- tz - lists.iana.org

defensive value for define on Solaris causes load of timerule to fail
by Kees Dekker June 12, 2017

June 12, 2017

Hi, The FILENAME_MAX macro, as used in the local_storage union in localtime.c, has on Solaris the very low/defensive value of 14 (from stdio.h). Please consider using PATH_MAX, from limits.h instead, as that value is (usually) 250 or beyond. Or anything else, that can hold reasonable lengths of file names. Loading zonefiles now fails on Solaris if a filename exceeds the used macro. Regards, Kees

2 2

Re: [tz] [PROPOSED] Clean up configuration of tzname etc.
by vanadovv＠hetnet.nl June 9, 2017

June 9, 2017

I found a small language error. Sorry I can't produce a diff/patch. The error can be found as follows: >----Origineel Bericht---- >Van : eggert(a)cs.ucla.edu >Datum : 09/06/2017 02:50 >Aan : tz(a)iana.org >Onderwerp : [tz] [PROPOSED] Clean up configuration of tzname etc. > ...... ...... >diff --git a/strftime.c b/strftime.c >index 4b9fd2c..d09a0db 100644 >--- a/strftime.c >+++ b/strftime.c ....... >@@ -513,20 +509,20 @@ label: > ** altzone and timezone), and the > ** magic might not have the correct > ** offset. Doing things correctly is >- ** tricky and requires disobeying C99; >+ ** tricky and requires disobeying the standard; > ** see GNU C strftime for details. > ** For now, punt and conform to the > ** standard, even though it's incorrect. > ** >- ** C99 says that %z must be replaced by the >- ** empty string if the time zone is not >+ ** C99 and later that %z must be replaced by >+ ** the empty string if the time zone is not ..... The error is: >+ ** C99 and later that %z must be replaced by This should be: >+ ** C99 and later say that %z must be replaced by --------------- Yours, Oscar van Vlijmen [ends]

2 1

1 s error in America/Adak and America/Nome
by Michael H Deckers June 9, 2017

June 9, 2017

The switch from 1867-10-18 to 1867-10-17 in Alaska is incorrectly represented as a jump from 1867-10-18 to 1867-10-17 + 00:00:01 h in the time scales for Adak and for Nome. Proposed corrections: -Zone America/Adak 12:13:21 - LMT 1867 Oct 18 +Zone America/Adak 12:13:22 - LMT 1867 Oct 18 -11:46:38 - LMT 1900 Aug 20 12:00 -Zone America/Nome 12:58:21 - LMT 1867 Oct 18 +Zone America/Nome 12:58:22 - LMT 1867 Oct 18 -11:01:38 - LMT 1900 Aug 20 12:00 Michael Deckers.

