2 Add support for use of the system timezone database, rather
3 than embedding a copy. Discussed upstream but was not desired.
6 r9: fix another compile error without --with-system-tzdata configured
7 r8: fix compile error without --with-system-tzdata configured
8 r7: improve check for valid timezone id to exclude directories
9 r6: fix fd leak in r5, fix country code/BC flag use in
10 timezone_identifiers_list() using system db,
11 fix use of PECL timezonedb to override system db,
12 r5: reverts addition of "System/Localtime" fake tzname.
13 updated for 5.3.0, parses zone.tab to pick up mapping between
14 timezone name, country code and long/lat coords
15 r4: added "System/Localtime" tzname which uses /etc/localtime
16 r3: fix a crash if /usr/share/zoneinfo doesn't exist (Raphael Geissert)
17 r2: add filesystem trawl to set up name alias index
20 --- a/ext/date/lib/parse_tz.c
21 +++ b/ext/date/lib/parse_tz.c
26 +#ifdef HAVE_SYSTEM_TZDATA
27 +#include <sys/mman.h>
28 +#include <sys/stat.h>
33 +#include "php_scandir.h"
44 +#ifndef HAVE_SYSTEM_TZDATA
45 #include "timezonedb.h"
50 #if (defined(__APPLE__) || defined(__APPLE_CC__)) && (defined(__BIG_ENDIAN__) || defined(__LITTLE_ENDIAN__))
51 # if defined(__LITTLE_ENDIAN__)
54 static void read_preamble(const unsigned char **tzf, timelib_tzinfo *tz)
59 + if (memcmp(tzf, "TZif", 4) == 0) {
68 tz->bc = (**tzf == '\1');
70 @@ -256,7 +276,397 @@ 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 = NULL;
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 +static uint32_t tz_hash(const char *str)
104 + const unsigned char *p = (const unsigned char *)str;
105 + uint32_t hash = 5381;
108 + while ((c = *p++) != '\0') {
109 + hash = (hash << 5) ^ hash ^ c;
112 + return hash % LOCINFO_HASH_SIZE;
115 +/* Parse an ISO-6709 date as used in zone.tab. Returns end of the
116 + * parsed string on success, or NULL on parse error. On success,
117 + * writes the parsed number to *result. */
118 +static char *parse_iso6709(char *p, double *result)
126 + else if (*p == '-')
132 + for (pend = p; *pend >= '0' && *pend <= '9'; pend++)
135 + /* Annoying encoding used by zone.tab has no decimal point, so use
136 + * the length to determine the format:
144 + if (len < 4 || len > 7) {
149 + v = (p[0] - '0') * 10.0 + (p[1] - '0');
151 + if (len == 5 || len == 7)
152 + v = v * 10.0 + (*p++ - '0');
154 + v += (10.0 * (p[0] - '0')
155 + + p[1] - '0') / 60.0;
159 + v += (10.0 * (p[0] - '0')
160 + + p[1] - '0') / 3600.0;
164 + /* Round to five decimal place, not because it's a good idea,
165 + * but, because the builtin data uses rounded data, so, match
167 + *result = round(v * sign * 100000.0) / 100000.0;
172 +/* This function parses the zone.tab file to build up the mapping of
173 + * timezone to country code and geographic location, and returns a
174 + * hash table. The hash table is indexed by the function:
176 + * tz_hash(timezone-name)
178 +static struct location_info **create_location_table(void)
180 + struct location_info **li, *i;
181 + char zone_tab[PATH_MAX];
185 + strncpy(zone_tab, ZONEINFO_PREFIX "/zone.tab", sizeof zone_tab);
187 + fp = fopen(zone_tab, "r");
192 + li = calloc(LOCINFO_HASH_SIZE, sizeof *li);
194 + while (fgets(line, sizeof line, fp)) {
195 + char *p = line, *code, *name, *comment;
197 + double latitude, longitude;
199 + while (isspace(*p))
202 + if (*p == '#' || *p == '\0' || *p == '\n')
205 + if (!isalpha(p[0]) || !isalpha(p[1]) || p[2] != '\t')
213 + /* coords => [+-][D]DDMM[SS][+-][D]DDMM[SS] */
214 + p = parse_iso6709(p, &latitude);
218 + p = parse_iso6709(p, &longitude);
223 + if (!p || *p != '\t') {
227 + /* name = string */
229 + while (*p != '\t' && *p && *p != '\n')
234 + /* comment = string */
236 + while (*p != '\t' && *p && *p != '\n')
239 + if (*p == '\n' || *p == '\t')
242 + hash = tz_hash(name);
243 + i = malloc(sizeof *i);
244 + memcpy(i->code, code, 2);
245 + strncpy(i->name, name, sizeof i->name);
246 + i->comment = strdup(comment);
247 + i->longitude = longitude;
248 + i->latitude = latitude;
249 + i->next = li[hash];
251 + /* printf("%s [%u, %f, %f]\n", name, hash, latitude, longitude); */
259 +/* Return location info from hash table, using given timezone name.
