') = HPRE, S('L') = BIGLPRE,
56 S('z') = ZTPRE, S('j') = JPRE, S('t') = ZTPRE,
57 }, { /* 1: l-prefixed */
58 S('d') = LONG, S('i') = LONG,
59 S('o') = ULONG, S('u') = ULONG, S('x') = ULONG, S('X') = ULONG,
60 S('e') = DBL, S('f') = DBL, S('g') = DBL, S('a') = DBL,
61 S('E') = DBL, S('F') = DBL, S('G') = DBL, S('A') = DBL,
62 S('c') = INT, S('s') = PTR, S('n') = PTR,
63 S('l') = LLPRE,
64 }, { /* 2: ll-prefixed */
65 S('d') = LLONG, S('i') = LLONG,
66 S('o') = ULLONG, S('u') = ULLONG,
67 S('x') = ULLONG, S('X') = ULLONG,
68 S('n') = PTR,
69 }, { /* 3: h-prefixed */
70 S('d') = SHORT, S('i') = SHORT,
71 S('o') = USHORT, S('u') = USHORT,
72 S('x') = USHORT, S('X') = USHORT,
73 S('n') = PTR,
74 S('h') = HHPRE,
75 }, { /* 4: hh-prefixed */
76 S('d') = CHAR, S('i') = CHAR,
77 S('o') = UCHAR, S('u') = UCHAR,
78 S('x') = UCHAR, S('X') = UCHAR,
79 S('n') = PTR,
80 }, { /* 5: L-prefixed */
81 S('e') = LDBL, S('f') = LDBL, S('g') = LDBL, S('a') = LDBL,
82 S('E') = LDBL, S('F') = LDBL, S('G') = LDBL, S('A') = LDBL,
83 S('n') = PTR,
84 }, { /* 6: z- or t-prefixed (assumed to be same size) */
85 S('d') = PDIFF, S('i') = PDIFF,
86 S('o') = SIZET, S('u') = SIZET,
87 S('x') = SIZET, S('X') = SIZET,
88 S('n') = PTR,
89 }, { /* 7: j-prefixed */
90 S('d') = IMAX, S('i') = IMAX,
91 S('o') = UMAX, S('u') = UMAX,
92 S('x') = UMAX, S('X') = UMAX,
93 S('n') = PTR,
94 }
95 };
96
97 #define OOB(x) ((unsigned)(x)-'A' > 'z'-'A')
98
99 union arg
100 {
101 uintmax_t i;
102 #if !defined(__wasilibc_printscan_no_floating_point)
103 #if defined(__wasilibc_printscan_no_long_double)
104 long_double f;
105 #else
106 long double f;
107 #endif
108 #endif
109 void *p;
110 };
111
pop_arg(union arg * arg,int type,va_list * ap)112 static void pop_arg(union arg *arg, int type, va_list *ap)
113 {
114 switch (type) {
115 case PTR: arg->p = va_arg(*ap, void *);
116 break; case INT: arg->i = va_arg(*ap, int);
117 break; case UINT: arg->i = va_arg(*ap, unsigned int);
118 break; case LONG: arg->i = va_arg(*ap, long);
119 break; case ULONG: arg->i = va_arg(*ap, unsigned long);
120 break; case ULLONG: arg->i = va_arg(*ap, unsigned long long);
121 break; case SHORT: arg->i = (short)va_arg(*ap, int);
122 break; case USHORT: arg->i = (unsigned short)va_arg(*ap, int);
123 break; case CHAR: arg->i = (signed char)va_arg(*ap, int);
124 break; case UCHAR: arg->i = (unsigned char)va_arg(*ap, int);
125 break; case LLONG: arg->i = va_arg(*ap, long long);
126 break; case SIZET: arg->i = va_arg(*ap, size_t);
127 break; case IMAX: arg->i = va_arg(*ap, intmax_t);
128 break; case UMAX: arg->i = va_arg(*ap, uintmax_t);
129 break; case PDIFF: arg->i = va_arg(*ap, ptrdiff_t);
130 break; case UIPTR: arg->i = (uintptr_t)va_arg(*ap, void *);
131 #if defined(__wasilibc_printscan_no_floating_point)
132 break; case DBL:
133 break; case LDBL: floating_point_not_supported();
134 #else
135 break; case DBL: arg->f = va_arg(*ap, double);
136 #if defined(__wasilibc_printscan_no_long_double)
137 break; case LDBL: long_double_not_supported();
138 #else
139 break; case LDBL: arg->f = va_arg(*ap, long double);
140 #endif
141 #endif
142 }
143 }
144
out(FILE * f,const wchar_t * s,size_t l)145 static void out(FILE *f, const wchar_t *s, size_t l)
146 {
147 while (l-- && !(f->flags & F_ERR)) fputwc(*s++, f);
148 }
149
getint(wchar_t ** s)150 static int getint(wchar_t **s) {
151 int i;
152 for (i=0; iswdigit(**s); (*s)++) {
153 if (i > INT_MAX/10U || **s-'0' > INT_MAX-10*i) i = -1;
154 else i = 10*i + (**s-'0');
155 }
156 return i;
157 }
158
159 static const char sizeprefix['y'-'a'] = {
160 ['a'-'a']='L', ['e'-'a']='L', ['f'-'a']='L', ['g'-'a']='L',
161 ['d'-'a']='j', ['i'-'a']='j', ['o'-'a']='j', ['u'-'a']='j', ['x'-'a']='j',
162 ['p'-'a']='j'
163 };
164
wprintf_core(FILE * f,const wchar_t * fmt,va_list * ap,union arg * nl_arg,int * nl_type)165 static int wprintf_core(FILE *f, const wchar_t *fmt, va_list *ap, union arg *nl_arg, int *nl_type)
166 {
