2 # http://opensource.org/licenses/MIT
4 Add support for use of the system timezone database, rather
5 than embedding a copy. Discussed upstream but was not desired.
7 (Few) upstream reports:
8 https://bugs.php.net/bug.php?id=53320
9 https://bugs.php.net/bug.php?id=54250
12 r15: adapt for timelib 2017.05beta7 (in 7.2.0RC1)
13 r14: improve check for valid tz file
14 r13: adapt for upstream changes to use PHP allocator
15 r12: adapt for upstream changes for new zic
16 r11: use canonical names to avoid more case sensitivity issues
17 round lat/long from zone.tab towards zero per builtin db
18 r10: make timezone case insensitive
19 r9: fix another compile error without --with-system-tzdata configured (Michael Heimpold)
20 r8: fix compile error without --with-system-tzdata configured
21 r7: improve check for valid timezone id to exclude directories
22 r6: fix fd leak in r5, fix country code/BC flag use in
23 timezone_identifiers_list() using system db,
24 fix use of PECL timezonedb to override system db,
25 r5: reverts addition of "System/Localtime" fake tzname.
26 updated for 5.3.0, parses zone.tab to pick up mapping between
27 timezone name, country code and long/lat coords
28 r4: added "System/Localtime" tzname which uses /etc/localtime
29 r3: fix a crash if /usr/share/zoneinfo doesn't exist (Raphael Geissert)
30 r2: add filesystem trawl to set up name alias index
33 diff -up ./ext/date/lib/parse_tz.c.systzdata ./ext/date/lib/parse_tz.c
34 --- ./ext/date/lib/parse_tz.c.systzdata 2017-08-22 09:40:38.000000000 +0200
35 +++ ./ext/date/lib/parse_tz.c 2017-08-22 12:16:00.370298079 +0200
38 #include "timelib_private.h"
40 +#ifdef HAVE_SYSTEM_TZDATA
41 +#include <sys/mman.h>
42 +#include <sys/stat.h>
47 +#include "php_scandir.h"
50 #define TIMELIB_SUPPORTS_V2DATA
51 #include "timezonedb.h"
56 #if (defined(__APPLE__) || defined(__APPLE_CC__)) && (defined(__BIG_ENDIAN__) || defined(__LITTLE_ENDIAN__))
57 # if defined(__LITTLE_ENDIAN__)
58 @@ -67,6 +80,11 @@ static int read_php_preamble(const unsig
62 + if (memcmp(*tzf, "TZif", 4) == 0) {
68 version = (*tzf)[3] - '0';
70 @@ -374,7 +392,429 @@ void timelib_dump_tzinfo(timelib_tzinfo
74 -static int seek_to_tz_position(const unsigned char **tzf, char *timezone, const timelib_tzdb *tzdb)
75 +#ifdef HAVE_SYSTEM_TZDATA
77 +#ifdef HAVE_SYSTEM_TZDATA_PREFIX
78 +#define ZONEINFO_PREFIX HAVE_SYSTEM_TZDATA_PREFIX
80 +#define ZONEINFO_PREFIX "/usr/share/zoneinfo"
83 +/* System timezone database pointer. */
84 +static const timelib_tzdb *timezonedb_system;
86 +/* Hash table entry for the cache of the zone.tab mapping table. */
87 +struct location_info {
89 + double latitude, longitude;
92 + struct location_info *next;
95 +/* Cache of zone.tab. */
96 +static struct location_info **system_location_table;
98 +/* Size of the zone.tab hash table; a random-ish prime big enough to
99 + * prevent too many collisions. */
100 +#define LOCINFO_HASH_SIZE (1021)
102 +/* Compute a case insensitive hash of str */
103 +static uint32_t tz_hash(const char *str)
105 + const unsigned char *p = (const unsigned char *)str;
106 + uint32_t hash = 5381;
109 + while ((c = tolower(*p++)) != '\0') {
110 + hash = (hash << 5) ^ hash ^ c;
113 + return hash % LOCINFO_HASH_SIZE;
116 +/* Parse an ISO-6709 date as used in zone.tab. Returns end of the
117 + * parsed string on success, or NULL on parse error. On success,
118 + * writes the parsed number to *result. */
119 +static char *parse_iso6709(char *p, double *result)
127 + else if (*p == '-')
133 + for (pend = p; *pend >= '0' && *pend <= '9'; pend++)
136 + /* Annoying encoding used by zone.tab has no decimal point, so use
137 + * the length to determine the format:
145 + if (len < 4 || len > 7) {
150 + v = (p[0] - '0') * 10.0 + (p[1] - '0');
152 + if (len == 5 || len == 7)
153 + v = v * 10.0 + (*p++ - '0');
155 + v += (10.0 * (p[0] - '0')
156 + + p[1] - '0') / 60.0;
160 + v += (10.0 * (p[0] - '0')
161 + + p[1] - '0') / 3600.0;
165 + /* Round to five decimal place, not because it's a good idea,
166 + * but, because the builtin data uses rounded data, so, match
168 + *result = trunc(v * sign * 100000.0) / 100000.0;
173 +/* This function parses the zone.tab file to build up the mapping of