2 1

[PROPOSED] Clean up configuration of tzname etc.
by Paul Eggert June 9, 2017

June 9, 2017

This documents a bit of software archaeology better. tzname is standardized by POSIX, whereas timezone and daylight are standardized by POSIX SI, so allow them to be configured separately. * Makefile, NEWS, Theory: Document this. * date.c, strftime.c, zdump.c (tzname): Remove decl; private.h declares tzname as needed now. * localtime.c (tzname, timezone, daylight, altzone): Also define if TZ_TIME_T. However, define only if the relevant other macro (HAVE_TZNAME, USG_COMPAT, ALTZONE) is set or defined. (update_tzname_etc, settzname): Set tzname only if HAVE_TZNAME. At the end, simplify ifdef by using TZ_TIME_T. * newctime.3: Document optional vars and fields more carefully. * private.h (USG_COMPAT): Set default from _XOPEN_VERSION. All uses changed to assume it is 1 or 0. (HAVE_TZNAME): Set default from that and from _POSIX_VERSION. (TZ_TIME_T): New macro. (daylight, timezone, tzname, altzone): Rename if needed to avoid collision with standard library. (tzname): Declare if (!HAVE_POSIX_DECLS || TZ_TIME_T) && HAVE_TZNAME. (timezone, daylight): Declare if (!HAVE_POSIX_DECLS || TZ_TIME_T) && USG_COMPAT, not if not defined and !HAVE_POSIX_DECLS. (altzone): Declare even if not defined. (tzsetwall, offtime, timegm, timelocal, timeoff, time2posix) (posix2time, posix2time_z, time2posix_z): Use TZ_TIME_T instead of (defined time_tz). * strftime.c (_fmt): * zdump.c (abbr): Use tzname only if HAVE_TZNAME. --- Makefile | 37 +++++++++++++++++++++++---------- NEWS | 8 +++++++ Theory | 6 +++--- date.c | 1 - localtime.c | 34 ++++++++++++++++-------------- newctime.3 | 46 ++++++++++++++++++++++++++--------------- private.h | 69 ++++++++++++++++++++++++++++++++++++++++++++++--------------- strftime.c | 38 ++++++++++++++++------------------ zdump.c | 9 ++++---- 9 files changed, 160 insertions(+), 88 deletions(-) diff --git a/Makefile b/Makefile index 476f259..e6d497d 100644 --- a/Makefile +++ b/Makefile @@ -164,6 +164,7 @@ LDLIBS= # not needed by the main-program tz code, which is single-threaded. # Append other compiler flags as needed, e.g., -pthread on GNU/Linux. # -Dtime_tz=\"T\" to use T as the time_t type, rather than the system time_t +# This is intended for internal use only; it mangles external names. # -DTZ_DOMAIN=\"foo\" to use "foo" for gettext domain name; default is "tz" # -DTZ_DOMAINDIR=\"/path\" to use "/path" for gettext directory; # the default is system-supplied, typically "/usr/lib/locale" @@ -199,12 +200,6 @@ GCC_DEBUG_FLAGS = -DGCC_LINT -g3 -O3 -fno-common \ -Wno-address -Wno-format-nonliteral -Wno-sign-compare \ -Wno-type-limits -Wno-unused-parameter # -# If you want to use System V compatibility code, add -# -DUSG_COMPAT -# to the end of the "CFLAGS=" line. This arrange for "timezone" and "daylight" -# variables to be kept up-to-date by the time conversion functions. Neither -# "timezone" nor "daylight" is described in X3J11's work. -# # If your system has a "GMT offset" field in its "struct tm"s # (or if you decide to add such a field in your system's "time.h" file), # add the name to a define such as @@ -216,6 +211,31 @@ GCC_DEBUG_FLAGS = -DGCC_LINT -g3 -O3 -fno-common \ # and define NO_TM_ZONE to suppress any guessing. These two fields are not # required by POSIX, but are widely available on GNU/Linux and BSD systems. # +# The next batch of options control support for external variables +# exported by tzcode. In practice these variables are less useful +# than TM_GMTOFF and TM_ZONE. However, most of them are standardized. +# # +# # To omit or support the external variable "tzname", add one of: +# # -DHAVE_TZNAME=0 +# # -DHAVE_TZNAME=1 +# # to the "CFLAGS=" line. "tzname" is required by POSIX 1988 and later. +# # If not defined, the code attempts to guess HAVE_TZNAME from other macros. +# # Warning: unless time_tz is also defined, HAVE_TZNAME=1 can cause +# # crashes when combined with some platforms' standard libraries, +# # presumably due to memory allocation issues. +# # +# # To omit or support the external variables "timezone" and "daylight", add +# # -DUSG_COMPAT=0 +# # -DUSG_COMPAT=1 +# # to the "CFLAGS=" line; "timezone" and "daylight" are inspired by +# # Unix Systems Group code and are required by POSIX 2008 (with XSI) and later. +# # If not defined, the code attempts to guess USG_COMPAT from other macros. +# # +# # To support the external variable "altzone", add +# # -DALTZONE +# # to the end of the "CFLAGS=" line; although "altzone" appeared in +# # System V Release 3.1 it has not been standardized. +# # If you want functions that were inspired by early versions of X3J11's work, # add # -DSTD_INSPIRED @@ -253,11 +273,6 @@ GCC_DEBUG_FLAGS = -DGCC_LINT -g3 -O3 -fno-common \ # -DALL_STATE # to the end of the "CFLAGS=" line. Storage is obtained by calling malloc. # -# If you want an "altzone" variable (a la System V Release 3.1), add -# -DALTZONE -# to the end of the "CFLAGS=" line. -# This variable is not described in X3J11's work. -# # NIST-PCTS:151-2, Version 1.4, (1993-12-03) is a test suite put # out by the National Institute of Standards and Technology # which claims to test C and Posix conformance. If you want to pass PCTS, add diff --git a/NEWS b/NEWS index 99f5d64..44135ae 100644 --- a/NEWS +++ b/NEWS @@ -57,6 +57,14 @@ Unreleased, experimental changes ordinary lines in leap second input. Formerly, zic sometimes warned about this undocumented usage and handled it incorrectly. + The new macro HAVE_TZNAME governs whether the tzname external + variable is exported, instead of USG_COMPAT. USG_COMPAT now + governs only the external variables "timezone" and "daylight". + This change is needed because the three variables are not in the + same category: although POSIX requires tzname, it specifies the + other two variables as optional. Also, USG_COMPAT is now 1 or 0: + if not defined, the code attempts to guess it from other macros. + Several minor changes have been made to the code to make it a bit easier to port to MS-Windows. (Thanks to Kees Dekker for reporting the problems.) diff --git a/Theory b/Theory index 2bfadc6..8aeccb2 100644 --- a/Theory +++ b/Theory @@ -581,9 +581,9 @@ Points of interest to folks with other systems: of GMT" value and a "daylight saving time in effect" flag) to a time zone abbreviation, and we refuse to guess. Programs that in the past used the timezone function may now examine - tzname[localtime(&clock)->tm_isdst] to learn the correct time - zone abbreviation to use. Alternatively, use - localtime(&clock)->tm_zone if this has been enabled. + localtime(&clock)->tm_zone (if TM_ZONE is defined) or + tzname[localtime(&clock)->tm_isdst] (if HAVE_TZNAME is defined) + to learn the correct time zone abbreviation to use. * The 4.2BSD gettimeofday function is not used in this package. This formerly let users obtain the current UTC offset and DST flag, diff --git a/date.c b/date.c index 269825d..0d64b8a 100644 --- a/date.c +++ b/date.c @@ -45,7 +45,6 @@ #if !HAVE_POSIX_DECLS extern char * optarg; extern int optind; -extern char * tzname[]; #endif static int retval = EXIT_SUCCESS; diff --git a/localtime.c b/localtime.c index 8f103aa..b23fe43 100644 --- a/localtime.c +++ b/localtime.c @@ -188,20 +188,21 @@ static int lcl_is_set; static struct tm tm; -#if !HAVE_POSIX_DECLS +#if !HAVE_POSIX_DECLS || TZ_TIME_T +# if HAVE_TZNAME char * tzname[2] = { (char *) wildabbr, (char *) wildabbr }; -# ifdef USG_COMPAT +# endif +# if USG_COMPAT long timezone; int daylight; # endif -#endif - -#ifdef ALTZONE +# ifdef ALTZONE long altzone; -#endif /* defined ALTZONE */ +# endif +#endif /* Initialize *S to a value based on GMTOFF, ISDST, and ABBRIND. */ static void @@ -263,8 +264,10 @@ detzcode64(const char *const codep) static void update_tzname_etc(struct state const *sp, struct ttinfo const *ttisp) { +#if HAVE_TZNAME tzname[ttisp->tt_isdst] = (char *) &sp->chars[ttisp->tt_abbrind]; -#ifdef USG_COMPAT +#endif +#if USG_COMPAT if (!ttisp->tt_isdst) timezone = - ttisp->tt_gmtoff; #endif @@ -280,16 +283,17 @@ settzname(void) register struct state * const sp = lclptr; register int i; - tzname[0] = tzname[1] = (char *) wildabbr; -#ifdef USG_COMPAT +#if HAVE_TZNAME + tzname[0] = tzname[1] = (char *) (sp ? wildabbr : gmt); +#endif +#if USG_COMPAT daylight = 0; timezone = 0; -#endif /* defined USG_COMPAT */ +#endif #ifdef ALTZONE altzone = 0; #endif /* defined ALTZONE */ if (sp == NULL) { - tzname[0] = tzname[1] = (char *) gmt; return; } /* @@ -304,10 +308,10 @@ settzname(void) &sp->ttis[ sp->types[i]]; update_tzname_etc(sp, ttisp); -#ifdef USG_COMPAT +#if USG_COMPAT if (ttisp->tt_isdst) daylight = 1; -#endif /* defined USG_COMPAT */ +#endif } } @@ -2286,9 +2290,9 @@ posix2time(time_t t) #endif /* defined STD_INSPIRED */ -#if defined time_tz || EPOCH_LOCAL || EPOCH_OFFSET != 0 +#if TZ_TIME_T -# ifndef USG_COMPAT +# if !USG_COMPAT # define daylight 0 # define timezone 0 # endif diff --git a/newctime.3 b/newctime.3 index 169c919..8b89935 100644 --- a/newctime.3 +++ b/newctime.3 @@ -7,7 +7,7 @@ asctime, ctime, difftime, gmtime, localtime, mktime \- convert date and time .el ds - \- .B #include <time.h> .PP -.B extern char *tzname[2]; +.BR "extern char *tzname[];" " /\(** (optional) \(**/" .PP .B char *ctime(time_t const *clock); .PP @@ -213,7 +213,7 @@ includes the following