introduce new wide scanf code and remove the last remnants of old scanf

at this point, strto* and all scanf family functions are using the new
unified integer and floating point parser/converter code.

the wide scanf is largely a wrapper for ordinary byte-based scanf;
since numbers can only contain ascii characters, only strings need to
be handled specially.

src/stdio/__scanf.c

@@ -1,483 +0,0 @@
-#include <stdio.h>
-#include <stdlib.h>
-#include <stdarg.h>
-#include <ctype.h>
-#include <wchar.h>
-#include <wctype.h>
-#include <limits.h>
-#include <string.h>
-#include <errno.h>
-#include <math.h>
-#include <float.h>
-
-#include "__scanf.h"
-
-static int read(rctx_t *r)
-{
-	if (--r->w < 0) return r->w = -1;
-	if (r->u) r->u = 0;
-	else r->read(r);
-	return r->c;
-}
-
-static void unread(rctx_t *r)
-{
-	if (r->c < 0 || r->w < 0) return;
-	r->w++;
-	r->u = 1;
-}
-
-#define SIZE_hh -2
-#define SIZE_h  -1
-#define SIZE_def 0
-#define SIZE_l   1
-#define SIZE_ll  2
-#define SIZE_L   3
-
-static void store_int(void *dest, int size, int neg, unsigned long long i)
-{
-	if (!dest) return;
-	if (neg) i = -i;
-	switch (size) {
-	case SIZE_hh:
-		*(char *)dest = i;
-		break;
-	case SIZE_h:
-		*(short *)dest = i;
-		break;
-	case SIZE_def:
-		*(int *)dest = i;
-		break;
-	case SIZE_l:
-		*(long *)dest = i;
-		break;
-	case SIZE_ll:
-		*(long long *)dest = i;
-		break;
-	}
-}
-
-static void *arg_n(va_list ap, unsigned int n)
-{
-	void *p;
-	unsigned int i;
-	va_list ap2;
-	va_copy(ap2, ap);
-	for (i=n; i>1; i--) va_arg(ap2, void *);
-	p = va_arg(ap2, void *);
-	va_end(ap2);
-	return p;
-}
-
-int __scanf(rctx_t *r, const wchar_t *fmt, va_list ap)
-{
-	int mode=0;
-	int width;
-	int size;
-	const wchar_t *p, *z;
-	int c, l, t, m;
-	long long dummy;
-	char *s;
-	wchar_t *wcs;
-	mbstate_t st;
-	int wide = r->wide;
-	void *dest=NULL;
-	int invert;
-	unsigned long long i=0;
-	int neg=0;
-	int matches=0;
-	long double f;
-	int (*is_space)(int) = r->is_space;
-
-	for (p=fmt; *p; ) {
-		if (is_space(*p)) {
-			do p++; while (is_space(*p));
-			do r->w=1; while (is_space(read(r)));
-			unread(r);
-			continue;
-		} else if (*p != '%' || p[1] == '%') {
-			if (*p == '%') p++;
-			r->w = 1;
-			if ((c = read(r)) < 0)
-				goto input_fail;
-			if (*p++ != c)
-				goto match_fail;
-			continue;
-		}
-		p++;
-		if (mode != 1) {
-			for (z=p; isdigit(*z); z++);
-			if (*z != '$' && *z != '*') {
-				if (mode == 0) mode = 1;
-				else goto fmt_fail;
-			} else if (*z != '*') {
-				int pos = 0;
-				mode = 2;
-				for (; p<z; p++) {
-					pos = 10*pos + *p - '0';
-				}
-				p++;
-				if (!pos) goto fmt_fail;
-				dest = arg_n(ap, pos);
-			}
-		}
-		if (*p == '*') {
-			dest = NULL;
-			p++;
-		} else if (mode == 1) {
-			dest = va_arg(ap, void *);
-		}
-		
-		if (!*p) goto fmt_fail;
-
-		width = 0;
-		for (; isdigit(*p); p++) {
-			width = 10*width + *p - '0';
-		}
-
-		size = 0;
-		switch (*p++) {
-		case 0:
-			goto fmt_fail;
-		case 'h':
-			if (*p == 'h') p++, size = SIZE_hh;
-			else size = SIZE_h;
-			break;
-		case 'l':
-			if (*p == 'l') p++, size = SIZE_ll;
-			else size = SIZE_l;
-			break;
-		case 'j':
-			size = SIZE_ll;
-			break;
-		case 'z':
-		case 't':
-			size = SIZE_l;
-			break;
-		case 'L':
-			size = SIZE_L;
-			break;
-		case 'd': case 'i': case 'o': case 'u': case 'x':
-		case 'a': case 'e': case 'f': case 'g':
-		case 'A': case 'E': case 'F': case 'G': case 'X':
-		case 's': case 'c': case '[':
