database/postgres
1
0
Fork 0
mirror of https://github.com/postgres/postgres.git synced 2026-01-27 21:43:08 +03:00
Code Activity
Files
fe07100e82b096d3c848cace790d4b4daf0c4131
postgres/src/include/utils/datetime.h
Tom Lane d7674c9fab Seek zone abbreviations in the IANA data before timezone_abbreviations. 
If a time zone abbreviation used in datetime input is defined in
the currently active timezone, use that definition in preference
to looking in the timezone_abbreviations list.  That allows us to
correctly handle abbreviations that have different meanings in
different timezones.  Also, it eliminates an inconsistency between
datetime input and datetime output: the non-ISO datestyles for
timestamptz have always printed abbreviations taken from the IANA
data, not from timezone_abbreviations.  Before this fix, it was
possible to demonstrate cases where casting a timestamp to text
and back fails or changes the value significantly because of that
inconsistency.

While this change removes the ability to override the IANA data about
an abbreviation known in the current zone, it's not clear that there's
any real use-case for doing so.  But it is clear that this makes life
a lot easier for dealing with abbreviations that have conflicts across
different time zones.

Also update the pg_timezone_abbrevs view to report abbreviations
that are recognized via the IANA data, and *not* report any
timezone_abbreviations entries that are thereby overridden.
Under the hood, there are now two SRFs, one that pulls the IANA
data and one that pulls timezone_abbreviations entries.  They're
combined by logic in the view.  This approach was useful for
debugging (since the functions can be called on their own).
While I don't intend to document the functions explicitly,
they might be useful to call directly.

Also improve DecodeTimezoneAbbrev's caching logic so that it can
cache zone abbreviations found in the IANA data.  Without that,
this patch would have caused a noticeable degradation of the
runtime of timestamptz_in.

Per report from Aleksander Alekseev and additional investigation.

Discussion: https://postgr.es/m/CAJ7c6TOATjJqvhnYsui0=CO5XFMF4dvTGH+skzB--jNhqSQu5g@mail.gmail.com
2025-01-16 14:11:19 -05:00