fields: .RS .PP .nf -.ta .5i +\w'long tm_gmtoff;\0\0'u +.ta 2n +\w'long tm_gmtoff;nn'u int tm_sec; /\(** seconds (0\*(en60) \(**/ int tm_min; /\(** minutes (0\*(en59) \(**/ int tm_hour; /\(** hours (0\*(en23) \(**/ @@ -223,21 +223,11 @@ includes the following fields: int tm_wday; /\(** day of week (Sunday = 0) \(**/ int tm_yday; /\(** day of year (0\*(en365) \(**/ int tm_isdst; /\(** is summer time in effect? \(**/ - char \(**tm_zone; /\(** abbreviation of time zone name \(**/ - long tm_gmtoff; /\(** offset from UT in seconds \(**/ + char \(**tm_zone; /\(** time zone abbreviation (optional) \(**/ + long tm_gmtoff; /\(** offset from UT in seconds (optional) \(**/ .fi .RE .PP -The -.I tm_zone -and -.I tm_gmtoff -fields exist, and are filled in, only if arrangements to do -so were made when the library containing these functions was -created. -There is no guarantee that these fields will continue to exist -in this form in future releases of this code. -.PP .I Tm_isdst is non-zero if summer time is in effect. .PP @@ -246,6 +236,21 @@ is the offset (in seconds) of the time represented from UT, with positive values indicating east of the Prime Meridian. The field's name is derived from Greenwich Mean Time, a precursor of UT. +.PP +In +.B struct tm +the +.I tm_zone +and +.I tm_gmtoff +fields exist, and are filled in, only if arrangements to do +so were made when the library containing these functions was +created. +Similarly, the +.B tzname +variable is optional. +There is no guarantee that these fields and this variable will +continue to exist in this form in future releases of this code. .SH FILES .ta \w'/usr/local/etc/zoneinfo/posixrules\0\0'u /usr/local/etc/zoneinfo time zone information directory @@ -277,12 +282,19 @@ and point to static data overwritten by each call. The +.B tzname +variable (once set) and the .B tm_zone field of a returned .B "struct tm" -points to a static array of characters, which -can be overwritten by later calls to -.IR tzset . +both point to an array of characters that +can be freed or overwritten by later calls to the functions +.IR localtime , +.IR tzfree , +and +.IR tzset , +if these functions affect the time zone information that specifies the +abbreviation in question. The remaining functions and data are thread-safe. .PP .IR Asctime , diff --git a/private.h b/private.h index 1e61592..ad9fcb9 100644 --- a/private.h +++ b/private.h @@ -153,6 +153,22 @@ # endif #endif +#ifndef USG_COMPAT +# ifndef _XOPEN_VERSION +# define USG_COMPAT 0 +# else +# define USG_COMPAT 1 +# endif +#endif + +#ifndef HAVE_TZNAME +# if _POSIX_VERSION < 198808 && !USG_COMPAT +# define HAVE_TZNAME 0 +# else +# define HAVE_TZNAME 1 +# endif +#endif + #ifndef F_OK #define F_OK 0 #endif /* !defined F_OK */ @@ -339,6 +355,12 @@ typedef unsigned long uintmax_t; ** typical platforms. */ #if defined time_tz || EPOCH_LOCAL || EPOCH_OFFSET != 0 +# define TZ_TIME_T 1 +#else +# define TZ_TIME_T 0 +#endif + +#if TZ_TIME_T # ifdef LOCALTIME_IMPLEMENTATION static time_t sys_time(time_t *x) { return time(x); } # endif @@ -393,6 +415,20 @@ typedef time_tz tz_time_t; # define tzset tz_tzset # undef tzsetwall # define tzsetwall tz_tzsetwall +# if HAVE_TZNAME +# undef tzname +# define tzname tz_tzname +# endif +# if USG_COMPAT +# undef daylight +# define daylight tz_daylight +# undef timezone +# define timezone tz_timezone +# endif +# ifdef ALTZONE +# undef altzone +# define altzone tz_altzone +# endif char *ctime(time_t const *); char *ctime_r(time_t const *, char *); @@ -422,18 +458,17 @@ extern char *asctime_r(struct tm const *restrict, char *restrict); extern char **environ; #endif -#if !HAVE_POSIX_DECLS -# ifdef USG_COMPAT -# ifndef timezone +#if TZ_TIME_T || !HAVE_POSIX_DECLS +# if HAVE_TZNAME +extern char *tzname[]; +# endif +# if USG_COMPAT extern long timezone; -# endif -# ifndef daylight extern int daylight; -# endif # endif #endif -#if defined ALTZONE && !defined altzone +#ifdef ALTZONE extern long altzone; #endif @@ -443,25 +478,25 @@ extern long altzone; */ #ifdef STD_INSPIRED -# if !defined tzsetwall || defined time_tz +# if TZ_TIME_T || !defined tzsetwall void tzsetwall(void); # endif -# if !defined offtime || defined time_tz +# if TZ_TIME_T || !defined offtime struct tm *offtime(time_t const *, long); # endif -# if !defined timegm || defined time_tz +# if TZ_TIME_T || !defined timegm time_t timegm(struct tm *); # endif -# if !defined timelocal || defined time_tz +# if TZ_TIME_T || !defined timelocal time_t timelocal(struct tm *); # endif -# if !defined timeoff || defined time_tz +# if TZ_TIME_T || !defined timeoff time_t timeoff(struct tm *, long); # endif -# if !defined time2posix || defined time_tz +# if TZ_TIME_T || !defined time2posix time_t time2posix(time_t); # endif -# if !defined posix2time || defined time_tz +# if TZ_TIME_T || !defined posix2time time_t posix2time(time_t); # endif #endif @@ -495,10 +530,10 @@ time_t mktime_z(timezone_t restrict, struct tm *restrict); timezone_t tzalloc(char const *); void tzfree(timezone_t); # ifdef STD_INSPIRED -# if !defined posix2time_z || defined time_tz +# if TZ_TIME_T || !defined posix2time_z time_t posix2time_z(timezone_t, time_t) ATTRIBUTE_PURE; # endif -# if !defined time2posix_z || defined time_tz +# if TZ_TIME_T || !defined time2posix_z time_t time2posix_z(timezone_t, time_t) ATTRIBUTE_PURE; # endif # endif @@ -649,7 +684,7 @@ char *ctime_r(time_t const *, char *); ** or ** isleap(a + b) == isleap(a % 400 + b % 400) ** This is true even if % means modulo rather than Fortran remainder -** (which is allowed by C89 but not C99). +** (which is allowed by C89 but not by C99 or later). ** We use this to avoid addition overflow problems. */ diff --git a/strftime.c b/strftime.c index 4b9fd2c..d09a0db 100644 --- a/strftime.c +++ b/strftime.c @@ -73,7 +73,7 @@ static const struct lc_time_T C_time_locale = { /* ** x_fmt - ** C99 requires this format. + ** C99 and later require this format. ** Using just numbers (as here) makes Quakers happier; ** it's also compatible with SVR4. */ @@ -81,7 +81,7 @@ static const struct lc_time_T C_time_locale = { /* ** c_fmt - ** C99 requires this format. + ** C99 and later require this format. ** Previously this code used "%D %X", but we now conform to C99. ** Note that ** "%a %b %d %H:%M:%S %Y" @@ -105,10 +105,6 @@ static char * _fmt(const char *, const struct tm *, char *, const char *, int *); static char * _yconv(int, int, bool, bool, char *, char const *); -#if !HAVE_POSIX_DECLS -extern char * tzname[]; -#endif - #ifndef YEAR_2000_NAME #define YEAR_2000_NAME "CHECK_STRFTIME_FORMATS_FOR_TWO_DIGIT_YEARS" #endif /* !defined YEAR_2000_NAME */ @@ -223,7 +219,7 @@ label: case 'E': case 'O': /* - ** C99 locale modifiers. + ** Locale modifiers of C99 and later. ** The sequences ** %Ec %EC %Ex %EX %Ey %EY ** %Od %oe %OH %OI %Om %OM @@ -483,19 +479,19 @@ label: case 'Z': #ifdef TM_ZONE pt = _add(t->TM_ZONE, pt, ptlim); -#else +#elif HAVE_TZNAME if (t->tm_isdst >= 0) pt = _add(tzname[t->tm_isdst != 0], pt, ptlim); #endif /* - ** C99 says that %Z must be replaced by the - ** empty string if the time zone is not - ** determinable. + ** C99 and later say that %Z must be + ** replaced by the empty string if the + ** time zone is not determinable. */ continue; case 'z': -#if defined TM_GMTOFF || defined USG_COMPAT || defined ALTZONE +#if defined TM_GMTOFF || USG_COMPAT || defined ALTZONE { long diff; char const * sign; @@ -505,7 +501,7 @@ label: diff = t->TM_GMTOFF; # else /* - ** C99 says that the UT offset must + ** C99 and later say that the UT offset must ** be computed by looking only at ** tm_isdst. This requirement is ** incorrect, since it means the code @@ -513,20 +509,20 @@ label: ** altzone and timezone), and the ** magic might not have the correct ** offset. Doing things correctly is - ** tricky and requires disobeying C99; + ** tricky and requires disobeying the standard; ** see GNU C strftime for details. ** For now, punt and conform to the ** standard, even though it's incorrect. ** - ** C99 says that %z must be replaced by the - ** empty string if the time zone is not + ** C99 and later that %z must be replaced by + ** the empty string if the time zone is not ** determinable, so output nothing if the ** appropriate variables are not available. */ if (t->tm_isdst < 0) continue; if (t->tm_isdst == 0) -# ifdef USG_COMPAT +# if USG_COMPAT diff = -timezone; # else continue; @@ -543,9 +539,11 @@ label: #ifdef TM_ZONE negative = t->TM_ZONE[0] == '-'; #else - negative - = (t->tm_isdst < 0 - || tzname[t->tm_isdst != 0][0] == '-'); + negative = t->tm_isdst < 0; +# if HAVE_TZNAME + if (tzname[t->tm_isdst != 0][0] == '-') + negative = true; +# endif #endif } if (negative) { diff --git a/zdump.c b/zdump.c index 7cde3d3..b69b1e9 100644 --- a/zdump.c +++ b/zdump.c @@ -77,7 +77,6 @@ extern int getopt(int argc, char * const argv[], const char * options); extern char * optarg; extern int optind; -extern char * tzname[]; #endif /* The minimum and maximum finite time values. */ @@ -997,9 +996,11 @@ abbr(struct tm const *tmp) #ifdef TM_ZONE return tmp->TM_ZONE; #else - return (0 <= tmp->tm_isdst && tzname[0 < tmp->tm_isdst] - ? tzname[0 < tmp->tm_isdst] - : ""); +# if HAVE_TZNAME + if (0 <= tmp->tm_isdst && tzname[0 < tmp->tm_isdst]) + return tzname[0 < tmp->tm_isdst]; +# endif + return ""; #endif } -- 2.9.4