260 + * Returns NULL if the name could not be found. */
261 +const struct location_info *find_zone_info(struct location_info **li,
264 + uint32_t hash = tz_hash(name);
265 + const struct location_info *l;
271 + for (l = li[hash]; l; l = l->next) {
272 + if (strcasecmp(l->name, name) == 0)
279 +/* Filter out some non-tzdata files and the posix/right databases, if
281 +static int index_filter(const struct dirent *ent)
283 + return strcmp(ent->d_name, ".") != 0
284 + && strcmp(ent->d_name, "..") != 0
285 + && strcmp(ent->d_name, "posix") != 0
286 + && strcmp(ent->d_name, "posixrules") != 0
287 + && strcmp(ent->d_name, "right") != 0
288 + && strstr(ent->d_name, ".tab") == NULL;
291 +static int sysdbcmp(const void *first, const void *second)
293 + const timelib_tzdb_index_entry *alpha = first, *beta = second;
295 + return strcmp(alpha->id, beta->id);
299 +/* Create the zone identifier index by trawling the filesystem. */
300 +static void create_zone_index(timelib_tzdb *db)
302 + size_t dirstack_size, dirstack_top;
303 + size_t index_size, index_next;
304 + timelib_tzdb_index_entry *db_index;
307 + /* LIFO stack to hold directory entries to scan; each slot is a
308 + * directory name relative to the zoneinfo prefix. */
309 + dirstack_size = 32;
310 + dirstack = malloc(dirstack_size * sizeof *dirstack);
312 + dirstack[0] = strdup("");
316 + db_index = malloc(index_size * sizeof *db_index);
320 + struct dirent **ents;
321 + char name[PATH_MAX], *top;
324 + /* Pop the top stack entry, and iterate through its contents. */
325 + top = dirstack[--dirstack_top];
326 + snprintf(name, sizeof name, ZONEINFO_PREFIX "/%s", top);
328 + count = php_scandir(name, &ents, index_filter, php_alphasort);
330 + while (count > 0) {
332 + const char *leaf = ents[count - 1]->d_name;
334 + snprintf(name, sizeof name, ZONEINFO_PREFIX "/%s/%s",
337 + if (strlen(name) && stat(name, &st) == 0) {
338 + /* Name, relative to the zoneinfo prefix. */
339 + const char *root = top;
341 + if (root[0] == '/') root++;
343 + snprintf(name, sizeof name, "%s%s%s", root,
344 + *root ? "/": "", leaf);
346 + if (S_ISDIR(st.st_mode)) {
347 + if (dirstack_top == dirstack_size) {
348 + dirstack_size *= 2;
349 + dirstack = realloc(dirstack,
350 + dirstack_size * sizeof *dirstack);
352 + dirstack[dirstack_top++] = strdup(name);
355 + if (index_next == index_size) {
357 + db_index = realloc(db_index,
358 + index_size * sizeof *db_index);
361 + db_index[index_next++].id = strdup(name);
365 + free(ents[--count]);
368 + if (count != -1) free(ents);
370 + } while (dirstack_top);
372 + qsort(db_index, index_next, sizeof *db_index, sysdbcmp);
374 + db->index = db_index;
375 + db->index_size = index_next;
380 +#define FAKE_HEADER "1234\0??\1??"
381 +#define FAKE_UTC_POS (7 - 4)
383 +/* Create a fake data segment for database 'sysdb'. */
384 +static void fake_data_segment(timelib_tzdb *sysdb,
385 + struct location_info **info)
390 + data = malloc(3 * sysdb->index_size + 7);
392 + p = mempcpy(data, FAKE_HEADER, sizeof(FAKE_HEADER) - 1);
394 + for (n = 0; n < sysdb->index_size; n++) {
395 + const struct location_info *li;
396 + timelib_tzdb_index_entry *ent;
398 + ent = (timelib_tzdb_index_entry *)&sysdb->index[n];
400 + /* Lookup the timezone name in the hash table. */
401 + if (strcmp(ent->id, "UTC") == 0) {
402 + ent->pos = FAKE_UTC_POS;
406 + li = find_zone_info(info, ent->id);
408 + /* If found, append the BC byte and the
409 + * country code; set the position for this
410 + * section of timezone data. */
411 + ent->pos = (p - data) - 4;
413 + *p++ = li->code[0];
414 + *p++ = li->code[1];
417 + /* If not found, the timezone data can
418 + * point at the header. */
423 + sysdb->data = (unsigned char *)data;
426 +/* Returns true if the passed-in stat structure describes a
427 + * probably-valid timezone file. */
428 +static int is_valid_tzfile(const struct stat *st)
430 + return S_ISREG(st->st_mode) && st->st_size > 20;
433 +/* Return the mmap()ed tzfile if found, else NULL. On success, the
434 + * length of the mapped data is placed in *length. */
435 +static char *map_tzfile(const char *timezone, size_t *length)
