date 3.2.1 → 3.5.1

data/ext/date/date_parse.c

@@ -7,6 +7,9 @@
 #include "ruby/re.h"
 #include <ctype.h>
 
+#undef strncasecmp
+#define strncasecmp STRNCASECMP
+
 RUBY_EXTERN VALUE rb_int_positive_pow(long x, unsigned long y);
 RUBY_EXTERN unsigned long ruby_scan_digits(const char *str, ssize_t len, int base, size_t *retlen, int *overflow);
 
@@ -253,6 +256,8 @@ s3e(VALUE hash, VALUE y, VALUE m, VALUE d, int bc)
 #define ABBR_DAYS "sun|mon|tue|wed|thu|fri|sat"
 #define ABBR_MONTHS "jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec"
 
+#define NUMBER "(?<!\\d)\\d"
+
 #ifdef TIGHT_PARSER
 #define VALID_DAYS "(?:" DAYS ")" "|(?:tues|wednes|thurs|thur|" ABBR_DAYS ")\\.?"
 #define VALID_MONTHS "(?:" MONTHS ")" "|(?:sept|" ABBR_MONTHS ")\\.?"
@@ -411,7 +416,6 @@ VALUE
 date_zone_to_diff(VALUE str)
 {
     VALUE offset = Qnil;
-    VALUE vbuf = 0;
     long l = RSTRING_LEN(str);
     const char *s = RSTRING_PTR(str);
 
@@ -437,16 +441,26 @@ date_zone_to_diff(VALUE str)
 	    l -= w;
 	    dst = 1;
 	}
+
 	{
+	    const char *zn = s;
 	    long sl = shrunk_size(s, l);
+	    char shrunk_buff[MAX_WORD_LENGTH]; /* no terminator to be added */
+	    const struct zone *z = 0;
+
+	    if (sl <= 0) {
+		sl = l;
+	    }
+	    else if (sl <= MAX_WORD_LENGTH) {
+		char *d = shrunk_buff;
+		sl = shrink_space(d, s, l);
+		zn = d;
+	    }
+
 	    if (sl > 0 && sl <= MAX_WORD_LENGTH) {
-		char *d = ALLOCV_N(char, vbuf, sl);
-		l = shrink_space(d, s, l);
-		s = d;
+		z = zonetab(zn, (unsigned int)sl);
 	    }
-	}
-	if (l > 0 && l <= MAX_WORD_LENGTH) {
-	    const struct zone *z = zonetab(s, (unsigned int)l);
+
 	    if (z) {
 		int d = z->offset;
 		if (dst)
@@ -455,6 +469,7 @@ date_zone_to_diff(VALUE str)
 		goto ok;
 	    }
 	}
+
 	{
 	    char *p;
 	    int sign = 0;
@@ -471,27 +486,53 @@ date_zone_to_diff(VALUE str)
 		s++;
 		l--;
 