1 0

[PATCH 1/2] * Makefile (check_character_set): Fix typo in filename.
by Paul Eggert June 8, 2017

June 8, 2017

--- Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Makefile b/Makefile index c3ef931..476f259 100644 --- a/Makefile +++ b/Makefile @@ -605,7 +605,7 @@ check_character_set: $(ENCHILADA) ! grep -Env $(SAFE_LINE) $(MANS) date.1 $(MANTXTS) \ $(MISC) $(SOURCES) $(WEB_PAGES) \ CONTRIBUTING LICENSE Makefile README \ - version tzdata.ic && \ + version tzdata.zi && \ ! grep -Env $(SAFE_SHARP_LINE) $(TDATA) backzone \ leapseconds yearistype.sh zone.tab && \ ! grep -Env $(OK_LINE) $(ENCHILADA) -- 2.7.4

1 1

[PATCH 1/2] Fix zic parsing of abbreviated line codes
by Paul Eggert June 2, 2017

June 2, 2017

* NEWS: Document this. * zic.8: Document more clearly that "Zone" etc. can be abbreviated. * zic.c (line_codes): Remove, replacing with ... (zi_line_codes, leap_line_codes): ... these new constants. (infile): Use them to distinguish context more accurately. Remove no-longer-applicable warning. --- NEWS | 9 +++++++++ zic.8 | 9 ++++++++- zic.c | 16 +++++++++------- 3 files changed, 26 insertions(+), 8 deletions(-) diff --git a/NEWS b/NEWS index 0a76781..99f5d64 100644 --- a/NEWS +++ b/NEWS @@ -48,6 +48,15 @@ Unreleased, experimental changes Also, zic warns about the undocumented usage with a "last-" prefix, e.g., "last-Fri". + Similarly, zic now accepts the unambiguous abbreviation "L" for + "Link" in ordinary context and for "Leap" in leap-second context. + Conversely, zic no longer accepts non-prefixes such as "La" as + abbreviations for words like "Leap". + + zic no longer accepts leap second lines in ordinary input, or + ordinary lines in leap second input. Formerly, zic sometimes + warned about this undocumented usage and handled it incorrectly. + Several minor changes have been made to the code to make it a bit easier to port to MS-Windows. (Thanks to Kees Dekker for reporting the problems.) diff --git a/zic.8 b/zic.8 index 6fee96d..ab95b08 100644 --- a/zic.8 +++ b/zic.8 @@ -147,7 +147,14 @@ Any line that is blank (after comment stripping) is ignored. Non-blank lines are expected to be of one of three types: rule lines, zone lines, and link lines. .PP -Names (such as month names) must be in English and are case insensitive. +Names must be in English and are case insensitive. +They appear in several contexts, and include month and weekday names +and keywords such as +.BR "maximum" , +.BR "only" , +.BR "Rolling" , +and +.BR "Zone" . A name can be abbreviated by omitting all but an initial prefix; any abbreviation must be unambiguous in context. .PP diff --git a/zic.c b/zic.c index 33cb4e7..ff71af6 100644 --- a/zic.c +++ b/zic.c @@ -298,10 +298,13 @@ struct lookup { static struct lookup const * byword(const char * string, const struct lookup * lp); -static struct lookup const line_codes[] = { +static struct lookup const zi_line_codes[] = { { "Rule", LC_RULE }, { "Zone", LC_ZONE }, { "Link", LC_LINK }, + { NULL, 0 } +}; +static struct lookup const leap_line_codes[] = { { "Leap", LC_LEAP }, { NULL, 0} }; @@ -1114,6 +1117,8 @@ infile(const char *name) } else if (wantcont) { wantcont = inzcont(fields, nfields); } else { + struct lookup const *line_codes + = name == leapsec ? leap_line_codes : zi_line_codes; lp = byword(fields[0], line_codes); if (lp == NULL) error(_("input line of unknown type")); @@ -1130,11 +1135,7 @@ infile(const char *name) wantcont = false; break; case LC_LEAP: - if (name != leapsec) - warning(_("%s: Leap line in non leap" - " seconds file %s"), - progname, name); - else inleap(fields, nfields); + inleap(fields, nfields); wantcont = false; break; default: /* "cannot happen" */ @@ -1586,7 +1587,8 @@ rulesub(struct rule *rp, const char *loyearp, const char *hiyearp, ** Day work. ** Accept things such as: ** 1 - ** last-Sunday + ** lastSunday + ** last-Sunday (undocumented; warn about this) ** Sun<=20 ** Sun>=7 */ -- 2.9.4

1 2

mend historical data of Europe/Dublin
by Michael Deckers June 1, 2017

June 1, 2017

Europe/Dublin currently has the two fall backs in 1946 and 1947 going from 02:00 to 01:00, one hour before the corresponding two jumps in Europe/London. This is wrong; both should go from 03:00 to 02:00. For 1947, the "Summer Time Order, 1947" at [http://www.irishstatutebook.ie/eli/1947/sro/71/made/en/print.html] gives the time of the switch as "two o'clock, West-European time", where "West-European time" means GMT according to the "Summer Time Act, 1925", section 4, at [http://www.irishstatutebook.ie/eli/1925/act/8/enacted/en/print.html] For 1946, the "Summer Time Order, 1926" at [http://www.irishstatutebook.ie/eli/1926/sro/919/made/en/print.html] specifies the time of the switch in the same manner. Hence I propose to change Zone Europe/Dublin as follows: # Zone NAME GMTOFF RULES FORMAT [UNTIL] - 0:00 1:00 IST 1946 Oct 6 2:00 + 0:00 1:00 IST 1946 Oct 6 2:00s 0:00 - GMT 1947 Mar 16 2:00 - 0:00 1:00 IST 1947 Nov 2 2:00 + 0:00 1:00 IST 1947 Nov 2 2:00s 0:00 - GMT 1948 Apr 18 2:00 Michael Deckers. --- This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

2 1

proposed changes for Win32 and a improved mktime() algorithm
by Kees Dekker May 30, 2017

May 30, 2017

Hello, I've a patch (diff/patch format) attached, which is the result of a longer period of testing (and delays due to other duties). The patch has been created against 2017B code. And I've tried to limit my changes to a minimum. The patch comprises changes for 5 files, where most of the changes are in private.h and localtime.c. The changes are (summarized list): 1. Win32/Visual Studio (2015) port. Related to including header files + handling of the Windows style backslash (by adding some macros) 2. Adding a faster (and more efficient) algorithm for mktime(). Instead of the binary search another approach is used, but also results in different ambiguous boundaries for switching from DST to non-DST. 3. Adding a possibility not to use putenv(). Putenv() may result in leaking memory, as it is unsure putenv() behaves on different platforms. As long as set & get of a TZ variable is within the TZ library, there is no need to use putenv()/getenv() at all. Also note that putenv/getenv on Windows are thread-unsafe (and I did not like to use the GetEnvironmentString/SetEnvironmentString() counterparts). 4. Solved some platform (e.g. HPUX) specific errors/warnings (initializing + some fixed char buffers). Indeed, making some compilers happy (there is more than gcc only). I'm not sure whether the way how I provide this patch is the recommended way. If another way is preferred, please let me know. If more information is needed, please ask. I added as much as possible the comments in the affected source files itself, but I can imagine that still some questions left. It is not my intention to see 'good luck with this patch'. If there is another preferred method to chop this patch in chunks, please let me know. Regards, Kees