437 + char fname[PATH_MAX];
442 + if (timezone[0] == '\0' || strstr(timezone, "..") != NULL) {
446 + snprintf(fname, sizeof fname, ZONEINFO_PREFIX "/%s", timezone);
448 + fd = open(fname, O_RDONLY);
451 + } else if (fstat(fd, &st) != 0 || !is_valid_tzfile(&st)) {
456 + *length = st.st_size;
457 + p = mmap(NULL, st.st_size, PROT_READ, MAP_SHARED, fd, 0);
460 + return p != MAP_FAILED ? p : NULL;
465 +static int inmem_seek_to_tz_position(const unsigned char **tzf, char *timezone, const timelib_tzdb *tzdb)
467 int left = 0, right = tzdb->index_size - 1;
468 #ifdef HAVE_SETLOCALE
469 @@ -295,36 +705,128 @@ static int seek_to_tz_position(const uns
473 +static int seek_to_tz_position(const unsigned char **tzf, char *timezone,
474 + char **map, size_t *maplen,
475 + const timelib_tzdb *tzdb)
477 +#ifdef HAVE_SYSTEM_TZDATA
478 + if (tzdb == timezonedb_system) {
481 + orig = map_tzfile(timezone, maplen);
482 + if (orig == NULL) {
486 + (*tzf) = (unsigned char *)orig ;
494 + return inmem_seek_to_tz_position(tzf, timezone, tzdb);
498 const timelib_tzdb *timelib_builtin_db(void)
500 +#ifdef HAVE_SYSTEM_TZDATA
501 + if (timezonedb_system == NULL) {
502 + timelib_tzdb *tmp = malloc(sizeof *tmp);
504 + tmp->version = "0.system";
506 + create_zone_index(tmp);
507 + system_location_table = create_location_table();
508 + fake_data_segment(tmp, system_location_table);
509 + timezonedb_system = tmp;
513 + return timezonedb_system;
515 return &timezonedb_builtin;
519 const timelib_tzdb_index_entry *timelib_timezone_builtin_identifiers_list(int *count)
521 +#ifdef HAVE_SYSTEM_TZDATA
522 + *count = timezonedb_system->index_size;
523 + return timezonedb_system->index;
525 *count = sizeof(timezonedb_idx_builtin) / sizeof(*timezonedb_idx_builtin);
526 return timezonedb_idx_builtin;
530 int timelib_timezone_id_is_valid(char *timezone, const timelib_tzdb *tzdb)
532 const unsigned char *tzf;
533 - return (seek_to_tz_position(&tzf, timezone, tzdb));
535 +#ifdef HAVE_SYSTEM_TZDATA
536 + if (tzdb == timezonedb_system) {
537 + char fname[PATH_MAX];
540 + if (timezone[0] == '\0' || strstr(timezone, "..") != NULL) {
544 + snprintf(fname, sizeof fname, ZONEINFO_PREFIX "/%s", timezone);
546 + return stat(fname, &st) == 0 && is_valid_tzfile(&st);
550 + return (inmem_seek_to_tz_position(&tzf, timezone, tzdb));
553 timelib_tzinfo *timelib_parse_tzfile(char *timezone, const timelib_tzdb *tzdb)
555 const unsigned char *tzf;
556 + char *memmap = NULL;
560 - if (seek_to_tz_position(&tzf, timezone, tzdb)) {
561 + if (seek_to_tz_position(&tzf, timezone, &memmap, &maplen, tzdb)) {
562 tmp = timelib_tzinfo_ctor(timezone);
564 read_preamble(&tzf, tmp);
565 read_header(&tzf, tmp);
566 read_transistions(&tzf, tmp);
567 read_types(&tzf, tmp);
568 - read_location(&tzf, tmp);
570 +#ifdef HAVE_SYSTEM_TZDATA
572 + const struct location_info *li;
574 + /* TZif-style - grok the location info from the system database,
577 + if ((li = find_zone_info(system_location_table, timezone)) != NULL) {
578 + tmp->location.comments = strdup(li->comment);
579 + strncpy(tmp->location.country_code, li->code, 2);
580 + tmp->location.longitude = li->longitude;
581 + tmp->location.latitude = li->latitude;
585 + strcpy(tmp->location.country_code, "??");
587 + tmp->location.comments = strdup("");
590 + /* Now done with the mmap segment - discard it. */
591 + munmap(memmap, maplen);
595 + /* PHP-style - use the embedded info. */
596 + read_location(&tzf, tmp);
601 --- a/ext/date/lib/timelib.m4
602 +++ b/ext/date/lib/timelib.m4
603 @@ -78,3 +78,17 @@ stdlib.h
605 dnl Check for strtoll, atoll
606 AC_CHECK_FUNCS(strtoll atoll strftime)
608 +PHP_ARG_WITH(system-tzdata, for use of system timezone data,
609 +[ --with-system-tzdata[=DIR] to specify use of system timezone data],
612 +if test "$PHP_SYSTEM_TZDATA" != "no"; then
613 + AC_DEFINE(HAVE_SYSTEM_TZDATA, 1, [Define if system timezone data is used])
615 + if test "$PHP_SYSTEM_TZDATA" != "yes"; then
616 + AC_DEFINE_UNQUOTED(HAVE_SYSTEM_TZDATA_PREFIX, "$PHP_SYSTEM_TZDATA",
617 + [Define for location of system timezone data])