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
31enum {
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
46static 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') = INT, S('C') = UINT,
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') = UINT, 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
99union 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
112static 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
145static void out(FILE *f, const wchar_t *s, size_t l)
146{
147 while (l-- && !ferror(f)) fputwc(*s++, f);
148}
149
150static void pad(FILE *f, int n, int fl)
151{
152 if ((fl & LEFT_ADJ) || !n || ferror(f)) return;
153 fprintf(f, "%*s", n, "");
154}
155
156static int getint(wchar_t **s) {
157 int i;
158 for (i=0; iswdigit(**s); (*s)++) {
159 if (i > INT_MAX/10U || **s-'0' > INT_MAX-10*i) i = -1;
160 else i = 10*i + (**s-'0');
161 }
162 return i;
163}
164
165static const char sizeprefix['y'-'a'] = {
166['a'-'a']='L', ['e'-'a']='L', ['f'-'a']='L', ['g'-'a']='L',
167['d'-'a']='j', ['i'-'a']='j', ['o'-'a']='j', ['u'-'a']='j', ['x'-'a']='j',
168['p'-'a']='j'
169};
170
171static int wprintf_core(FILE *f, const wchar_t *fmt, va_list *ap, union arg *nl_arg, int *nl_type)
172{
173 wchar_t *a, *z, *s=(wchar_t *)fmt;
174 unsigned l10n=0, fl;
175 int w, p, xp;
176 union arg arg;
177 int argpos;
178 unsigned st, ps;
179 int cnt=0, l=0;
180 int i;
181 int t;
182 char *bs;
183 char charfmt[16];
184 wchar_t wc;
185
186 for (;;) {
187 /* This error is only specified for snprintf, but since it's
188 * unspecified for other forms, do the same. Stop immediately
189 * on overflow; otherwise %n could produce wrong results. */
190 if (l > INT_MAX - cnt) goto overflow;
191
192 /* Update output count, end loop when fmt is exhausted */
193 cnt += l;
194 if (!*s) break;
195
196 /* Handle literal text and %% format specifiers */
197 for (a=s; *s && *s!='%'; s++);
198 for (z=s; s[0]=='%' && s[1]=='%'; z++, s+=2);
199 if (z-a > INT_MAX-cnt) goto overflow;
200 l = z-a;
201 if (f) out(f, a, l);
202 if (l) continue;
203
204 if (iswdigit(s[1]) && s[2]=='$') {
205 l10n=1;
206 argpos = s[1]-'0';
207 s+=3;
208 } else {
209 argpos = -1;
210 s++;
211 }
212
213 /* Read modifier flags */
214 for (fl=0; (unsigned)*s-' '<32 && (FLAGMASK&(1U<<*s-' ')); s++)
215 fl |= 1U<<*s-' ';
216
217 /* Read field width */
218 if (*s=='*') {
219 if (iswdigit(s[1]) && s[2]=='$') {
220 l10n=1;
221 nl_type[s[1]-'0'] = INT;
222 w = nl_arg[s[1]-'0'].i;
223 s+=3;
224 } else if (!l10n) {
225 w = f ? va_arg(*ap, int) : 0;
226 s++;
227 } else goto inval;
228 if (w<0) fl|=LEFT_ADJ, w=-w;
229 } else if ((w=getint(&s))<0) goto overflow;
230
231 /* Read precision */
232 if (*s=='.' && s[1]=='*') {
233 if (isdigit(s[2]) && s[3]=='$') {
234 nl_type[s[2]-'0'] = INT;
235 p = nl_arg[s[2]-'0'].i;
236 s+=4;
237 } else if (!l10n) {
238 p = f ? va_arg(*ap, int) : 0;
239 s+=2;
240 } else goto inval;
241 xp = (p>=0);
242 } else if (*s=='.') {
243 s++;
244 p = getint(&s);
245 xp = 1;
246 } else {
247 p = -1;
248 xp = 0;
249 }
250
251 /* Format specifier state machine */
252 st=0;
253 do {
254 if (OOB(*s)) goto inval;
255 ps=st;
256 st=states[st]S(*s++);
257 } while (st-1<STOP);
258 if (!st) goto inval;
259
260 /* Check validity of argument type (nl/normal) */
261 if (st==NOARG) {
262 if (argpos>=0) goto inval;
263 } else {
264 if (argpos>=0) nl_type[argpos]=st, arg=nl_arg[argpos];
265 else if (f) pop_arg(&arg, st, ap);
266 else return 0;
267 }
268
269 if (!f) continue;
270
271 /* Do not process any new directives once in error state. */
272 if (ferror(f)) return -1;
273
274 t = s[-1];
275 if (ps && (t&15)==3) t&=~32;