+#define out_of_range(v, min, max) ((v) < (min) || (max) < (v))
 		hour = STRTOUL(s, &p, 10);
 		if (*p == ':') {
+		    if (out_of_range(hour, 0, 23)) return Qnil;
 		    s = ++p;
 		    min = STRTOUL(s, &p, 10);
+		    if (out_of_range(min, 0, 59)) return Qnil;
 		    if (*p == ':') {
 			s = ++p;
 			sec = STRTOUL(s, &p, 10);
+			if (out_of_range(sec, 0, 59)) return Qnil;
 		    }
-		    goto num;
 		}
-		if (*p == ',' || *p == '.') {
-		    char *e = 0;
-		    p++;
-		    min = STRTOUL(p, &e, 10) * 3600;
+		else if (*p == ',' || *p == '.') {
+		    /* fractional hour */
+		    size_t n;
+		    int ov;
+		    /* no over precision for offset; 10**-7 hour = 0.36
+		     * milliseconds should be enough. */
+		    const size_t max_digits = 7; /* 36 * 10**7 < 32-bit FIXNUM_MAX */
+
+		    if (out_of_range(hour, 0, 23)) return Qnil;
+
+		    n = (s + l) - ++p;
+		    if (n > max_digits) n = max_digits;
+		    sec = ruby_scan_digits(p, n, 10, &n, &ov);
+		    if ((p += n) < s + l && *p >= ('5' + !(sec & 1)) && *p <= '9') {
+			/* round half to even */
+			sec++;
+		    }
+		    sec *= 36;
 		    if (sign) {
 			hour = -hour;
-			min = -min;
+			sec = -sec;
+		    }
+		    if (n <= 2) {
+			/* HH.nn or HH.n */
+			if (n == 1) sec *= 10;
+			offset = INT2FIX(sec + hour * 3600);
+		    }
+		    else {
+			VALUE denom = rb_int_positive_pow(10, (int)(n - 2));
+			offset = f_add(rb_rational_new(INT2FIX(sec), denom), INT2FIX(hour * 3600));
+			if (rb_rational_den(offset) == INT2FIX(1)) {
+			    offset = rb_rational_num(offset);
+			}
 		    }
-		    offset = rb_rational_new(INT2FIX(min),
-					     rb_int_positive_pow(10, (int)(e - p)));
-		    offset = f_add(INT2FIX(hour * 3600), offset);
 		    goto ok;
 		}
 		else if (l > 2) {
@@ -504,18 +545,16 @@ date_zone_to_diff(VALUE str)
 			min  = ruby_scan_digits(&s[2 - l % 2], 2, 10, &n, &ov);
 		    if (l >= 5)
 			sec  = ruby_scan_digits(&s[4 - l % 2], 2, 10, &n, &ov);
-		    goto num;
 		}
-	      num:
 		sec += min * 60 + hour * 3600;
 		if (sign) sec = -sec;
 		offset = INT2FIX(sec);
+#undef out_of_range
 	    }
 	}
     }
     RB_GC_GUARD(str);
   ok:
-    ALLOCV_END(vbuf);
     return offset;
 }
 
@@ -652,24 +691,27 @@ parse_time(VALUE str, VALUE hash)
 {
     static const char pat_source[] =
 		"("
+		   "" NUMBER "+\\s*"
 		   "(?:"
-		     "\\d+\\s*:\\s*\\d+"
 		     "(?:"
+		       ":\\s*\\d+"
+		       "(?:"
 #ifndef TIGHT_PARSER
-		       "\\s*:\\s*\\d+(?:[,.]\\d*)?"
+		         "\\s*:\\s*\\d+(?:[,.]\\d*)?"
 #else
-		       "\\s*:\\s*\\d+(?:[,.]\\d+)?"
+		         "\\s*:\\s*\\d+(?:[,.]\\d+)?"
 #endif
+		       ")?"
+		     "|"
+		       "h(?:\\s*\\d+m?(?:\\s*\\d+s?)?)?"
+		     ")"
+		     "(?:"
+		       "\\s*"
+		       "[ap](?:m\\b|\\.m\\.)"
 		     ")?"
 		   "|"
-		     "\\d+\\s*h(?:\\s*\\d+m?(?:\\s*\\d+s?)?)?"
-		   ")"
-		   "(?:"
-		     "\\s*"
 		     "[ap](?:m\\b|\\.m\\.)"
-		   ")?"
-		 "|"
-		   "\\d+\\s*[ap](?:m\\b|\\.m\\.)"
+		   ")"
 		 ")"
 		 "(?:"
 		   "\\s*"
@@ -691,6 +733,9 @@ parse_time(VALUE str, VALUE hash)
 #endif
 }
 
+#define BEGIN_ERA "\\b"
+#define END_ERA "(?!(?<!\\.)[a-z])"
+
 #ifdef TIGHT_PARSER
 static int
 parse_era1_cb(VALUE m, VALUE hash)
@@ -702,7 +747,7 @@ static int
 parse_era1(VALUE str, VALUE hash)
 {
     static const char pat_source[] =
-	"(a(?:d|\\.d\\.))";
+	BEGIN_ERA "(a(?:d\\b|\\.d\\.))" END_ERA;
     static VALUE pat = Qnil;
 
     REGCOMP_I(pat);
@@ -724,8 +769,9 @@ parse_era2_cb(VALUE m, VALUE hash)
 static int
 parse_era2(VALUE str, VALUE hash)
 {
-    static const char pat_source[] =
-	"(c(?:e|\\.e\\.)|b(?:ce|\\.c\\.e\\.)|b(?:c|\\.c\\.))";
+    static const char pat_source[] = BEGIN_ERA
+	"(c(?:e\\b|\\.e\\.)|b(?:ce\\b|\\.c\\.e\\.)|b(?:c\\b|\\.c\\.))"
+	END_ERA;
     static VALUE pat = Qnil;
 