-		case 'S': case 'C':
-		case 'p': case 'n':
-			p--;
-			break;
-		default:
-			goto fmt_fail;
-		}
-
-		t = *p++;
-
-		switch (t) {
-		case 'C':
-		case 'c':
-			if (width < 1) width = 1;
-		case 's':
-			if (size == SIZE_l) t &= ~0x20;
-		case 'd': case 'i': case 'o': case 'u': case 'x':
-		case 'a': case 'e': case 'f': case 'g':
-		case 'A': case 'E': case 'F': case 'G': case 'X':
-		case '[': case 'S':
-		case 'p': case 'n':
-			if (width < 1) width = INT_MAX;
-			break;
-		default:
-			goto fmt_fail;
-		}
-
-		r->w = width;
-
-		if (t != 'n') {
-			if (read(r) < 0) goto input_fail;
-			unread(r);
-		}
-
-		switch (t) {
-		case 'n':
-			store_int(dest, size, 0, r->l - r->u);
-			/* do not increment match count, etc! */
-			continue;
-		case 'C':
-			wcs = dest ? dest : (void *)&dummy;
-			st = (mbstate_t){ 0 };
-			while ((c=read(r)) >= 0) {
-				if (wide) {
-					if (dest) *wcs++ = c;
-				} else {
-					char ch = c;
-					switch (mbrtowc(wcs, &ch, 1, &st)) {
-					case -1:
-						goto enc_fail;
-					case -2:
-						break;
-					default:
-						if (dest) wcs++;
-					}
-				}
-			}
-			if (r->w > 0) goto match_fail;
-			break;
-		case 'c':
-			s = dest ? dest : (void *)&dummy;
-			while ((c=read(r)) >= 0) {
-				if (wide) {
-					if ((l=wctomb(s, c)) < 0)
-						goto enc_fail;
-					if (dest) s += l;
-				} else {
-					if (dest) *s++ = c;
-				}
-			}
-			if (r->w > 0) goto match_fail;
-			break;
-		case '[':
-			wcs = dest ? dest : (void *)&dummy;
-			s = dest ? dest : (void *)&dummy;
-			if (!wide && size == SIZE_l) st = (mbstate_t){ 0 };
-
-			if (*p == '^') p++, invert = 1;
-			else invert = 0;
-
-			if (wide) {
-				for (m=0; (c=read(r)) >= 0; m=1) {
-					for (z=p; *z && *z != c && (*z != ']' || z==p); z++);
-					if (!*z) goto fmt_fail;
-					if (*z == c && (*z != ']' || z==p)) {
-						if (invert) break;
-					} else {
-						if (!invert) break;
-					}
-					if (size == SIZE_l) {
-						if (dest) *wcs++ = c;
-					} else {
-						if ((l=wctomb(s, c)) < 0)
-							goto enc_fail;
-						if (dest) s += l;
-					}
-				}
-				for (p++; *p && *p != ']'; p++);
-				p++;
-			} else {
-				unsigned char scanset[257];
-				memset(scanset, invert, sizeof scanset);
-				scanset[0] = 0;
-				for (z=p; *z && (*z != ']' || z==p); z++)
-					scanset[1+*z] = 1-invert;
-				if (!*z) goto fmt_fail;
-				p=z+1;
-				c=0;
-				for (m=0; scanset[(c=read(r))+1]; m=1) {
-					if (size == SIZE_l) {
-						char ch = c;
-						switch (mbrtowc(wcs, &ch, 1, &st)) {
-						case -1:
-							goto enc_fail;
-						case -2:
-							break;
-						default:
-							if (dest) wcs++;
-						}
-					} else {
-						if (dest) *s++ = c;
-					}
-				}
-			}
-			if (!m) goto match_fail;
-			if (dest) {
-				if (size == SIZE_l) *wcs++ = 0;
-				else *s++ = 0;
-			}
-			break;
-		default:
-			/* read unlimited number of spaces, then reset width */
-			do r->w = 1; while (is_space(c = read(r)));
-			if (c < 0) goto input_fail;
-			unread(r);
-			r->w = width;
-		}
-
-		switch (t) {
-		case 'p':
-		case 'X':
-			t = 'x';
-		case 'd':
-		case 'i':
-		case 'o':
-		case 'u':
-		case 'x':
-			i = m = neg = 0;
-			if ((c=read(r)) == '-') neg=1;
-			else if (c != '+') unread(r);
-			switch (t) {
-			case 'i':
-			case 'x':
-				if ((c=read(r)) != '0') {
-					if (t == 'i') t = 'd';
-					unread(r);
-					break;
-				}
-				m = 1;
-				if (((c=read(r))|0x20) != 'x') {
-					if (t == 'i') t = 'o';
-					unread(r);