370 lines
11 KiB
C
Raw Blame History

/*-------------------------------------------------------------------------
*
* datetime.h
* Definitions for date/time support code.
* The support code is shared with other date data types,
* including date, and time.
*
*
* Portions Copyright (c) 1996-2025, PostgreSQL Global Development Group
* Portions Copyright (c) 1994, Regents of the University of California
*
* src/include/utils/datetime.h
*
*-------------------------------------------------------------------------
*/
#ifndef DATETIME_H
#define DATETIME_H
#include "utils/timestamp.h"
/* this struct is declared in utils/tzparser.h: */
struct tzEntry;
/* ----------------------------------------------------------------
* time types + support macros
*
* String definitions for standard time quantities.
*
* These strings are the defaults used to form output time strings.
* Other alternative forms are hardcoded into token tables in datetime.c.
* ----------------------------------------------------------------
*/
#define DAGO "ago"
#define DCURRENT "current"
#define EPOCH "epoch"
#define INVALID "invalid"
#define EARLY "-infinity"
#define LATE "infinity"
#define NOW "now"
#define TODAY "today"
#define TOMORROW "tomorrow"
#define YESTERDAY "yesterday"
#define ZULU "zulu"
#define DMICROSEC "usecond"
#define DMILLISEC "msecond"
#define DSECOND "second"
#define DMINUTE "minute"
#define DHOUR "hour"
#define DDAY "day"
#define DWEEK "week"
#define DMONTH "month"
#define DQUARTER "quarter"
#define DYEAR "year"
#define DDECADE "decade"
#define DCENTURY "century"
#define DMILLENNIUM "millennium"
#define DA_D "ad"
#define DB_C "bc"
#define DTIMEZONE "timezone"
/*
* Fundamental time field definitions for parsing.
*
* Meridian: am, pm, or 24-hour style.
* Millennium: ad, bc
*/
#define AM 0
#define PM 1
#define HR24 2
#define AD 0
#define BC 1
/*
* Field types for time decoding.
*
* Can't have more of these than there are bits in an unsigned int
* since these are turned into bit masks during parsing and decoding.
*
* Furthermore, the values for YEAR, MONTH, DAY, HOUR, MINUTE, SECOND
* must be in the range 0..14 so that the associated bitmasks can fit
* into the left half of an INTERVAL's typmod value. Since those bits
* are stored in typmods, you can't change them without initdb!
*/
#define RESERV 0
#define MONTH 1
#define YEAR 2
#define DAY 3
#define JULIAN 4
#define TZ 5 /* fixed-offset timezone abbreviation */
#define DTZ 6 /* fixed-offset timezone abbrev, DST */
#define DYNTZ 7 /* dynamic timezone abbreviation */
#define IGNORE_DTF 8
#define AMPM 9
#define HOUR 10
#define MINUTE 11
#define SECOND 12
#define MILLISECOND 13
#define MICROSECOND 14
#define DOY 15
#define DOW 16
#define UNITS 17
#define ADBC 18
/* these are only for relative dates */
#define AGO 19
#define ABS_BEFORE 20
#define ABS_AFTER 21
/* generic fields to help with parsing */
#define ISODATE 22
#define ISOTIME 23
/* these are only for parsing intervals */
#define WEEK 24
#define DECADE 25
#define CENTURY 26
#define MILLENNIUM 27
/* hack for parsing two-word timezone specs "MET DST" etc */
#define DTZMOD 28 /* "DST" as a separate word */
/* reserved for unrecognized string values */
#define UNKNOWN_FIELD 31
/*
* Token field definitions for time parsing and decoding.
*
* Some field type codes (see above) use these as the "value" in datetktbl[].
* These are also used for bit masks in DecodeDateTime and friends
* so actually restrict them to within [0,31] for now.
* - thomas 97/06/19
* Not all of these fields are used for masks in DecodeDateTime
* so allow some larger than 31. - thomas 1997-11-17
*
* Caution: there are undocumented assumptions in the code that most of these
* values are not equal to IGNORE_DTF nor RESERV. Be very careful when