     REGCOMP_I(pat);
@@ -829,7 +875,7 @@ parse_eu(VALUE str, VALUE hash)
 		FPW_COM FPT_COM
 #endif
 #ifndef TIGHT_PARSER
-		"('?\\d+)[^-\\d\\s]*"
+		"('?" NUMBER "+)[^-\\d\\s]*"
 #else
 		"(\\d+)(?:(?:st|nd|rd|th)\\b)?"
 #endif
@@ -842,7 +888,11 @@ parse_eu(VALUE str, VALUE hash)
 		 "(?:"
 		   "\\s*"
 #ifndef TIGHT_PARSER
-		   "(c(?:e|\\.e\\.)|b(?:ce|\\.c\\.e\\.)|a(?:d|\\.d\\.)|b(?:c|\\.c\\.))?"
+		   "(?:"
+		     BEGIN_ERA
+		     "(c(?:e|\\.e\\.)|b(?:ce|\\.c\\.e\\.)|a(?:d|\\.d\\.)|b(?:c|\\.c\\.))"
+		     END_ERA
+		   ")?"
 		   "\\s*"
 		   "('?-?\\d+(?:(?:st|nd|rd|th)\\b)?)"
 #else
@@ -919,8 +969,8 @@ parse_us(VALUE str, VALUE hash)
 		COM_FPT
 #endif
 		 "(?:"
-		   "\\s*,?"
-		   "\\s*"
+		   "\\s*+,?"
+		   "\\s*+"
 #ifndef TIGHT_PARSER
 		   "(c(?:e|\\.e\\.)|b(?:ce|\\.c\\.e\\.)|a(?:d|\\.d\\.)|b(?:c|\\.c\\.))?"
 		   "\\s*"
@@ -967,7 +1017,7 @@ parse_iso(VALUE str, VALUE hash)
 {
     static const char pat_source[] =
 #ifndef TIGHT_PARSER
-	"('?[-+]?\\d+)-(\\d+)-('?-?\\d+)"
+	"('?[-+]?" NUMBER "+)-(\\d+)-('?-?\\d+)"
 #else
 	BOS
 	FPW_COM FPT_COM
@@ -1321,7 +1371,7 @@ parse_vms11(VALUE str, VALUE hash)
 {
     static const char pat_source[] =
 #ifndef TIGHT_PARSER
-	"('?-?\\d+)-(" ABBR_MONTHS ")[^-/.]*"
+	"('?-?" NUMBER "+)-(" ABBR_MONTHS ")[^-/.]*"
 	"-('?-?\\d+)"
 #else
 	BOS
@@ -1416,7 +1466,7 @@ parse_sla(VALUE str, VALUE hash)
 {
     static const char pat_source[] =
 #ifndef TIGHT_PARSER
-	"('?-?\\d+)/\\s*('?\\d+)(?:\\D\\s*('?-?\\d+))?"
+	"('?-?" NUMBER "+)/\\s*('?\\d+)(?:\\D\\s*('?-?\\d+))?"
 #else
 	BOS
 	FPW_COM FPT_COM
@@ -1524,7 +1574,7 @@ parse_dot(VALUE str, VALUE hash)
 {
     static const char pat_source[] =
 #ifndef TIGHT_PARSER
-	"('?-?\\d+)\\.\\s*('?\\d+)\\.\\s*('?-?\\d+)"
+	"('?-?" NUMBER "+)\\.\\s*('?\\d+)\\.\\s*('?-?\\d+)"
 #else
 	BOS
 	FPW_COM FPT_COM
@@ -1684,7 +1734,7 @@ parse_mday(VALUE str, VALUE hash)
 {
     static const char pat_source[] =
 #ifndef TIGHT_PARSER
-	"(\\d+)(st|nd|rd|th)\\b"
+	"(" NUMBER "+)(st|nd|rd|th)\\b"
 #else
 	BOS
 	FPW_COM FPT_COM
@@ -1922,7 +1972,7 @@ parse_ddd(VALUE str, VALUE hash)
 #ifdef TIGHT_PARSER
 		BOS
 #endif
-		"([-+]?)(\\d{2,14})"
+		"([-+]?)(" NUMBER "{2,14})"
 		  "(?:"
 		    "\\s*"
 		    "t?"