11 43

[PROPOSED] New file tzdata.zi for text-format tzdata
by Paul Eggert May 26, 2017

May 26, 2017

* NEWS: Document this. * Makefile (TZDATA_TEXT, TZDATA_ZI_DEPS): New macros. (TABDATA): Use them. (DATA): Redo to not include tzdata.zi, since it is distributed as part of tzdb, not tzdata. (AWK_SCRIPTS): Add zishrink.awk. (ENCHILADA): Add tzdata.zi. (install): Install tzdata.zi and leapseconds, if requested by TZDATA_TEXT. (tzdata.zi): New rule. (install_data, $(TZS_NEW), zonenames): Use tzdata.zi instead of its inputs. (check_character_set, check_links): Check tzdata.zi. (clean): Remove tzdata.zi. (set-timestamps.out): Set tzdata.zi’s time stamp. (check_public): Check that tzdata.zi can be compiled. * zishrink.awk: New file. --- Makefile | 58 +++++++++++++++++++++++++------------- NEWS | 19 ++++++++++--- zishrink.awk | 91 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 145 insertions(+), 23 deletions(-) create mode 100644 zishrink.awk diff --git a/Makefile b/Makefile index 6373920..e037b7c 100644 --- a/Makefile +++ b/Makefile @@ -92,6 +92,17 @@ LIBDIR= $(TOPDIR)/lib REDO= posix_right +# To install data in text form that has all the information of the binary data, +# (optionally incorporating leap second information), use +# TZDATA_TEXT= tzdata.zi leapseconds +# To install text data without leap second information (e.g., because +# REDO='posix_only'), use +# TZDATA_TEXT= tzdata.zi +# To avoid installing text data, use +# TZDATA_TEXT= + +TZDATA_TEXT= leapseconds tzdata.zi + # For backward-compatibility links for old zone names, use # BACKWARD= backward pacificnew # To omit these links, use @@ -395,18 +406,19 @@ YDATA= $(PRIMARY_YDATA) etcetera $(BACKWARD) NDATA= systemv factory TDATA= $(YDATA) $(NDATA) ZONETABLES= zone1970.tab zone.tab -TABDATA= iso3166.tab leapseconds $(ZONETABLES) +TABDATA= iso3166.tab $(TZDATA_TEXT) $(ZONETABLES) LEAP_DEPS= leapseconds.awk leap-seconds.list -DATA= $(YDATA) $(NDATA) backzone $(TABDATA) \ - leap-seconds.list yearistype.sh -AWK_SCRIPTS= checklinks.awk checktab.awk leapseconds.awk +TZDATA_ZI_DEPS= zishrink.awk $(TDATA) $(PACKRATDATA) +DATA= $(YDATA) $(NDATA) backzone iso3166.tab leap-seconds.list \ + leapseconds yearistype.sh $(ZONETABLES) +AWK_SCRIPTS= checklinks.awk checktab.awk leapseconds.awk zishrink.awk MISC= $(AWK_SCRIPTS) zoneinfo2tdf.pl TZS_YEAR= 2050 TZS= to$(TZS_YEAR).tzs TZS_NEW= to$(TZS_YEAR)new.tzs TZS_DEPS= $(PRIMARY_YDATA) asctime.c localtime.c \ private.h tzfile.h zdump.c zic.c -ENCHILADA= $(COMMON) $(DOCS) $(SOURCES) $(DATA) $(MISC) $(TZS) +ENCHILADA= $(COMMON) $(DOCS) $(SOURCES) $(DATA) $(MISC) $(TZS) tzdata.zi # Consult these files when deciding whether to rebuild the 'version' file. # This list is not the same as the output of 'git ls-files', since @@ -443,7 +455,7 @@ install: all $(DATA) $(REDO) $(MANS) $(DESTDIR)$(MANDIR)/man3 $(DESTDIR)$(MANDIR)/man5 \ $(DESTDIR)$(MANDIR)/man8 $(ZIC_INSTALL) -l $(LOCALTIME) -p $(POSIXRULES) - cp -f iso3166.tab $(ZONETABLES) $(DESTDIR)$(TZDIR)/. + cp -f $(TABDATA) $(DESTDIR)$(TZDIR)/. cp tzselect zic zdump $(DESTDIR)$(ETCDIR)/. cp libtz.a $(DESTDIR)$(LIBDIR)/. $(RANLIB) $(DESTDIR)$(LIBDIR)/libtz.a @@ -464,6 +476,13 @@ version: $(VERSION_DEPS) printf '%s\n' "$$V" >$@.out mv $@.out $@ +# This file can be tailored by setting BACKWARD, PACKRATDATA, etc. +tzdata.zi: $(TZDATA_ZI_DEPS) + $(AWK) -v PACKRATDATA='$(PACKRATDATA)' \ + -f zishrink.awk \ + $(TDATA) $(PACKRATDATA) >$@.out + mv $@.out $@ + version.h: version VERSION=`cat version` && printf '%s\n' \ 'static char const PKGVERSION[]="($(PACKAGE)) ";' \ @@ -498,10 +517,8 @@ INSTALLARGS = \ ZIC='$(ZIC)' # 'make install_data' installs one set of tz binary files. -# It can be tailored by setting LEAPSECONDS, PACKRATDATA, etc. -install_data: zic leapseconds yearistype $(PACKRATDATA) $(TDATA) - $(ZIC_INSTALL) $(TDATA) - $(AWK) '/^Rule/' $(TDATA) | $(ZIC_INSTALL) - $(PACKRATDATA) +install_data: zic leapseconds yearistype tzdata.zi + $(ZIC_INSTALL) tzdata.zi posix_only: $(MAKE) $(INSTALLARGS) LEAPSECONDS= install_data @@ -538,14 +555,14 @@ posix_packrat: zones: $(REDO) -$(TZS_NEW): $(TDATA) zdump zic +$(TZS_NEW): tzdata.zi zdump zic mkdir -p tzs.dir - $(zic) -d tzs.dir $(TDATA) + $(zic) -d tzs.dir tzdata.zi $(AWK) '/^Link/{print $$1 "\t" $$2 "\t" $$3}' \ - $(TDATA) | LC_ALL=C sort >$@.out + tzdata.zi | LC_ALL=C sort >$@.out wd=`pwd` && \ zones=`$(AWK) -v wd="$$wd" \ - '/^Zone/{print wd "/tzs.dir/" $$2}' $(TDATA) \ + '/^Zone/{print wd "/tzs.dir/" $$2}' tzdata.zi \ | LC_ALL=C sort` && \ ./zdump -i -c $(TZS_YEAR) $$zones >>$@.out sed 's,^TZ=".*tzs\.dir/,TZ=",' $@.out >$@.sed.out @@ -589,7 +606,8 @@ check_character_set: $(ENCHILADA) sharp='#' && \ ! grep -Env $(SAFE_LINE) $(MANS) date.1 $(MANTXTS) \ $(MISC) $(SOURCES) $(WEB_PAGES) \ - CONTRIBUTING LICENSE Makefile README version && \ + CONTRIBUTING LICENSE Makefile README \ + version tzdata.ic && \ ! grep -Env $(SAFE_SHARP_LINE) $(TDATA) backzone \ leapseconds yearistype.sh zone.tab && \ ! grep -Env $(OK_LINE) $(ENCHILADA) @@ -613,6 +631,7 @@ check_sorted: backward backzone iso3166.tab zone.tab zone1970.tab check_links: checklinks.awk $(TDATA) $(AWK) -f checklinks.awk $(TDATA) + $(AWK) -f checklinks.awk tzdata.zi check_tables: checktab.awk $(PRIMARY_YDATA) $(ZONETABLES) for tab in $(ZONETABLES); do \ @@ -630,7 +649,7 @@ clean_misc: rm -f core *.o *.out \ date tzselect version.h zdump zic yearistype libtz.a clean: clean_misc - rm -fr *.dir tzdb-*/ $(TZS_NEW) + rm -fr *.dir tzdata.zi tzdb-*/ $(TZS_NEW) maintainer-clean: clean @echo 'This command is intended for maintainers to use; it' @@ -684,6 +703,7 @@ set-timestamps.out: $(ENCHILADA) touch -cmr `ls -t $$file workman.sh | sed 1q` $$file.txt || \ exit; \ done + touch -cmr `ls -t $(TZDATA_ZI_DEPS) | sed 1q` tzdata.zi touch -cmr `ls -t $(TZS_DEPS) | sed 1q` $(TZS) touch -cmr `ls -t $(VERSION_DEPS) | sed 1q` version touch $@ @@ -695,7 +715,7 @@ check_public: $(MAKE) maintainer-clean $(MAKE) "CFLAGS=$(GCC_DEBUG_FLAGS)" ALL mkdir -p public.dir - for i in $(TDATA) ; do \ + for i in $(TDATA) tzdata.zi; do \ $(zic) -v -d public.dir $$i 2>&1 || exit; \ done $(zic) -v -d public.dir $(TDATA) @@ -789,8 +809,8 @@ typecheck: $(MAKE) clean ; \ done -zonenames: $(TDATA) - @$(AWK) '/^Zone/ { print $$2 } /^Link/ { print $$3 }' $(TDATA) +zonenames: tzdata.zi + @$(AWK) '/^Zone/ { print $$2 } /^Link/ { print $$3 }' tzdata.zi asctime.o: private.h tzfile.h date.o: private.h diff --git a/NEWS b/NEWS index d22d561..8c81caa 100644 --- a/NEWS +++ b/NEWS @@ -10,6 +10,21 @@ Unreleased, experimental changes Add 7 s to the UT offset in Asia/Yangon before 1920. + Changes to build procedure + + To support applications that prefer to read time zone data in text + form, two zic input files tzdata.zi and leapseconds are now + installed by default. The commands 'zic tzdata.zi' and 'zic -L + leapseconds tzdata.zi' can reproduce the tzdata binary files + without and with leap seconds, respectively. To prevent these two + new files from being installed, use 'make TZDATA_TEXT=', and to + suppress leap seconds from the tzdata text installation, use 'make + TZDATA_TEXT=tzdata.zi'. + + 'make BACKWARD=' now suppresses backward-compatibility names + like 'US/Pacific' that are defined in the 'backward' and + 'pacificnew' files. + Changes to code zic and the reference runtime now reject multiple leap seconds @@ -27,10 +42,6 @@ Unreleased, experimental changes Also, zic warns about the undocumented usage with a "last-" prefix, e.g., "last-Fri". - 'make BACKWARD=' now suppresses backward-compatibility names - like 'US/Pacific' that are defined in the 'backward' and - 'pacificnew' files. - Several minor changes have been made to the code to make it a bit easier to port to MS-Windows. (Thanks to Kees Dekker for reporting the problems.) diff --git a/zishrink.awk b/zishrink.awk new file mode 100644 index 0000000..42240f1 --- /dev/null +++ b/zishrink.awk @@ -0,0 +1,91 @@ +# Convert tzdata source into a smaller version of itself. + +# Contributed by Paul Eggert. This file is in the public domain. + +# This is not a general-purpose converter; it is designed for current tzdata. +# 'zic' should treat this script's output as if it were identical to +# this script's input. + +BEGIN { + print "# This zic input file is in the public domain." + if (PACKRATDATA) { + while (0 < (getline line <PACKRATDATA)) { + if (split(line, field)) { + if (field[1] == "Zone") packrat_zone[field[2]] = 1 + if (field[1] == "Link") packrat_zone[field[3]] = 1 + } + } + close (PACKRATDATA) + } +} + +# Remove comments, normalize spaces, and append a space to each line. +/^[[:space:]]*[^#[:space:]]/ { + line = $0 + sub(/#.*/, "", line) + line = line " " + gsub(/[[:space:]]+/, " ", line) + + # SystemV rules are not needed. + if (line ~ /^Rule SystemV /) next + + # Replace FooAsia rules with the same rules without "Asia", as they + # are duplicates. + if (n = match(line, /[^ ]Asia /)) { + if (line ~ /^Rule /) next + line = substr(line, 1, n) substr(line, n + 5) + } + + # Abbreviate times. + while (n = match(line, /[: ]0+[0-9]/)) { + line = substr(line, 1, n) substr(line, n + RLENGTH - 1) + } + while (n = match(line, /:0[^:]/)) { + line = substr(line, 1, n - 1) substr(line, n + 2) + } + + # Abbreviate weekday names. Do not abbreviate "Sun" and "Sat", as + # pre-2017c zic erroneously diagnoses "Su" and "Sa" as ambiguous. + while (n = match(line, /[ l]Mon[<>]/)) { + line = substr(line, 1, n + 1) substr(line, n + 4) + } + while (n = match(line, /[ l]Tue[<>]/)) { + line = substr(line, 1, n + 2) substr(line, n + 4) + } + while (n = match(line, /[ l]Wed[<>]/)) { + line = substr(line, 1, n + 1) substr(line, n + 4) + } + while (n = match(line, /[ l]Thu[<>]/)) { + line = substr(line, 1, n + 2) substr(line, n + 4) + } + while (n = match(line, /[ l]Fri[<>]/)) { + line = substr(line, 1, n + 1) substr(line, n + 4) + } + + # Abbreviate "only" and month names. + gsub(/ only /, " o ", line) + gsub(/ Jan /, " Ja ", line) + gsub(/ Feb /, " F ", line) + gsub(/ Apr /, " Ap ", line) + gsub(/ Aug /, " Au ", line) + gsub(/ Sep /, " S ", line) + gsub(/ Oct /, " O ", line) + gsub(/ Nov /, " N ", line) + gsub(/ Dec /, " D ", line) + + # Strip leading and trailing space. + sub(/^ /, "", line) + sub(/ $/, "", line) + + # Remove unnecessary trailing zero fields. + sub(/ 0+$/, "", line) + + # Output lines unless they are later overridden in PACKRATDATA. + if (FILENAME != PACKRATDATA && line ~ /^[LZ]/) { + overridden = 0 + split(line, field) + overridden = packrat_zone[field[2 + (field[1] == "Link")]] + } + if (!overridden) + print line +} -- 2.9.4