* renumbering values in either of these apparently-independent lists :-(
*/
#define DTK_NUMBER 0
#define DTK_STRING 1
#define DTK_DATE 2
#define DTK_TIME 3
#define DTK_TZ 4
#define DTK_AGO 5
#define DTK_SPECIAL 6
#define DTK_EARLY 9
#define DTK_LATE 10
#define DTK_EPOCH 11
#define DTK_NOW 12
#define DTK_YESTERDAY 13
#define DTK_TODAY 14
#define DTK_TOMORROW 15
#define DTK_ZULU 16
#define DTK_DELTA 17
#define DTK_SECOND 18
#define DTK_MINUTE 19
#define DTK_HOUR 20
#define DTK_DAY 21
#define DTK_WEEK 22
#define DTK_MONTH 23
#define DTK_QUARTER 24
#define DTK_YEAR 25
#define DTK_DECADE 26
#define DTK_CENTURY 27
#define DTK_MILLENNIUM 28
#define DTK_MILLISEC 29
#define DTK_MICROSEC 30
#define DTK_JULIAN 31
#define DTK_DOW 32
#define DTK_DOY 33
#define DTK_TZ_HOUR 34
#define DTK_TZ_MINUTE 35
#define DTK_ISOYEAR 36
#define DTK_ISODOW 37
/*
* Bit mask definitions for time parsing.
*/
#define DTK_M(t) (0x01 << (t))
/* Convenience: a second, plus any fractional component */
#define DTK_ALL_SECS_M (DTK_M(SECOND) | DTK_M(MILLISECOND) | DTK_M(MICROSECOND))
#define DTK_DATE_M (DTK_M(YEAR) | DTK_M(MONTH) | DTK_M(DAY))
#define DTK_TIME_M (DTK_M(HOUR) | DTK_M(MINUTE) | DTK_ALL_SECS_M)
/*
* Working buffer size for input and output of interval, timestamp, etc.
* Inputs that need more working space will be rejected early. Longer outputs
* will overrun buffers, so this must suffice for all possible output. As of
* this writing, interval_out() needs the most space at ~90 bytes.
*/
#define MAXDATELEN 128
/* maximum possible number of fields in a date string */
#define MAXDATEFIELDS 25
/* only this many chars are stored in datetktbl */
#define TOKMAXLEN 10
/* keep this struct small; it gets used a lot */
typedef struct
{
char token[TOKMAXLEN + 1]; /* always NUL-terminated */
char type; /* see field type codes above */
int32 value; /* meaning depends on type */
} datetkn;
/* one of its uses is in tables of time zone abbreviations */
typedef struct TimeZoneAbbrevTable
{
Size tblsize; /* size in bytes of TimeZoneAbbrevTable */
int numabbrevs; /* number of entries in abbrevs[] array */
datetkn abbrevs[FLEXIBLE_ARRAY_MEMBER];
/* DynamicZoneAbbrev(s) may follow the abbrevs[] array */
} TimeZoneAbbrevTable;
/* auxiliary data for a dynamic time zone abbreviation (non-fixed-offset) */
typedef struct DynamicZoneAbbrev
{
pg_tz *tz; /* NULL if not yet looked up */
char zone[FLEXIBLE_ARRAY_MEMBER]; /* NUL-terminated zone name */
} DynamicZoneAbbrev;
/* FMODULO()
* Macro to replace modf(), which is broken on some platforms.
* t = input and remainder
* q = integer part
* u = divisor
*/
#define FMODULO(t,q,u) \
do { \
(q) = (((t) < 0) ? ceil((t) / (u)) : floor((t) / (u))); \
if ((q) != 0) (t) -= rint((q) * (u)); \
} while(0)
/* TMODULO()
* Like FMODULO(), but work on the timestamp datatype (now always int64).
* We assume that int64 follows the C99 semantics for division (negative
* quotients truncate towards zero).
*/
#define TMODULO(t,q,u) \
do { \
(q) = ((t) / (u)); \
if ((q) != 0) (t) -= ((q) * (u)); \
} while(0)
/*
* Date/time validation
* Include check for leap year.
*/
extern PGDLLIMPORT const char *const months[]; /* months (3-char
* abbreviations) */
extern PGDLLIMPORT const char *const days[]; /* days (full names) */
extern PGDLLIMPORT const int day_tab[2][13];
/*
* These are the rules for the Gregorian calendar, which was adopted in 1582.
* However, we use this calculation for all prior years as well because the
* SQL standard specifies use of the Gregorian calendar. This prevents the
* date 1500-02-29 from being stored, even though it is valid in the Julian
* calendar.
*/
#define isleap(y) (((y) % 4) == 0 && (((y) % 100) != 0 || ((y) % 400) == 0))
/*
* Datetime input parsing routines (ParseDateTime, DecodeDateTime, etc)