-					break;
-				}
-				t = 'x';
-				m = 0;
-			}
-		}
-		
-		switch (t) {
-		case 'd':
-		case 'u':
-			for (; isdigit(c=read(r)); m=1)
-				i = 10*i + c-'0';
-			goto int_finish;
-		case 'o':
-			for (; (unsigned)(c=read(r))-'0' < 8; m=1)
-				i = (i<<3) + c-'0';
-			goto int_finish;
-		case 'x':
-			for (; ; m=1) {
-				if (isdigit(c=read(r))) {
-					i = (i<<4) + c-'0';
-				} else if ((unsigned)(c|0x20)-'a' < 6) {
-					i = (i<<4) + (c|0x20)-'a'+10;
-				} else break;
-			}
-		int_finish:
-			if (!m) goto match_fail;
-			store_int(dest, size, neg, i);
-			break;
-		case 'a':
-		case 'e':
-		case 'f':
-		case 'g':
-			f = 0.0;
-			neg = m = 0;
-			if ((c=read(r)) == '-') neg=1;
-			else if (c != '+') unread(r);
-			/* FIXME: check for INF/NAN strings here */
-			if (read(r)=='0' && (m=1, (read(r)|0x20) == 'x'))
-				goto hexfloat;
-			else unread(r);
-			for (; isdigit(c=read(r)); m=1)
-				f = 10.0 * f + (c-'0');
-			if (c=='.') {
-				double mag = 10.0;
-				for (; isdigit(c=read(r)); mag*=10.0)
-					f += (c-'0')/mag;
-			}
-			if ((c|0x20)=='e') {
-				int ex=0, en=0;
-				m = 0;
-				if ((c=read(r))=='-') en=1;
-				else if (c!='+') unread(r);
-				for (; isdigit(c=read(r)); m=1)
-					if (ex < LDBL_MAX_10_EXP)
-						ex = 10 * ex + (c-'0');
-				if (ex > LDBL_MAX_10_EXP)
-					f = en ? 0 : INFINITY;
-				else {
-					if (en) while (ex--) f/=10.0;
-					else while (ex--) f*=10.0;
-				}
-			}
-			goto writefloat;
-hexfloat:
-			m = 0;
-			for (; isxdigit(c=read(r)); m=1)
-				if (isdigit(c)) f = 16.0*f + (c-'0');
-				else f = 16.0*f + ((c|32)-'a'+10);
-			if (c=='.') {
-				double mag = 1/16.0;
-				for (; isxdigit(c=read(r)); mag*=1/16.0)
-					if (isdigit(c)) f += (c-'0')*mag;
-					else f += ((c|32)-'a'+10)*mag;
-			}
-			if ((c|0x20)=='p') {
-				int ex=0, en=0;
-				m = 0;
-				if ((c=read(r))=='-') en=1;
-				else if (c!='+') unread(r);
-				for (; isdigit(c=read(r)); m=1)
-					if (ex < LDBL_MAX_EXP)
-						ex = 10 * ex + (c-'0');
-				if (ex > LDBL_MAX_EXP)
-					f = en ? 0 : INFINITY;
-				else {
-					if (en) while (ex--) f*=0.5;
-					else while (ex--) f*=2.0;
-				}
-			}
-writefloat:
-			if (!m) goto match_fail;
-			if (neg) f *= -1.0;
-			if (dest) switch (size) {
-			case SIZE_def:
-				*(float *)dest = f;
-				break;
-			case SIZE_l:
-				*(double *)dest = f;
-				break;
-			case SIZE_L:
-				*(long double *)dest = f;
-				break;
-			}
-			break;
-		case 'S':
-			wcs = dest ? dest : (void *)&dummy;
-			st = (mbstate_t){ 0 };
-			while((c=read(r)) >= 0) {
-				if (wide) {
-					if (is_space(c)) break;
-					if (dest) *wcs++ = c;
-				} else {
-					char ch = c;
-					if (is_space(c)) break;
-					switch (mbrtowc(wcs, &ch, 1, &st)) {
-					case -1:
-						goto enc_fail;
-					case -2:
-						break;
-					default:
-						if (dest) wcs++;
-					}
-				}
-			}
-			if (dest) *wcs++ = 0;
-			break;
-		case 's':
-			s = dest ? dest : (void *)&dummy;
-			while((c=read(r)) >= 0) {
-				if (wide) {
-					if (is_space(c)) break;
-					if ((l=wctomb(s, c)) < 0)
-						goto enc_fail;
-					if (dest) s += l;
-				} else {
-					if (is_space(c)) break;
-					if (dest) *s++ = c;
-				}
-			}
-			if (dest) *s++ = 0;
-			break;
-		}
-
-		/* unread will do nothing if field width was exhausted */
-		unread(r);
-		if (dest) matches++;
-	}
-	return matches;
-enc_fail:
-	errno = EILSEQ;
-fmt_fail:
-input_fail:
-	if (!matches) matches--;
-match_fail:
-	unread(r);
-	return matches;
-}