2 4

[PATCH] Shorten an overlong abbrevation in 'backzone'
by Paul Eggert May 26, 2017

May 26, 2017

* backzone (Africa/Luanda): Use 'LMT' for what was apparently Luanda Mean Time. '+005204' is so long that it now provokes a zic warning. --- backzone | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/backzone b/backzone index 9ce7831..a630129 100644 --- a/backzone +++ b/backzone @@ -181,7 +181,7 @@ Zone Africa/Lome 0:04:52 - LMT 1893 # with the date that it took effect, namely 1912-01-01. # Zone Africa/Luanda 0:52:56 - LMT 1892 - 0:52:04 - +005204 1912 Jan 1 + 0:52:04 - LMT 1912 Jan 1 # Luanda Mean Time? 1:00 - WAT # Democratic Republic of the Congo (east) -- 2.9.4

1 0

U.S. DST causes an estimated 30 traffic-accident deaths/year
by Paul Eggert May 25, 2017

May 25, 2017

Austin Smith of Miami University analyzed data from the U.S. Fatality Analysis Reporting System both before and after the 2007 daylight-saving rule change, and found that the spring-forward transition increased fatal crash risk by 5% to 6.5%, with four statistical tests suggesting that the increase is explained by sleep deprivation. Smith concludes, "Back-of-the-envelope calculations suggest that the spring transition into DST caused over 30 deaths annually at a social cost of $275 million." I assume this estimate is for just the U.S. Smith AC. Spring forward at your own risk: daylight saving time and fatal vehicle crashes. Am Econ J: Appl Econ. 2016;8(2):65-91. http://dx.doi.org/10.1257/app.20140100

1 0

[PATCH] Clarify LICENSE file
by Paul Eggert May 25, 2017

May 25, 2017

* LICENSE: Reword LICENSE so that it makes sense when only data files are distributed. --- LICENSE | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/LICENSE b/LICENSE index 148eb23..8ba4399 100644 --- a/LICENSE +++ b/LICENSE @@ -1,4 +1,5 @@ -With a few exceptions, all files in the tz code and data (including -this one) are in the public domain. The exceptions are date.c, -newstrftime.3, and strftime.c, which contain material derived from BSD -and which use the BSD 3-clause license. +Unless specified below, all files in the tz code and data (including +this LICENSE file) are in the public domain. + +If the files date.c, newstrftime.3, and strftime.c are present, they +contain material derived from BSD and use the BSD 3-clause license. -- 2.9.4

1 0

input needed on creation of a new sub-package for raw zone data
by Patsy Franklin May 24, 2017

May 24, 2017

Hi, I'd appreciate community input on this plan. Ideally if others decide to also ship a similar subpackage we can use a common naming convention and install directory. We are planning to ship a new subpackage for users who want to have access to the raw zone data files e.g. leapseconds, and in a pre-determined install location e.g. /usr/share/zoneinfo/zonedata/. The raw zone data is useful for users designing their own interfaces on top of this data. The broadest flexibility is offered by the raw zone data, and while the compiled binary data is versioned and more stable, some users have expressed a request for the raw zone data. The understanding is that the raw zone data format may change. In summary: Our current plan is to use -zonedata in naming the subpackage, For example, tzdata-zonedata-2017b-1.el7. We plan to install the files in this subpackage under /usr/share/zoneinfo/zonedata/. Just as an example we would ship the following files: LICENSE version africa antarctica asia australasia europe northamerica southamerica pacificnew etcetera backward systemv factory backzone iso3166.tab leapseconds leap-seconds.list zone1970.tab zone.tab Thank you! Patsy

5 11

[PROPOSED 1/5] New Makefile macro BACKWARD
by Paul Eggert May 24, 2017

May 24, 2017

* NEWS: Document this. * Makefile (BACKWARD): New macro. (YDATA, INSTALLARGS): Use it. --- Makefile | 10 +++++++++- NEWS | 4 ++++ 2 files changed, 13 insertions(+), 1 deletion(-) diff --git a/Makefile b/Makefile index c62d46f..6373920 100644 --- a/Makefile +++ b/Makefile @@ -92,6 +92,13 @@ LIBDIR= $(TOPDIR)/lib REDO= posix_right +# For backward-compatibility links for old zone names, use +# BACKWARD= backward pacificnew +# To omit these links, use +# BACKWARD= + +BACKWARD= backward pacificnew + # If you want out-of-scope and often-wrong data from the file 'backzone', use # PACKRATDATA= backzone # To omit this data, use @@ -384,7 +391,7 @@ WEB_PAGES= tz-art.htm tz-how-to.html tz-link.htm DOCS= $(MANS) date.1 $(MANTXTS) $(WEB_PAGES) PRIMARY_YDATA= africa antarctica asia australasia \ europe northamerica southamerica -YDATA= $(PRIMARY_YDATA) pacificnew etcetera backward +YDATA= $(PRIMARY_YDATA) etcetera $(BACKWARD) NDATA= systemv factory TDATA= $(YDATA) $(NDATA) ZONETABLES= zone1970.tab zone.tab @@ -482,6 +489,7 @@ leapseconds: $(LEAP_DEPS) # Arguments to pass to submakes of install_data. # They can be overridden by later submake arguments. INSTALLARGS = \ + BACKWARD=$(BACKWARD) \ DESTDIR=$(DESTDIR) \ LEAPSECONDS='$(LEAPSECONDS)' \ PACKRATDATA='$(PACKRATDATA)' \ diff --git a/NEWS b/NEWS index 4e8091b..dcd49eb 100644 --- a/NEWS +++ b/NEWS @@ -20,6 +20,10 @@ Unreleased, experimental changes in civil timekeeping. (Thanks to Robert Elz and Bradley White for noticing glitches in the code that uncovered this problem.) + 'make BACKWARD=' now suppresses backward-compatibility names + like 'US/Pacific' that are defined in the 'backward' and + 'pacificnew' files. + Several minor changes have been made to the code to make it a bit easier to port to MS-Windows. (Thanks to Kees Dekker for reporting the problems.) -- 2.9.4

1 4

Fwd: input needed on creation of a new sub-package for raw zone data
by Patsy Franklin May 23, 2017