* return zero or a positive value on success. On failure, they return
* one of these negative code values. DateTimeParseError may be used to
* produce a suitable error report. For some of these codes,
* DateTimeParseError requires additional information, which is carried
* in struct DateTimeErrorExtra.
*/
#define DTERR_BAD_FORMAT (-1)
#define DTERR_FIELD_OVERFLOW (-2)
#define DTERR_MD_FIELD_OVERFLOW (-3) /* triggers hint about DateStyle */
#define DTERR_INTERVAL_OVERFLOW (-4)
#define DTERR_TZDISP_OVERFLOW (-5)
#define DTERR_BAD_TIMEZONE (-6)
#define DTERR_BAD_ZONE_ABBREV (-7)
typedef struct DateTimeErrorExtra
{
/* Needed for DTERR_BAD_TIMEZONE and DTERR_BAD_ZONE_ABBREV: */
const char *dtee_timezone; /* incorrect time zone name */
/* Needed for DTERR_BAD_ZONE_ABBREV: */
const char *dtee_abbrev; /* relevant time zone abbreviation */
} DateTimeErrorExtra;
/* Result codes for DecodeTimezoneName() */
#define TZNAME_FIXED_OFFSET 0
#define TZNAME_DYNTZ 1
#define TZNAME_ZONE 2
extern void GetCurrentDateTime(struct pg_tm *tm);
extern void GetCurrentTimeUsec(struct pg_tm *tm, fsec_t *fsec, int *tzp);
extern void j2date(int jd, int *year, int *month, int *day);
extern int date2j(int year, int month, int day);
extern int ParseDateTime(const char *timestr, char *workbuf, size_t buflen,
char **field, int *ftype,
int maxfields, int *numfields);
extern int DecodeDateTime(char **field, int *ftype, int nf,
int *dtype, struct pg_tm *tm, fsec_t *fsec, int *tzp,
DateTimeErrorExtra *extra);
extern int DecodeTimezone(const char *str, int *tzp);
extern int DecodeTimeOnly(char **field, int *ftype, int nf,
int *dtype, struct pg_tm *tm, fsec_t *fsec, int *tzp,
DateTimeErrorExtra *extra);
extern int DecodeInterval(char **field, int *ftype, int nf, int range,
int *dtype, struct pg_itm_in *itm_in);
extern int DecodeISO8601Interval(char *str,
int *dtype, struct pg_itm_in *itm_in);
extern void DateTimeParseError(int dterr, DateTimeErrorExtra *extra,
const char *str, const char *datatype,
struct Node *escontext);
extern int DetermineTimeZoneOffset(struct pg_tm *tm, pg_tz *tzp);
extern int DetermineTimeZoneAbbrevOffset(struct pg_tm *tm, const char *abbr, pg_tz *tzp);
extern int DetermineTimeZoneAbbrevOffsetTS(TimestampTz ts, const char *abbr,
pg_tz *tzp, int *isdst);
extern void EncodeDateOnly(struct pg_tm *tm, int style, char *str);
extern void EncodeTimeOnly(struct pg_tm *tm, fsec_t fsec, bool print_tz, int tz, int style, char *str);
extern void EncodeDateTime(struct pg_tm *tm, fsec_t fsec, bool print_tz, int tz, const char *tzn, int style, char *str);
extern void EncodeInterval(struct pg_itm *itm, int style, char *str);
extern void EncodeSpecialTimestamp(Timestamp dt, char *str);
extern int ValidateDate(int fmask, bool isjulian, bool is2digits, bool bc,
struct pg_tm *tm);
extern int DecodeTimezoneAbbrev(int field, const char *lowtoken,
int *ftype, int *offset, pg_tz **tz,
DateTimeErrorExtra *extra);
extern int DecodeSpecial(int field, const char *lowtoken, int *val);
extern int DecodeUnits(int field, const char *lowtoken, int *val);
extern int DecodeTimezoneName(const char *tzname, int *offset, pg_tz **tz);
extern pg_tz *DecodeTimezoneNameToTz(const char *tzname);
extern int DecodeTimezoneAbbrevPrefix(const char *str,
int *offset, pg_tz **tz);
extern void ClearTimeZoneAbbrevCache(void);
extern int j2day(int date);
extern struct Node *TemporalSimplify(int32 max_precis, struct Node *node);
extern bool CheckDateTokenTables(void);
extern TimeZoneAbbrevTable *ConvertTimeZoneAbbrevs(struct tzEntry *abbrevs,
int n);
extern void InstallTimeZoneAbbrevs(TimeZoneAbbrevTable *tbl);
extern bool AdjustTimestampForTypmod(Timestamp *time, int32 typmod,
struct Node *escontext);
#endif /* DATETIME_H */
View Git Blame Copy Permalink