data/ext/date/date_strptime.c

@@ -7,31 +7,21 @@
 #include "ruby/re.h"
 #include <ctype.h>
 
+#undef strncasecmp
+#define strncasecmp STRNCASECMP
+
 static const char *day_names[] = {
     "Sunday", "Monday", "Tuesday", "Wednesday",
     "Thursday", "Friday", "Saturday",
-    "Sun", "Mon", "Tue", "Wed",
-    "Thu", "Fri", "Sat"
 };
+static const int ABBREVIATED_DAY_NAME_LENGTH = 3;
 
 static const char *month_names[] = {
     "January", "February", "March", "April",
     "May", "June", "July", "August", "September",
     "October", "November", "December",
-    "Jan", "Feb", "Mar", "Apr", "May", "Jun",
-    "Jul", "Aug", "Sep", "Oct", "Nov", "Dec"
-};
-
-static const char *merid_names[] = {
-    "am", "pm",
-    "a.m.", "p.m."
-};
-
-static const char *extz_pats[] = {
-    ":z",
-    "::z",
-    ":::z"
 };
+static const int ABBREVIATED_MONTH_NAME_LENGTH = 3;
 
 #define sizeof_array(o) (sizeof o / sizeof o[0])
 
@@ -75,7 +65,7 @@ num_pattern_p(const char *s)
 #define NUM_PATTERN_P() num_pattern_p(&fmt[fi + 1])
 
 static long
-read_digits(const char *s, VALUE *n, size_t width)
+read_digits(const char *s, size_t slen, VALUE *n, size_t width)
 {
     size_t l;
 
@@ -83,7 +73,7 @@ read_digits(const char *s, VALUE *n, size_t width)
         return 0;
 
     l = 0;
-    while (ISDIGIT(s[l])) {
+    while (l < slen && ISDIGIT(s[l])) {
         if (++l == width) break;
     }
 
@@ -131,9 +121,10 @@ do { \
 #define READ_DIGITS(n,w) \
 do { \
     size_t l; \
-    l = read_digits(&str[si], &n, w); \
-    if (l == 0) \
+    l = read_digits(&str[si], slen - si, &n, w); \
+    if (l == 0) { \
 	fail();	\
+    } \
     si += l; \
 } while (0)
 
@@ -161,6 +152,12 @@ do { \
 
 VALUE date_zone_to_diff(VALUE);
 
+static inline int
+head_match_p(size_t len, const char *name, const char *str, size_t slen, size_t si)
+{
+    return slen - si >= len && strncasecmp(name, &str[si], len) == 0;
+}
+
 static size_t
 date__strptime_internal(const char *str, size_t slen,
 			const char *fmt, size_t flen, VALUE hash)
@@ -168,9 +165,18 @@ date__strptime_internal(const char *str, size_t slen,
     size_t si, fi;
     int c;
 
+#define HEAD_MATCH_P(len, name) head_match_p(len, name, str, slen, si)
     si = fi = 0;
 
     while (fi < flen) {
+	if (isspace((unsigned char)fmt[fi])) {
+	    while (si < slen && isspace((unsigned char)str[si]))
+		si++;
+	    while (++fi < flen && isspace((unsigned char)fmt[fi]));
+	    continue;
+	}
+
+	if (si >= slen) fail();
 
 	switch (fmt[fi]) {
 	  case '%':
@@ -194,12 +200,11 @@ date__strptime_internal(const char *str, size_t slen,
 		{
 		    int i;
 
-		    for (i = 0; i < (int)sizeof_array(extz_pats); i++)
-			if (strncmp(extz_pats[i], &fmt[fi],
-					strlen(extz_pats[i])) == 0) {
-			    fi += i;
-			    goto again;
-			}
+		    for (i = 1; i < 3 && fi + i < flen && fmt[fi+i] == ':'; ++i);
+		    if (fmt[fi+i] == 'z') {
+			fi += i - 1;
+			goto again;
+		    }
 		    fail();
 		}
 