174 + * timezone to country code and geographic location, and returns a
175 + * hash table. The hash table is indexed by the function:
177 + * tz_hash(timezone-name)
179 +static struct location_info **create_location_table(void)
181 + struct location_info **li, *i;
182 + char zone_tab[PATH_MAX];
186 + strncpy(zone_tab, ZONEINFO_PREFIX "/zone.tab", sizeof zone_tab);
188 + fp = fopen(zone_tab, "r");
193 + li = calloc(LOCINFO_HASH_SIZE, sizeof *li);
195 + while (fgets(line, sizeof line, fp)) {
196 + char *p = line, *code, *name, *comment;
198 + double latitude, longitude;
200 + while (isspace(*p))
203 + if (*p == '#' || *p == '\0' || *p == '\n')
206 + if (!isalpha(p[0]) || !isalpha(p[1]) || p[2] != '\t')
214 + /* coords => [+-][D]DDMM[SS][+-][D]DDMM[SS] */
215 + p = parse_iso6709(p, &latitude);
219 + p = parse_iso6709(p, &longitude);
224 + if (!p || *p != '\t') {
228 + /* name = string */
230 + while (*p != '\t' && *p && *p != '\n')
235 + /* comment = string */
237 + while (*p != '\t' && *p && *p != '\n')
240 + if (*p == '\n' || *p == '\t')
243 + hash = tz_hash(name);
244 + i = malloc(sizeof *i);
245 + memcpy(i->code, code, 2);
246 + strncpy(i->name, name, sizeof i->name);
247 + i->comment = strdup(comment);
248 + i->longitude = longitude;
249 + i->latitude = latitude;
250 + i->next = li[hash];
252 + /* printf("%s [%u, %f, %f]\n", name, hash, latitude, longitude); */
260 +/* Return location info from hash table, using given timezone name.
261 + * Returns NULL if the name could not be found. */
262 +const struct location_info *find_zone_info(struct location_info **li,
265 + uint32_t hash = tz_hash(name);
266 + const struct location_info *l;
272 + for (l = li[hash]; l; l = l->next) {
273 + if (strcasecmp(l->name, name) == 0)
280 +/* Filter out some non-tzdata files and the posix/right databases, if
282 +static int index_filter(const struct dirent *ent)
284 + return strcmp(ent->d_name, ".") != 0
285 + && strcmp(ent->d_name, "..") != 0
286 + && strcmp(ent->d_name, "posix") != 0
287 + && strcmp(ent->d_name, "posixrules") != 0
288 + && strcmp(ent->d_name, "right") != 0
289 + && strstr(ent->d_name, ".list") == NULL
290 + && strstr(ent->d_name, ".tab") == NULL;
293 +static int sysdbcmp(const void *first, const void *second)
295 + const timelib_tzdb_index_entry *alpha = first, *beta = second;
297 + return strcasecmp(alpha->id, beta->id);
301 +/* Create the zone identifier index by trawling the filesystem. */
302 +static void create_zone_index(timelib_tzdb *db)
304 + size_t dirstack_size, dirstack_top;
305 + size_t index_size, index_next;
306 + timelib_tzdb_index_entry *db_index;
309 + /* LIFO stack to hold directory entries to scan; each slot is a
310 + * directory name relative to the zoneinfo prefix. */
311 + dirstack_size = 32;
312 + dirstack = malloc(dirstack_size * sizeof *dirstack);
314 + dirstack[0] = strdup("");
318 + db_index = malloc(index_size * sizeof *db_index);
322 + struct dirent **ents;
323 + char name[PATH_MAX], *top;
326 + /* Pop the top stack entry, and iterate through its contents. */
327 + top = dirstack[--dirstack_top];
328 + snprintf(name, sizeof name, ZONEINFO_PREFIX "/%s", top);
330 + count = php_scandir(name, &ents, index_filter, php_alphasort);
332 + while (count > 0) {
334 + const char *leaf = ents[count - 1]->d_name;
336 + snprintf(name, sizeof name, ZONEINFO_PREFIX "/%s/%s",
339 + if (strlen(name) && stat(name, &st) == 0) {
340 + /* Name, relative to the zoneinfo prefix. */
341 + const char *root = top;
343 + if (root[0] == '/') root++;
345 + snprintf(name, sizeof name, "%s%s%s", root,
346 + *root ? "/": "", leaf);
348 + if (S_ISDIR(st.st_mode)) {
349 + if (dirstack_top == dirstack_size) {
350 + dirstack_size *= 2;
351 + dirstack = realloc(dirstack,
352 + dirstack_size * sizeof *dirstack);
354 + dirstack[dirstack_top++] = strdup(name);
357 + if (index_next == index_size) {
359 + db_index = realloc(db_index,
360 + index_size * sizeof *db_index);
363 + db_index[index_next++].id = strdup(name);
367 + free(ents[--count]);
370 + if (count != -1) free(ents);
372 + } while (dirstack_top);
374 + qsort(db_index, index_next, sizeof *db_index, sysdbcmp);
376 + db->index = db_index;
377 + db->index_size = index_next;
382 +#define FAKE_HEADER "1234\0??\1??"