May 23, 2017

Sorry, should have replied to all on this one. ---------- Forwarded message ---------- From: Patsy Franklin <pfrankli(a)redhat.com> Date: Tue, May 23, 2017 at 9:26 AM Subject: Re: [tz] input needed on creation of a new sub-package for raw zone data To: Brian.Inglis(a)systematicsw.ab.ca Hi Brian, Thanks for the feedback! On Mon, May 22, 2017 at 1:54 PM, Brian Inglis <Brian.Inglis(a)systematicsw.ab. ca> wrote: > On 2017-05-22 07:03, Patsy Franklin wrote: > > I'd appreciate community input on this plan. Ideally if others decide > > to also ship a similar subpackage we can use a common naming > > convention and install directory. > > > > We are planning to ship a new subpackage for users who want to have > > access to the raw zone data files e.g. leapseconds, and in a > > pre-determined install location e.g. /usr/share/zoneinfo/zonedata/. > > How about using /usr/share/tzdata{,-src}/ maybe with subdirectories > tzdata-2017b, etc. > > We really want to avoid using src to prevent any confusion between the full src rpm that we ship and this subset of the zone data. > > The raw zone data is useful for users designing their own interfaces > > on top of this data. The broadest flexibility is offered by the raw > > zone data, and while the compiled binary data is versioned and more > > stable, some users have expressed a request for the raw zone data. > > The understanding is that the raw zone data format may change. > > > > In summary: > > Our current plan is to use -zonedata in naming the subpackage, For > > example, tzdata-zonedata-2017b-1.el7. > > How about tzdata-src-2017b...? > > We already provide tzdata-2017b-1.el7.src.rpm which contains the full set of sources. The location of these sources is defined by the installer. > We plan to install the files in this subpackage under > > /usr/share/zoneinfo/zonedata/. > > Adding zonedata does not add any useful information about the contents > and storing under /usr/share/zoneinfo/ could be confusing as there are > already subdirectories posix and right holding copies of standard and > leapsecond compensated binary data files, so TZ=posix/... and > TZ=right/... are valid zones, and users may expect something similar to > happen using TZ=zonedata/... which would be unfortunate. > > That's a good point. I hadn't considered that case. We chose zonedata to clarify that it was only the zone data - not zic, zdump, binaries, etc. It was requested that we provide this subset of files in a consistent location at install time which is why we grouped it under the zoneinfo directory. Is there a more appropriate place to install it? Do you have any other suggestions for a name since I think src is ambiguous for our purposes? > Just as an example we would ship the following files: > > LICENSE > > version > > africa > > antarctica > > asia > > australasia > > europe > > northamerica > > southamerica > > pacificnew > > etcetera > > backward > > systemv > > factory > > backzone > > iso3166.tab > > leapseconds > > leap-seconds.list > > zone1970.tab > > zone.tab > > README, CONTRIBUTING, NEWS, and Theory should be included, also > tz-how-to.html which documents how to define and use the sources. > leap-seconds.list should be a (symbolic?) link to the canonical file > version leap-seconds.<timestamp> e.g. leap-seconds.3692908800. > We ship README and Theory with the base package. CONTRIBUTING, NEWS and tz-how-to.html is shipped with our src package. Do we still need to duplicate these in the optional subpackages? BTW, I forgot to mention that our spec file will not allow the new subpackage to be installed without the corresponding version of the base package. The base package does not require the optional new subpackage but will check to insure that it is in sync if it is installed. Do we still need to version the leap-seconds file if it is tied to a specific tzdata version? Thanks! Patsy > > -- > Take care. Thanks, Brian Inglis, Calgary, Alberta, Canada >

2 1

Looking for a way to cut down binary zoneinfo size. Does anyone have the way to?
by Viktor Sergiienko May 19, 2017

May 19, 2017

Hello, I'm developing for a pretty small-sized embedded system, and 2.9M of binaries is a bit too much for us. The system never uses a date before 2007. I estimate I can slim it down too 500K by cutting down the old data. 1. I made a script to cut down the source data: all the rules and all the zonelines before 2001. Compiling this data, I get a lot of errors like: > "northamerica", line 2537: %s in ruleless zone > "northamerica", line 2618: unruly zone because some zones are left without rules, but with %s format in them. 2. I tried to modify zic itself, but apparently understanding it will take me days. Just setting min_year to 2001 results in bunch of > can't determine time zone abbreviation to use just after until time Does anyone have a ready solution? Thanks!

7 11

[PROPOSED 1/2] Rename ‘lint’ to ‘GCC_LINT’
by Paul Eggert May 18, 2017

May 18, 2017

* Makefile (GCC_DEBUG_FLAGS): * private.h (INITIALIZE): Rename ‘lint’ to ‘GCC_LINT’, as the macro is intended for GCC, not for traditional ‘lint’. --- Makefile | 2 +- private.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/Makefile b/Makefile index e7c6ac0..bcfdcb2 100644 --- a/Makefile +++ b/Makefile @@ -159,7 +159,7 @@ LDLIBS= # (or some other number) to set the maximum time zone abbreviation length # that zic will accept without a warning (the default is 6) # $(GCC_DEBUG_FLAGS) if you are using recent GCC and want lots of checking -GCC_DEBUG_FLAGS = -Dlint -g3 -O3 -fno-common -fstrict-aliasing \ +GCC_DEBUG_FLAGS = -DGCC_LINT -g3 -O3 -fno-common -fstrict-aliasing \ -Wall -Wextra \ -Wbad-function-cast -Wcast-align -Wdate-time \ -Wdeclaration-after-statement \ diff --git a/private.h b/private.h index e9eb211..6886185 100644 --- a/private.h +++ b/private.h @@ -566,7 +566,7 @@ static time_t const time_t_max = MAXVAL(time_t, TYPE_BIT(time_t)); ** INITIALIZE(x) */ -#ifdef lint +#ifdef GCC_LINT # define INITIALIZE(x) ((x) = 0) #else # define INITIALIZE(x) -- 2.9.4

1 1

Morocco changes May 27th 2017 instead of May 21th
by Info - avisched.com May 17, 2017

May 17, 2017

Morocco is about to switch on May 27th back to UTC from +1. The current release doesn't reflect that (May 21st). Is there any update planned before the actual transition? Many thanks! Martin

5 5

Asia/Harbin, Asia/Chongqing and Asia/Kashgar
by Phake Nick May 13, 2017

May 13, 2017

iirc, according to past conversation on the mailing list, the reason why these timezones that are not in use for many years already still exists in tz database is because of legacy reason, however given that Asia/Ishigaki were deleted from this database in 1999 when it's found to be not in use, is there any reason preventing the same to be done to Asia/Harbin, Asia/Chongqing, and Asia/Kashgar?

2 1

gcc 8 demanded patch (from NetBSD)
by Robert Elz May 11, 2017

May 11, 2017

The following patch was recently added to NetBSD, to appease GCC 8 (apparently) ... "hit" is a bool, and it seems gcc 8 does not like ++ operators applied to one of those. (deleted parts of the diff which just showed the version ID string being updated...) kre Index: src/lib/libc/time/localtime.c diff -u src/lib/libc/time/localtime.c:1.106 src/lib/libc/time/localtime.c:1.107 --- src/lib/libc/time/localtime.c:1.106 Sat Mar 11 18:23:14 2017 +++ src/lib/libc/time/localtime.c Tue May 9 02:30:49 2017 @@ -1716,7 +1716,7 @@ timesub(const time_t *timep, int_fast32_ sp->lsis[i - 1].ls_trans + 1 && sp->lsis[i].ls_corr == sp->lsis[i - 1].ls_corr + 1) { - ++hit; + hit = true; --i; } }

8 13

TZDEFRULESTRING is increasingly out-of-date
by Tom Lane May 7, 2017

May 7, 2017

Due to a script's unintended failure to supply a -p option to zic, I recently had my nose rubbed in the fact that localtime.c still has this as the fallback when the "posixrules" link can't be loaded: /* ** The DST rules to use if TZ has no rules and we can't load TZDEFRULES. ** We default to US rules as of 1999-08-17. ** POSIX 1003.1 section 8.1.1 says that the default DST rules are ** implementation dependent; for historical reasons, US rules are a ** common default. */ #ifndef TZDEFRULESTRING #define TZDEFRULESTRING ",M4.1.0,M10.5.0" #endif /* !defined TZDEFDST */ It's now over ten years since the US last followed those rules; isn't it time to update TZDEFRULESTRING to match current law? I can't really think of any situation where the current default would be of any value now. Apologies if this policy is being followed intentionally; but I couldn't find any discussion of the point by googling. regards, tom lane

2 1

Potted history of timekeeping from the BBC
by John Haxby April 21, 2017

April 21, 2017

This is from a series of things that made the modern economy, it says on the web page: Tick tock: The importance of knowing the right time: http://www.bbc.co.uk/news/business-39129620 In one short page it manages to mention everything but DST :) and it includes a picture of Harrison No 1. (If you're ever near Greenwich I recommend visiting the Royal Observatory, taking the obligatory foot-in-both-hemispheres selfie and then going to look at the clocks.) >From the BBC web page though, I particularly noticed this: > Some financiers recently calculated it was worth spending $300m (£247m) drilling through mountains between Chicago and New York to lay fibre-optic cables in a slightly straighter line. That sped up communication between the two cities' exchanges by three milliseconds. jch