@@ -209,10 +214,12 @@ date__strptime_internal(const char *str, size_t slen,
 		    int i;
 
 		    for (i = 0; i < (int)sizeof_array(day_names); i++) {
-			size_t l = strlen(day_names[i]);
-			if (strncasecmp(day_names[i], &str[si], l) == 0) {
+			const char *day_name = day_names[i];
+			size_t l = strlen(day_name);
+			if (HEAD_MATCH_P(l, day_name) ||
+			    HEAD_MATCH_P(l = ABBREVIATED_DAY_NAME_LENGTH, day_name)) {
 			    si += l;
-			    set_hash("wday", INT2FIX(i % 7));
+			    set_hash("wday", INT2FIX(i));
 			    goto matched;
 			}
 		    }
@@ -225,10 +232,12 @@ date__strptime_internal(const char *str, size_t slen,
 		    int i;
 
 		    for (i = 0; i < (int)sizeof_array(month_names); i++) {
-			size_t l = strlen(month_names[i]);
-			if (strncasecmp(month_names[i], &str[si], l) == 0) {
+			const char *month_name = month_names[i];
+			size_t l = strlen(month_name);
+			if (HEAD_MATCH_P(l, month_name) ||
+			    HEAD_MATCH_P(l = ABBREVIATED_MONTH_NAME_LENGTH, month_name)) {
 			    si += l;
-			    set_hash("mon", INT2FIX((i % 12) + 1));
+			    set_hash("mon", INT2FIX(i + 1));
 			    goto matched;
 			}
 		    }
@@ -402,18 +411,19 @@ date__strptime_internal(const char *str, size_t slen,
 
 	      case 'P':
 	      case 'p':
+		if (slen - si < 2) fail();
 		{
-		    int i;
-
-		    for (i = 0; i < 4; i++) {
-			size_t l = strlen(merid_names[i]);
-			if (strncasecmp(merid_names[i], &str[si], l) == 0) {
-			    si += l;
-			    set_hash("_merid", INT2FIX((i % 2) == 0 ? 0 : 12));
-			    goto matched;
-			}
+		    char c = str[si];
+		    const int hour = (c == 'P' || c == 'p') ? 12 : 0;
+		    if (!hour && !(c == 'A' || c == 'a')) fail();
+		    if ((c = str[si+1]) == '.') {
+			if (slen - si < 4 || str[si+3] != '.') fail();
+			c = str[si += 2];
 		    }
-		    fail();
+		    if (!(c == 'M' || c == 'm')) fail();
+		    si += 2;
+		    set_hash("_merid", INT2FIX(hour));
+		    goto matched;
 		}
 
 	      case 'Q':
@@ -587,7 +597,7 @@ date__strptime_internal(const char *str, size_t slen,
 
 		    b = rb_backref_get();
 		    rb_match_busy(b);
-		    m = f_match(pat, rb_usascii_str_new2(&str[si]));
+		    m = f_match(pat, rb_usascii_str_new(&str[si], slen - si));
 
 		    if (!NIL_P(m)) {
 			VALUE s, l, o;
@@ -619,22 +629,13 @@ date__strptime_internal(const char *str, size_t slen,
 		if (str[si] != '%')
 		    fail();
 		si++;
-		if (fi < flen)
-		    if (str[si] != fmt[fi])
+		if (fi < flen) {
+		    if (si >= slen || str[si] != fmt[fi])
 			fail();
-		si++;
+		    si++;
+		}
 		goto matched;
 	    }
-	  case ' ':
-	  case '\t':
-	  case '\n':
-	  case '\v':
-	  case '\f':
-	  case '\r':
-	    while (isspace((unsigned char)str[si]))
-		si++;
-	    fi++;
-	    break;
 	  default:
 	  ordinal:
 	    if (str[si] != fmt[fi])

data/ext/date/extconf.rb

@@ -3,6 +3,8 @@ require 'mkmf'
 
 config_string("strict_warnflags") {|w| $warnflags += " #{w}"}
 
+append_cflags("-Wno-compound-token-split-by-macro") if RUBY_VERSION < "2.7."
+have_func("rb_category_warn")
 with_werror("", {:werror => true}) do |opt, |
   have_var("timezone", "time.h", opt)
   have_var("altzone", "time.h", opt)

data/ext/date/prereq.mk

@@ -1,7 +1,7 @@
 .SUFFIXES: .list
 
 .list.h:
-	gperf --ignore-case -C -c -P -p -j1 -i 1 -g -o -t -N $(*F) $< \
+	gperf --ignore-case -L ANSI-C -C -c -P -p -j1 -i 1 -g -o -t -N $(*F) $< \
 	| sed -f $(top_srcdir)/tool/gperf.sed \
 	> $(@F)