383 +#define FAKE_UTC_POS (7 - 4)
385 +/* Create a fake data segment for database 'sysdb'. */
386 +static void fake_data_segment(timelib_tzdb *sysdb,
387 + struct location_info **info)
392 + data = malloc(3 * sysdb->index_size + 7);
394 + p = mempcpy(data, FAKE_HEADER, sizeof(FAKE_HEADER) - 1);
396 + for (n = 0; n < sysdb->index_size; n++) {
397 + const struct location_info *li;
398 + timelib_tzdb_index_entry *ent;
400 + ent = (timelib_tzdb_index_entry *)&sysdb->index[n];
402 + /* Lookup the timezone name in the hash table. */
403 + if (strcmp(ent->id, "UTC") == 0) {
404 + ent->pos = FAKE_UTC_POS;
408 + li = find_zone_info(info, ent->id);
410 + /* If found, append the BC byte and the
411 + * country code; set the position for this
412 + * section of timezone data. */
413 + ent->pos = (p - data) - 4;
415 + *p++ = li->code[0];
416 + *p++ = li->code[1];
419 + /* If not found, the timezone data can
420 + * point at the header. */
425 + sysdb->data = (unsigned char *)data;
428 +/* Returns true if the passed-in stat structure describes a
429 + * probably-valid timezone file. */
430 +static int is_valid_tzfile(const struct stat *st, int fd)
434 + if (read(fd, buf, 20)!=20) {
437 + lseek(fd, SEEK_SET, 0);
438 + if (memcmp(buf, "TZif", 4)) {
442 + return S_ISREG(st->st_mode) && st->st_size > 20;
445 +/* To allow timezone names to be used case-insensitively, find the
446 + * canonical name for this timezone, if possible. */
447 +static const char *canonical_tzname(const char *timezone)
449 + if (timezonedb_system) {
450 + timelib_tzdb_index_entry *ent, lookup;
452 + lookup.id = (char *)timezone;
454 + ent = bsearch(&lookup, timezonedb_system->index,
455 + timezonedb_system->index_size, sizeof lookup,
465 +/* Return the mmap()ed tzfile if found, else NULL. On success, the
466 + * length of the mapped data is placed in *length. */
467 +static char *map_tzfile(const char *timezone, size_t *length)
469 + char fname[PATH_MAX];
474 + if (timezone[0] == '\0' || strstr(timezone, "..") != NULL) {
478 + snprintf(fname, sizeof fname, ZONEINFO_PREFIX "/%s", canonical_tzname(timezone));
480 + fd = open(fname, O_RDONLY);
483 + } else if (fstat(fd, &st) != 0 || !is_valid_tzfile(&st, fd)) {
488 + *length = st.st_size;
489 + p = mmap(NULL, st.st_size, PROT_READ, MAP_SHARED, fd, 0);
492 + return p != MAP_FAILED ? p : NULL;
497 +static int inmem_seek_to_tz_position(const unsigned char **tzf, char *timezone, const timelib_tzdb *tzdb)
499 int left = 0, right = tzdb->index_size - 1;
500 #ifdef HAVE_SETLOCALE
501 @@ -419,9 +859,48 @@ static int seek_to_tz_position(const uns
505 +static int seek_to_tz_position(const unsigned char **tzf, char *timezone,
506 + char **map, size_t *maplen,
507 + const timelib_tzdb *tzdb)
509 +#ifdef HAVE_SYSTEM_TZDATA
510 + if (tzdb == timezonedb_system) {
513 + orig = map_tzfile(timezone, maplen);
514 + if (orig == NULL) {
518 + (*tzf) = (unsigned char *)orig;
525 + return inmem_seek_to_tz_position(tzf, timezone, tzdb);
529 const timelib_tzdb *timelib_builtin_db(void)
531 +#ifdef HAVE_SYSTEM_TZDATA
532 + if (timezonedb_system == NULL) {
533 + timelib_tzdb *tmp = malloc(sizeof *tmp);
535 + tmp->version = "0.system";
537 + create_zone_index(tmp);
538 + system_location_table = create_location_table();
539 + fake_data_segment(tmp, system_location_table);
540 + timezonedb_system = tmp;