10 10

Re: [tz] leapseconds.awk, LEAP directives and references
by Chris Woodbury April 21, 2017

April 21, 2017

In the course of reviewing the AWK script leapseconds.awk, I was disturbed by the fact that it (1) operated on the comments (probably not a good idea) in the NIST produced leap-seconds.list file rather than the NTP timestamp itself and (2) that it assumed that leap seconds would only occur in December and June. The regulation that governs leap seconds, ITU-R TF.460-6 <https://www.itu.int/rec/R-REC-TF.460-6-200202-I/>, actually allows leap seconds to occur at the end of any UTC month. Annex 1, Section 2: Leap Seconds 2.1 A positive or negative leap-second should be the last second of a UTC month, but first preference should be given to the end of December and June, and second preference to the end of March and September. I note, in the event of a (highly unlikely) negative leap second, the LEAP directive generated is for 23:59:59 which, in the case of a negative leap second, doesn't exist. I reviewed the C code and, it seems to me, the transaction added should actually be for the second before, 23:59:58, as the transaction itself causes 23:59:59 to be skippedaltogether. (Kinda hard to skip an event if you're in it.) I bring this up as I've alrady written the changes necessary but I want to verify the correctness of this last beforesubmitting them. Also, I've an AWK script that uses the IERS file Leap_Seconds.dat rather than the NIST leap-seconds.listfile. Leap_Seconds.dat is a straight list of leap-seconds. USNO MAIA also has straight list of leap-seconds in a different format. See URLs: <http://hpiers.obspm.fr/iers/bul/bulc/Leap_Seconds.dat> <http://maia.usno.navy.mil/ser7/leapsec.dat> Either of these would, particularly the IERS file as it has few but excellent comments, be a better source of leap-second information than the once removed NIST file leap-seconds.list; which is purposed for another task. Lastly, thanks to Dr. Eggert for including the reference I gave him to the AAPT Resource Letter that included Terry Quinn's article in the NIST Time and Frequency Division's PDF archive (1168). Anyone wanting to know about Time and Frequency Measurement might want to read the other articles as well as Dr. Quinn's. Another good, comprehensive article is "The leap second: its history and possible future" by R.A. Nelson, et al, from Metrogia Vol 38, pp. 509-529 (2001). PDF available (with permission) at: <https://www.cl.cam.ac.uk/~mgk25/time/metrologia-leapsecond.pdf>

9 16

[PROPOSED] Use railway time for India from 1870-1941
by Paul Eggert April 20, 2017

April 20, 2017

From: Paul Eggert <eggert(a)Penguin.CS.UCLA.EDU> * NEWS: Document this. * asia (Asia/Yangon): Add 7 s to times from 1880 to 1920. (Asia/Kolkata): Switch from HMT to MMT in 1870, and from MMT to IST in 1906. --- NEWS | 11 +++++++++++ asia | 59 +++++++++++++++++++++++++++++++++++++++++++++++++++++------ 2 files changed, 64 insertions(+), 6 deletions(-) diff --git a/NEWS b/NEWS index f986cf6..ad099f4 100644 --- a/NEWS +++ b/NEWS @@ -1,5 +1,16 @@ News for the tz database +Unreleased, experimental changes + + Changes to past time stamps + + Use railway time for Asia/Kolkata before 1941, by switching to + Madras local time (UT +052110) in 1870, then to IST (UT +0530) in + 1906. Also, treat 1941-2's +0630 as DST, like 1942-5. + + Add 7 s to the UT offset in Asia/Yangon before 1920. + + Release 2017b - 2017-03-17 07:30:38 -0700 Briefly: Haiti has resumed DST. diff --git a/asia b/asia index 35774c6..f027071 100644 --- a/asia +++ b/asia @@ -258,9 +258,15 @@ Zone Asia/Brunei 7:39:40 - LMT 1926 Mar # Bandar Seri Begawan # Milne says 6:24:40 was the meridian of the time ball observatory at Rangoon. +# From Paul Eggert (2017-04-20): +# Page 27 of Reed & Low (cited for Asia/Kolkata) says "Rangoon local time is +# used upon the railways and telegraphs of Burma, and is 6h. 24m. 47s. ahead +# of Greenwich." This refers to the period before Burma's transition to +0630, +# a transition for which Shanks is the only source. + # Zone NAME GMTOFF RULES FORMAT [UNTIL] -Zone Asia/Yangon 6:24:40 - LMT 1880 # or Rangoon - 6:24:40 - RMT 1920 # Rangoon Mean Time? +Zone Asia/Yangon 6:24:47 - LMT 1880 # or Rangoon + 6:24:47 - RMT 1920 # Rangoon local time 6:30 - +0630 1942 May 9:00 - +09 1945 May 3 6:30 - +0630 @@ -887,14 +893,55 @@ Zone Asia/Dili 8:22:20 - LMT 1912 Jan 1 # local time in favor of Indian Standard Time.... Journalists called this # dispute the "Battle of the Clocks." It lasted nearly half a century. +# From Paul Eggert (2017-04-20): +# Good luck trying to nail down old timekeeping records in India. +# "... in the nineteenth century ... Madras Observatory took its magnetic +# measurements in Göttingen time, its meterological measurements on Madras +# (local) time, dropped its time ball on Greenwich (ocean navigator's) time, +# and distributed civil (local time)." -- Bartky IR. Selling the time: +# 19th-century timekeeping in america. Stanford U Press (2000), 247 note 19. +# "A more potent cause of resistance to the general adoption of the present +# standard time lies in the fact that it is Madras time. The citizen of +# Bombay, proud of being 'primus in Indis' and of Calcutta, equally proud of +# his city being the Capital of India, and - for a part of the year - the Seat +# of the Supreme Government, alike look down on Madras, and refuse to change +# the time they are using, for that of what they regard as a benighted +# Presidency; while Madras, having for long given the standard time to the +# rest of India, would resist the adoption of any other Indian standard in its +# place." -- Oldham RD. On Time in India: a suggestion for its improvement. +# Proceedings of the Asiatic Society of Bengal (April 1899), 49-55. +# +# "In 1870 ... Madras time - 'now used by the telegraph and regulated from the +# only government observatory' - was suggested as a standard railway time, +# first to be dopted on the Great Indian Peninsular Railway (GIPR).... +# Calcutta, Bombay, and Karachi, were to be allowed to continue with their +# local time for civil purposes." - Prasad R. Tracks of Change: Railways and +# Everyday Life in Colonial India. Cambridge University Press (2016), 145. +# +# Reed S, Low F. The Indian Year Book 1936-37. Bennett, Coleman, pp 27-8. +# https://archive.org/details/in.ernet.dli.2015.282212 +# This lists +052110 as Madras local time used in railways, and says that on +# 1906-01-01 railways and telegraphs in India switched to +0530. Some +# municipalities retained their former time, and the time in Calcutta +# continued to depend on whether you were at the railway station or at +# government offices. Government time was at +055320 (according to Shanks) or +# at +0554 (according to the Indian Year Book). Railway time is more +# appropriate for our purposes, as it was better documented, it is what we do +# elsewhere (e.g., Europe/London before 1880), and after 1906 it was +# consistent in the region now identified by Asia/Kolkata. So, use railway +# time for 1870-1941. Shanks is our only (and dubious) source for the +# 1941-1945 data. + # Zone NAME GMTOFF RULES FORMAT [UNTIL] -Zone Asia/Kolkata 5:53:28 - LMT 1880 # Kolkata - 5:53:20 - HMT 1941 Oct # Howrah Mean Time? - 6:30 - +0630 1942 May 15 +Zone Asia/Kolkata 5:53:28 - LMT 1854 Jun 28 # Kolkata + 5:53:20 - HMT 1870 # Howrah Mean Time? + 5:21:10 - MMT 1906 Jan 1 # Madras local time + 5:30 - IST 1941 Oct + 5:30 1:00 +0630 1942 May 15 5:30 - IST 1942 Sep 5:30 1:00 +0630 1945 Oct 15 5:30 - IST -# The following are like Asia/Kolkata: +# Since 1970 the following are like Asia/Kolkata: # Andaman Is # Lakshadweep (Laccadive, Minicoy and Amindivi Is) # Nicobar Is -- 2.9.3

1 0

Mouradconstantine@icloud.com
by Mourad Constantine April 13, 2017

April 13, 2017

Sent from my iPhone

1 0