167 wchar_t *a, *z, *s=(wchar_t *)fmt;
168 unsigned l10n=0, fl;
169 int w, p, xp;
170 union arg arg;
171 int argpos;
172 unsigned st, ps;
173 int cnt=0, l=0;
174 int i;
175 int t;
176 char *bs;
177 char charfmt[16];
178 wchar_t wc;
179
180 for (;;) {
181 /* This error is only specified for snprintf, but since it's
182 * unspecified for other forms, do the same. Stop immediately
183 * on overflow; otherwise %n could produce wrong results. */
184 if (l > INT_MAX - cnt) goto overflow;
185
186 /* Update output count, end loop when fmt is exhausted */
187 cnt += l;
188 if (!*s) break;
189
190 /* Handle literal text and %% format specifiers */
191 for (a=s; *s && *s!='%'; s++);
192 for (z=s; s[0]=='%' && s[1]=='%'; z++, s+=2);
193 if (z-a > INT_MAX-cnt) goto overflow;
194 l = z-a;
195 if (f) out(f, a, l);
196 if (l) continue;
197
198 if (iswdigit(s[1]) && s[2]=='$') {
199 l10n=1;
200 argpos = s[1]-'0';
201 s+=3;
202 } else {
203 argpos = -1;
204 s++;
205 }
206
207 /* Read modifier flags */
208 for (fl=0; (unsigned)*s-' '<32 && (FLAGMASK&(1U<<*s-' ')); s++)
209 fl |= 1U<<*s-' ';
210
211 /* Read field width */
212 if (*s=='*') {
213 if (iswdigit(s[1]) && s[2]=='$') {
214 l10n=1;
215 nl_type[s[1]-'0'] = INT;
216 w = nl_arg[s[1]-'0'].i;
217 s+=3;
218 } else if (!l10n) {
219 w = f ? va_arg(*ap, int) : 0;
220 s++;
221 } else goto inval;
222 if (w<0) fl|=LEFT_ADJ, w=-w;
223 } else if ((w=getint(&s))<0) goto overflow;
224
225 /* Read precision */
226 if (*s=='.' && s[1]=='*') {
227 if (isdigit(s[2]) && s[3]=='$') {
228 nl_type[s[2]-'0'] = INT;
229 p = nl_arg[s[2]-'0'].i;
230 s+=4;
231 } else if (!l10n) {
232 p = f ? va_arg(*ap, int) : 0;
233 s+=2;
234 } else goto inval;
235 xp = (p>=0);
236 } else if (*s=='.') {
237 s++;
238 p = getint(&s);
239 xp = 1;
240 } else {
241 p = -1;
242 xp = 0;
243 }
244
245 /* Format specifier state machine */
246 st=0;
247 do {
248 if (OOB(*s)) goto inval;
249 ps=st;
250 st=states[st]S(*s++);
251 } while (st-1<STOP);
252 if (!st) goto inval;
253
254 /* Check validity of argument type (nl/normal) */
255 if (st==NOARG) {
256 if (argpos>=0) goto inval;
257 } else {
258 if (argpos>=0) nl_type[argpos]=st, arg=nl_arg[argpos];
259 else if (f) pop_arg(&arg, st, ap);
260 else return 0;
261 }
262
263 if (!f) continue;
264 t = s[-1];
265 if (ps && (t&15)==3) t&=~32;
266
267 switch (t) {
268 case 'n':
269 switch(ps) {
270 case BARE: *(int *)arg.p = cnt; break;
271 case LPRE: *(long *)arg.p = cnt; break;
272 case LLPRE: *(long long *)arg.p = cnt; break;
273 case HPRE: *(unsigned short *)arg.p = cnt; break;
274 case HHPRE: *(unsigned char *)arg.p = cnt; break;
275 case ZTPRE: *(size_t *)arg.p = cnt; break;
276 case JPRE: *(uintmax_t *)arg.p = cnt; break;
277 }
278 continue;
279 case 'c':
280 if (w<1) w=1;
281 if (w>1 && !(fl&LEFT_ADJ)) fprintf(f, "%*s", w-1, "");
282 fputwc(btowc(arg.i), f);
283 if (w>1 && (fl&LEFT_ADJ)) fprintf(f, "%*s", w-1, "");
284 l = w;
285 continue;
286 case 'C':
287 fputwc(arg.i, f);
288 l = 1;
289 continue;
290 case 'S':
291 a = arg.p;
292 z = a + wcsnlen(a, p<0 ? INT_MAX : p);
293 if (p<0 && *z) goto overflow;
294 p = z-a;
295 if (w<p) w=p;
296 if (!(fl&LEFT_ADJ)) fprintf(f, "%*s", w-p, "");
297 out(f, a, p);
298 if ((fl&LEFT_ADJ)) fprintf(f, "%*s", w-p, "");
299 l=w;
300 continue;
301 case 'm':
302 arg.p = strerror(errno);
303 case 's':
304 if (!arg.p) arg.p = "(null)";
305 bs = arg.p;
306 for (i=l=0; l<(p<0?INT_MAX:p) && (i=mbtowc(&wc, bs, MB_LEN_MAX))>0; bs+=i, l++);
307 if (i<0) return -1;
308 if (p<0 && *bs) goto overflow;
309 p=l;
310 if (w<p) w=p;
311 if (!(fl&LEFT_ADJ)) fprintf(f, "%*s", w-p, "");
312 bs = arg.p;
313 while (l--) {
314 i=mbtowc(&wc, bs, MB_LEN_MAX);
315 bs+=i;
316 fputwc(wc, f);
317 }
318 if ((fl&LEFT_ADJ)) fprintf(f, "%*s", w-p, "");
319 l=w;
320 continue;
321 }
322
323 if (xp && p<0) goto overflow;
324 #if defined(__wasilibc_printscan_no_long_double)
325 // Omit the 'L' modifier for floating-point cases.