276
277 switch (t) {
278 case 'n':
279 switch(ps) {
280 case BARE: *(int *)arg.p = cnt; break;
281 case LPRE: *(long *)arg.p = cnt; break;
282 case LLPRE: *(long long *)arg.p = cnt; break;
283 case HPRE: *(unsigned short *)arg.p = cnt; break;
284 case HHPRE: *(unsigned char *)arg.p = cnt; break;
285 case ZTPRE: *(size_t *)arg.p = cnt; break;
286 case JPRE: *(uintmax_t *)arg.p = cnt; break;
287 }
288 continue;
289 case 'c':
290 case 'C':
291 if (w<1) w=1;
292 pad(f, w-1, fl);
293 out(f, &(wchar_t){t=='C' ? arg.i : btowc(arg.i)}, 1);
294 pad(f, w-1, fl^LEFT_ADJ);
295 l = w;
296 continue;
297 case 'S':
298 a = arg.p;
299 z = a + wcsnlen(a, p<0 ? INT_MAX : p);
300 if (p<0 && *z) goto overflow;
301 p = z-a;
302 if (w<p) w=p;
303 pad(f, w-p, fl);
304 out(f, a, p);
305 pad(f, w-p, fl^LEFT_ADJ);
306 l=w;
307 continue;
308 case 'm':
309 arg.p = strerror(errno);
310 case 's':
311 if (!arg.p) arg.p = "(null)";
312 bs = arg.p;
313 for (i=l=0; l<(p<0?INT_MAX:p) && (i=mbtowc(&wc, bs, MB_LEN_MAX))>0; bs+=i, l++);
314 if (i<0) return -1;
315 if (p<0 && *bs) goto overflow;
316 p=l;
317 if (w<p) w=p;
318 pad(f, w-p, fl);
319 bs = arg.p;
320 while (l--) {
321 i=mbtowc(&wc, bs, MB_LEN_MAX);
322 bs+=i;
323 out(f, &wc, 1);
324 }
325 pad(f, w-p, fl^LEFT_ADJ);
326 l=w;
327 continue;
328 }
329
330 if (xp && p<0) goto overflow;
331#if defined(__wasilibc_printscan_no_long_double)
332 // Omit the 'L' modifier for floating-point cases.
333 switch (t|32) {
334 case 'a': case 'e': case 'f': case 'g':
335 snprintf(charfmt, sizeof charfmt, "%%%s%s%s%s%s*.*%c",
336 "#"+!(fl & ALT_FORM),
337 "+"+!(fl & MARK_POS),
338 "-"+!(fl & LEFT_ADJ),
339 " "+!(fl & PAD_POS),
340 "0"+!(fl & ZERO_PAD),
341 t);
342
343 l = fprintf(f, charfmt, w, p, arg.f);
344 break;
345 case 'd': case 'i': case 'o': case 'u': case 'x': case 'p':
346 snprintf(charfmt, sizeof charfmt, "%%%s%s%s%s%s*.*%c%c",
347 "#"+!(fl & ALT_FORM),
348 "+"+!(fl & MARK_POS),
349 "-"+!(fl & LEFT_ADJ),
350 " "+!(fl & PAD_POS),
351 "0"+!(fl & ZERO_PAD),
352 sizeprefix[(t|32)-'a'], t);
353
354 l = fprintf(f, charfmt, w, p, arg.i);
355 break;
356 }
357#else
358 snprintf(charfmt, sizeof charfmt, "%%%s%s%s%s%s*.*%c%c",
359 "#"+!(fl & ALT_FORM),
360 "+"+!(fl & MARK_POS),
361 "-"+!(fl & LEFT_ADJ),
362 " "+!(fl & PAD_POS),
363 "0"+!(fl & ZERO_PAD),
364 sizeprefix[(t|32)-'a'], t);
365
366 switch (t|32) {
367#if !defined(__wasilibc_printscan_no_floating_point)
368 case 'a': case 'e': case 'f': case 'g':
369 l = fprintf(f, charfmt, w, p, arg.f);
370 break;
371#endif
372 case 'd': case 'i': case 'o': case 'u': case 'x': case 'p':
373 l = fprintf(f, charfmt, w, p, arg.i);
374 break;
375 }
376#endif
377 }
378
379 if (f) return cnt;
380 if (!l10n) return 0;
381
382 for (i=1; i<=NL_ARGMAX && nl_type[i]; i++)
383 pop_arg(nl_arg+i, nl_type[i], ap);
384 for (; i<=NL_ARGMAX && !nl_type[i]; i++);
385 if (i<=NL_ARGMAX) return -1;
386 return 1;
387
388inval:
389 errno = EINVAL;
390 return -1;
391overflow:
392 errno = EOVERFLOW;
393 return -1;
394}
395
396int vfwprintf(FILE *restrict f, const wchar_t *restrict fmt, va_list ap)
397{
398 va_list ap2;
399 int nl_type[NL_ARGMAX+1] = {0};
400 union arg nl_arg[NL_ARGMAX+1];
401 int olderr;
402 int ret;
403
404 /* the copy allows passing va_list* even if va_list is an array */
405 va_copy(ap2, ap);
406 if (wprintf_core(0, fmt, &ap2, nl_arg, nl_type) < 0) {
407 va_end(ap2);
408 return -1;
409 }
410
411 FLOCK(f);
412 fwide(f, 1);
413 olderr = f->flags & F_ERR;
414 f->flags &= ~F_ERR;
415 ret = wprintf_core(f, fmt, &ap2, nl_arg, nl_type);
416 if (ferror(f)) ret = -1;
417 f->flags |= olderr;
418 FUNLOCK(f);
419 va_end(ap2);
420 return ret;
421}