src/stdio/__scanf.h

@@ -1,16 +0,0 @@
-#include <wchar.h>
-
-typedef struct rctx
-{
-	void (*read)(struct rctx *);
-	void *opaque;
-	int wide;
-	int (*is_space)();
-	int l;
-	int e;
-	int c;
-	int u;
-	int w;
-} rctx_t;
-
-int __scanf(rctx_t *, const wchar_t *, va_list);

src/stdio/vfwscanf.c

@@ -1,28 +1,299 @@
 #include <stdio.h>
-#include <string.h>
-#include <errno.h>
+#include <stdlib.h>
+#include <stdarg.h>
+#include <ctype.h>
 #include <wchar.h>
 #include <wctype.h>
+#include <limits.h>
+#include <string.h>
+#include <errno.h>
+#include <math.h>
+#include <float.h>
 
 #include "stdio_impl.h"
-#include "__scanf.h"
+#include "shgetc.h"
+#include "intscan.h"
+#include "floatscan.h"
+
+#define SIZE_hh -2
+#define SIZE_h  -1
+#define SIZE_def 0
+#define SIZE_l   1
+#define SIZE_L   2
+#define SIZE_ll  3
+
+static void store_int(void *dest, int size, unsigned long long i)
+{
+	if (!dest) return;
+	switch (size) {
+	case SIZE_hh:
+		*(char *)dest = i;
+		break;
+	case SIZE_h:
+		*(short *)dest = i;
+		break;
+	case SIZE_def:
+		*(int *)dest = i;
+		break;
+	case SIZE_l:
+		*(long *)dest = i;
+		break;
+	case SIZE_ll:
+		*(long long *)dest = i;
+		break;
+	}
+}
+
+static void *arg_n(va_list ap, unsigned int n)
+{
+	void *p;
+	unsigned int i;
+	va_list ap2;
+	va_copy(ap2, ap);
+	for (i=n; i>1; i--) va_arg(ap2, void *);
+	p = va_arg(ap2, void *);
+	va_end(ap2);
+	return p;
+}
 
-static void f_read(rctx_t *r)
+static int in_set(const wchar_t *set, int c)
 {
-	FILE *f = r->opaque;
-	if ((r->c = fgetwc(f)) >= 0) r->l++;
+	int j;
+	const wchar_t *p = set;
+	if (*p == '-') {
+		if (c=='-') return 1;
+		p++;
+	} else if (*p == ']') {
+		if (c==']') return 1;
+		p++;
+	}
+	for (; *p && *p != ']'; p++) {
+		if (*p=='-' && p[1] && p[1] != ']')
+			for (j=p++[-1]; j<*p; j++)
+				if (c==j) return 1;
+		if (c==*p) return 1;
+	}
+	return 0;
 }
 