543 + return timezonedb_system;
545 return &timezonedb_builtin;
549 const timelib_tzdb_index_entry *timelib_timezone_identifiers_list(timelib_tzdb *tzdb, int *count)
550 @@ -433,7 +912,30 @@ const timelib_tzdb_index_entry *timelib_
551 int timelib_timezone_id_is_valid(char *timezone, const timelib_tzdb *tzdb)
553 const unsigned char *tzf;
554 - return (seek_to_tz_position(&tzf, timezone, tzdb));
556 +#ifdef HAVE_SYSTEM_TZDATA
557 + if (tzdb == timezonedb_system) {
558 + char fname[PATH_MAX];
561 + if (timezone[0] == '\0' || strstr(timezone, "..") != NULL) {
565 + if (system_location_table) {
566 + if (find_zone_info(system_location_table, timezone) != NULL) {
567 + /* found in cache */
572 + snprintf(fname, sizeof fname, ZONEINFO_PREFIX "/%s", canonical_tzname(timezone));
574 + return stat(fname, &st) == 0 && is_valid_tzfile(&st, 0);
578 + return (inmem_seek_to_tz_position(&tzf, timezone, tzdb));
581 static int skip_64bit_preamble(const unsigned char **tzf, timelib_tzinfo *tz)
582 @@ -475,12 +977,14 @@ static timelib_tzinfo* timelib_tzinfo_ct
583 timelib_tzinfo *timelib_parse_tzfile(char *timezone, const timelib_tzdb *tzdb, int *error_code)
585 const unsigned char *tzf;
586 + char *memmap = NULL;
590 int transitions_result, types_result;
591 unsigned int type; /* TIMELIB_TZINFO_PHP or TIMELIB_TZINFO_ZONEINFO */
593 - if (seek_to_tz_position(&tzf, timezone, tzdb)) {
594 + if (seek_to_tz_position(&tzf, timezone, &memmap, &maplen, tzdb)) {
595 tmp = timelib_tzinfo_ctor(timezone);
597 version = read_preamble(&tzf, tmp, &type);
598 @@ -503,6 +1007,29 @@ timelib_tzinfo *timelib_parse_tzfile(cha
599 timelib_tzinfo_dtor(tmp);
603 +#ifdef HAVE_SYSTEM_TZDATA
605 + const struct location_info *li;
607 + /* TZif-style - grok the location info from the system database,
610 + if ((li = find_zone_info(system_location_table, timezone)) != NULL) {
611 + tmp->location.comments = timelib_strdup(li->comment);
612 + strncpy(tmp->location.country_code, li->code, 2);
613 + tmp->location.longitude = li->longitude;
614 + tmp->location.latitude = li->latitude;
618 + set_default_location_and_comments(&tzf, tmp);
621 + /* Now done with the mmap segment - discard it. */
622 + munmap(memmap, maplen);
625 if (version == 2 || version == 3) {
626 if (!skip_64bit_preamble(&tzf, tmp)) {
627 /* 64 bit preamble is not in place */
628 @@ -520,6 +1047,9 @@ timelib_tzinfo *timelib_parse_tzfile(cha
630 set_default_location_and_comments(&tzf, tmp);
632 +#ifdef HAVE_SYSTEM_TZDATA
636 *error_code = TIMELIB_ERROR_NO_SUCH_TIMEZONE;
638 diff -up ./ext/date/lib/timelib.m4.systzdata ./ext/date/lib/timelib.m4
639 --- ./ext/date/lib/timelib.m4.systzdata 2017-08-22 09:40:38.000000000 +0200
640 +++ ./ext/date/lib/timelib.m4 2017-08-22 11:32:29.357799927 +0200
641 @@ -81,3 +81,16 @@ io.h
643 dnl Check for strtoll, atoll
644 AC_CHECK_FUNCS(strtoll atoll strftime gettimeofday)
646 +PHP_ARG_WITH(system-tzdata, for use of system timezone data,
647 +[ --with-system-tzdata[=DIR] to specify use of system timezone data],
650 +if test "$PHP_SYSTEM_TZDATA" != "no"; then
651 + AC_DEFINE(HAVE_SYSTEM_TZDATA, 1, [Define if system timezone data is used])
653 + if test "$PHP_SYSTEM_TZDATA" != "yes"; then
654 + AC_DEFINE_UNQUOTED(HAVE_SYSTEM_TZDATA_PREFIX, "$PHP_SYSTEM_TZDATA",
655 + [Define for location of system timezone data])