326 switch (t|32) {
327 case 'a': case 'e': case 'f': case 'g':
328 snprintf(charfmt, sizeof charfmt, "%%%s%s%s%s%s*.*%c",
329 "#"+!(fl & ALT_FORM),
330 "+"+!(fl & MARK_POS),
331 "-"+!(fl & LEFT_ADJ),
332 " "+!(fl & PAD_POS),
333 "0"+!(fl & ZERO_PAD),
334 t);
335
336 l = fprintf(f, charfmt, w, p, arg.f);
337 break;
338 case 'd': case 'i': case 'o': case 'u': case 'x': case 'p':
339 snprintf(charfmt, sizeof charfmt, "%%%s%s%s%s%s*.*%c%c",
340 "#"+!(fl & ALT_FORM),
341 "+"+!(fl & MARK_POS),
342 "-"+!(fl & LEFT_ADJ),
343 " "+!(fl & PAD_POS),
344 "0"+!(fl & ZERO_PAD),
345 sizeprefix[(t|32)-'a'], t);
346
347 l = fprintf(f, charfmt, w, p, arg.i);
348 break;
349 }
350 #else
351 snprintf(charfmt, sizeof charfmt, "%%%s%s%s%s%s*.*%c%c",
352 "#"+!(fl & ALT_FORM),
353 "+"+!(fl & MARK_POS),
354 "-"+!(fl & LEFT_ADJ),
355 " "+!(fl & PAD_POS),
356 "0"+!(fl & ZERO_PAD),
357 sizeprefix[(t|32)-'a'], t);
358
359 switch (t|32) {
360 #if !defined(__wasilibc_printscan_no_floating_point)
361 case 'a': case 'e': case 'f': case 'g':
362 l = fprintf(f, charfmt, w, p, arg.f);
363 break;
364 #endif
365 case 'd': case 'i': case 'o': case 'u': case 'x': case 'p':
366 l = fprintf(f, charfmt, w, p, arg.i);
367 break;
368 }
369 #endif
370 }
371
372 if (f) return cnt;
373 if (!l10n) return 0;
374
375 for (i=1; i<=NL_ARGMAX && nl_type[i]; i++)
376 pop_arg(nl_arg+i, nl_type[i], ap);
377 for (; i<=NL_ARGMAX && !nl_type[i]; i++);
378 if (i<=NL_ARGMAX) return -1;
379 return 1;
380
381 inval:
382 errno = EINVAL;
383 return -1;
384 overflow:
385 errno = EOVERFLOW;
386 return -1;
387 }
388
vfwprintf(FILE * restrict f,const wchar_t * restrict fmt,va_list ap)389 int vfwprintf(FILE *restrict f, const wchar_t *restrict fmt, va_list ap)
390 {
391 va_list ap2;
392 int nl_type[NL_ARGMAX] = {0};
393 union arg nl_arg[NL_ARGMAX];
394 int olderr;
395 int ret;
396
397 /* the copy allows passing va_list* even if va_list is an array */
398 va_copy(ap2, ap);
399 if (wprintf_core(0, fmt, &ap2, nl_arg, nl_type) < 0) {
400 va_end(ap2);
401 return -1;
402 }
403
404 FLOCK(f);
405 fwide(f, 1);
406 olderr = f->flags & F_ERR;
407 f->flags &= ~F_ERR;
408 ret = wprintf_core(f, fmt, &ap2, nl_arg, nl_type);
409 if (f->flags & F_ERR) ret = -1;
410 f->flags |= olderr;
411 FUNLOCK(f);
412 va_end(ap2);
413 return ret;
414 }
415