+#if 1
+#undef getwc
+#define getwc(f) \
+	((f)->rpos < (f)->rend && *(f)->rpos < 128 ? *(f)->rpos++ : (getwc)(f))
+
+#undef ungetwc
+#define ungetwc(c,f) \
+	((f)->rend && (c)<128 ? *--(f)->rpos : ungetwc((c),(f)))
+#endif
+
 int vfwscanf(FILE *f, const wchar_t *fmt, va_list ap)
 {
-	rctx_t r = { f_read, (void *)f, 1, iswspace };
-	int result;
+	int width;
+	int size;
+	int alloc;
+	const wchar_t *p;
+	int c, t;
+	char *s;
+	wchar_t *wcs;
+	void *dest=NULL;
+	int invert;
+	int matches=0;
+	off_t pos = 0, cnt;
+	static const char size_pfx[][3] = { "hh", "h", "", "l", "L", "ll" };
+	char tmp[3*sizeof(int)+10];
 
-	result = __scanf(&r, fmt, ap);
+	FLOCK(f);
 
-	if (r.u && r.c >= 0) {
-		ungetwc(r.c, f);
-	}
+	for (p=fmt; *p; p++) {
+
+		if (iswspace(*p)) {
+			while (iswspace(p[1])) p++;
+			while (iswspace((c=getwc(f)))) pos++;
+			ungetwc(c, f);
+			continue;
+		}
+		if (*p != '%' || p[1] == '%') {
+			p += *p=='%';
+			c = getwc(f);
+			if (c!=*p) {
+				ungetwc(c, f);
+				if (c<0) goto input_fail;
+				goto match_fail;
+			}
+			pos++;
+			continue;
+		}
+
+		p++;
+		if (*p=='*') {
+			dest = 0; p++;
+		} else if (iswdigit(*p) && p[1]=='$') {
+			dest = arg_n(ap, *p-'0'); p+=2;
+		} else {
+			dest = va_arg(ap, void *);
+		}
+
+		for (width=0; iswdigit(*p); p++) {
+			width = 10*width + *p - '0';
+		}
 
-	return result;
+		if (*p=='m') {
+			alloc = 1;
+			p++;
+		} else {
+			alloc = 0;
+		}
+
+		size = SIZE_def;
+		switch (*p++) {
+		case 'h':
+			if (*p == 'h') p++, size = SIZE_hh;
+			else size = SIZE_h;
+			break;
+		case 'l':
+			if (*p == 'l') p++, size = SIZE_ll;
+			else size = SIZE_l;
+			break;
+		case 'j':
+			size = SIZE_ll;
+			break;
+		case 'z':
+		case 't':
+			size = SIZE_l;
+			break;
+		case 'L':
+			size = SIZE_L;
+			break;
+		case 'd': case 'i': case 'o': case 'u': case 'x':
+		case 'a': case 'e': case 'f': case 'g':
+		case 'A': case 'E': case 'F': case 'G': case 'X':
+		case 's': case 'c': case '[':
+		case 'S': case 'C':
+		case 'p': case 'n':
+			p--;
+			break;
+		default:
+			goto fmt_fail;
+		}
+
+		t = *p;
+
+		/* Transform ls,lc -> S,C */
+		if (size==SIZE_l && (t&15)==3) t&=~32;
+
+		if (t != 'n' && t != '[' && (t|32) != 'c') {
+			while (iswspace((c=getwc(f)))) pos++;
+			if (c < 0) goto input_fail;
+			ungetwc(c, f);
+		}
+
+		switch (t) {
+		case 'n':
+			store_int(dest, size, pos);
+			/* do not increment match count, etc! */
+			continue;
+
+		case 'c':
+			if (width < 1) width = 1;
+			s = dest;
+			for (; width && (c=getwc(f)) >= 0; width--) {
+				int l = wctomb(s?s:tmp, c);
+				if (l<0) goto input_fail;
+				if (s) s+=l;
+				pos++;
+			}
+			if (width) goto match_fail;
+			break;
+
+		case 'C':
+			if (width < 1) width = 1;
+			wcs = dest;
+			for (; width && (c=getwc(f)) >= 0; width--)
+				pos++, wcs && (*wcs++ = c);
+			if (width) goto match_fail;
+			break;
+
+		case 's':
+			s = dest;
+			while (!iswspace(c=getwc(f)) && c!=EOF) {
+				int l = wctomb(s?s:tmp, c);
+				if (l<0) goto input_fail;
+				if (s) s+=l;
+				pos++;
+			}
+			if (s) *s = 0;
+			break;
+
+		case 'S':
+			wcs = dest;
+			while (!iswspace(c=getwc(f)) && c!=EOF)
+				pos++, *wcs++ = c;
+			if (wcs) *wcs = 0;
+			break;
+
+		case '[':
+			s = (size == SIZE_def) ? dest : 0;
+			wcs = (size == SIZE_l) ? dest : 0;
+
+			if (*++p == '^') p++, invert = 1;
+			else invert = 0;
+
+			int gotmatch = 0;
+
+			for (;;) {
+				if ((c=getwc(f))<0) break;
+				if (in_set(p, c) == invert)
+					break;
+				if (wcs) {
+					*wcs++ = c;
+				} else if (size != SIZE_l) {
+					int l = wctomb(s?s:tmp, c);
+					if (l<0) goto input_fail;
+					if (s) s+=l;
+				}
+				pos++;
+				gotmatch=1;
+			}
+			ungetwc(c, f);
+
+			if (!gotmatch) goto match_fail;
+
+			if (*p==']') p++;
+			while (*p!=']') {
+				if (!*p) goto fmt_fail;
+				p++;
+			}
+
+			if (wcs) *wcs++ = 0;
+			if (s) *s++ = 0;
+			break;
+
+		case 'd': case 'i': case 'o': case 'u': case 'x':
+		case 'a': case 'e': case 'f': case 'g':
+		case 'A': case 'E': case 'F': case 'G': case 'X':
+		case 'p':
+			if (width < 1) width = 0;
+			snprintf(tmp, sizeof tmp, "%.*s%.0d%s%c%%lln",
+				1+!dest, "%*", width, size_pfx[size+2], t);
+			cnt = 0;
+			if (fscanf(f, tmp, dest?dest:&cnt, &cnt) == -1)
+				goto input_fail;
+			else if (!cnt)
+				goto match_fail;
+			pos += cnt;
+			break;
+		default:
+			goto fmt_fail;
+		}
+
+		if (dest) matches++;
+	}
+	if (0) {
+fmt_fail:
+input_fail:
+		if (!matches) matches--;
+	}
+match_fail:
+	FUNLOCK(f);
+	return matches;
 }

src/stdio/vswscanf.c

@@ -1,19 +1,35 @@
-#include <stdio.h>
-#include <string.h>
-#include <wchar.h>
-#include <wctype.h>
+#include "stdio_impl.h"
 
-#include "__scanf.h"
-
-static void s_read(rctx_t *r)
+static size_t wstring_read(FILE *f, unsigned char *buf, size_t len)
 {
-	wchar_t *s = r->opaque;
-	if (!s[r->l]) r->c = -1;
-	else r->c = s[r->l++];
+	const wchar_t *src = f->cookie;
+	size_t k;
+
+	if (!src) return 0;
+
+	k = wcsrtombs((void *)f->buf, &src, f->buf_size, 0);
+	if (k==(size_t)-1) {
+		f->rpos = f->rend = 0;
+		return 0;
+	}
+
+	f->rpos = f->buf;
+	f->rend = f->buf + k;
+	f->cookie = (void *)src;
+
+	if (!len) return 0;
+
+	*buf = *f->rpos++;
+	return 1;
 }
 
 int vswscanf(const wchar_t *s, const wchar_t *fmt, va_list ap)
 {
-	rctx_t r = { s_read, (void *)s, 1, iswspace };
-	return __scanf(&r, fmt, ap);
+	unsigned char buf[256];
+	FILE f = {
+		.buf = buf, .buf_size = sizeof buf,
+		.cookie = (void *)s,
+		.read = wstring_read, .lock = -1
+	};
+	return vfwscanf(&f, fmt, ap);
 }