- tz - lists.iana.org

"time zone" vs "timezone" in documentation
by Paul Eggert July 4, 2018

July 4, 2018

[I am cc'ing this to the tz mailing list (and changing the subject line) to give tz readers a heads-up about the terminology issue. The question here is: should tzdb man pages say "time zone" or "timezone"? For context please see Michael Kerrisk's email <https://marc.info/?m=152964652828320>, which says: > When I inherited the project, the pages used a mixture of "time zone" > and "timezone", with the former predominant. I nevertheless > standardized on the latter, and although I don't recall for sure, I > suspect it was because that is the spelling used in POSIX. (As an > aside, there's an argument that--because POSIX--tzdb might want to > consider switching spellings.) I'm not religious about the particular > choice (although I have naturally now got used to the particular > choice I made some years ago), but I did make that choice because I > want consistency within the project, and I'm reluctant to introduce > inconsistency. and my response <https://marc.info/?m=152964676331902>, which says: > I deferred to POSIX for "timestamp" versus "time stamp", but > "timezone" is a bridge too far for me. ] On second thought, perhaps I was too hasty. We could distinguish "time zone" in the usual English-language sense (a set of geographic locations that currently share the same standard time offset from UTC) from "timezone" in the POSIX sense (a history and predicted future of UTC offsets, abbreviations and isdst flags). If so, the tzdb documentation could be more careful about using "time zone" for the former and "timezone" for the latter, and this would make for fewer changes to the GNU/Linux man-pages for tzdb. I can look into this and propose an updated set of tzdb-related patches accordingly. If I were designing the terminology from scratch, I wouldn't specify two nearly-identical phrases "time zone" and "timezone" to mean such different things. However, the phrases do have the advantage of existing practice (common English usage and POSIX, respectively).

9 11

[PROPOSED] * tz-link.html (Precision timekeeping): Mention Huygens.
by Paul Eggert June 29, 2018

June 29, 2018

--- tz-link.html | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/tz-link.html b/tz-link.html index 548fbba..7ed1efd 100644 --- a/tz-link.html +++ b/tz-link.html @@ -813,11 +813,15 @@ title="Network Time Protocol">NTP</abbr>: The Network Time Protocol</a> (Internet <abbr>RFC</abbr> 5905) discusses how to synchronize clocks of Internet hosts.</li> +<li>The <a href="https://www.usenix.org/system/files/conference/nsdi18/nsdi18-geng.pdf"><font style="font-variant: small-caps">Huygens</font></a> +family of software algorithms can achieve accuracy to a few tens of +nanoseconds in scalable server farms without special hardware.</li> <li>The <a href="https://www.nist.gov/intelligent-systems-division/ieee-1588">Precision Time Protocol</a> (<abbr title="Institute of Electrical and Electronics Engineers">IEEE</abbr> 1588) -can achieve submicrosecond clock accuracy on a local area network.</li> +can achieve submicrosecond clock accuracy on a local area network +with special-purpose hardware.</li> <li><a href="https://tools.ietf.org/html/rfc4833">Timezone Options for <abbr title="Dynamic Host Configuration Protocol">DHCP</abbr></a> -- 2.17.1

1 0

California will vote on permanent daylight
by Steve Allen June 29, 2018

June 29, 2018

The governor approved AB 807 https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=201720… Daylight time in California was established by the voters, and must be repealed by the voters, so nothing happens until after an election decides whether to ask the state legislature to consider how permanent daylight time could be achieved. -- Steve Allen <sla(a)ucolick.org> WGS-84 (GPS) UCO/Lick Observatory--ISB 260 Natural Sciences II, Room 165 Lat +36.99855 1156 High Street Voice: +1 831 459 3046 Lng -122.06015 Santa Cruz, CA 95064 http://www.ucolick.org/~sla/ Hgt +250 m

1 0

[PROPOSED] Make Makefile safer for parallelism
by Paul Eggert June 26, 2018

June 26, 2018

Although this still does not fully work with GNU Make, it’s better than what we had. * Makefile (TIME_T_ALTERNATIVES_HEAD) (TIME_T_ALTERNATIVES_TAIL): New macros. (TIME_T_ALTERNATIVES): Use them. (TZS_CUTOFF_FLAG): New macro. ($(ZDS)): Use it, so that submakes can use -t instead of -c cutoffs. (EIGHT_YARDS): New macro. (ENCHILADA): Use it. (INSTALL_DATA_DEPS, TZS_NEW_DEPS): New macros. (install_data, posix_only, right_only, posix_packrat, $(TZS_NEW)): Use them to avoid duplicate work with GNU make -j. ($(TZS_NEW), check_zishrink_posix, check_zishrink_right, check_public) (check_time_t_alternatives): Use rule-specific temporary directories so that other rules can run in parallel. ($(TZS_NEW)): Check for sort failure. ($(TZS)): Simply touch it if it doesn’t exist, so that it’s empty. (check_tzs): Treat empty $(TZS) as nonexistent. (check_character_set, check_white_space, check_name_lengths) (check_sorted, check_links, check_tables, check_tzs, check_web) (check_zishrink, check_public): No longer phony. (check_zishrink_posix, check_zishrink_right) (typecheck_long_long, typecheck_unsigned): New rules, so that they can be done in parallel. (clean_misc): Clean the timestamp files that are now made due to rules no longer being phony. (set-timestamps.out): Do not set timestamp of $(TZS). (set-tzs-timestamp.out): New rule. (check_time_t_alternatives): Break into subrules, one for each alternative, so that they can be done in parallel. (tarballs, rearguard_tarballs, traditional_tarballs) (signatures, rearguard_signatures, traditional_signatures): Depend on set-timestamps.out and rearguard.zi. (tzdb-$(VERSION).tar.lz): Depend on set-tzs-timestamp.out. (TYPECHECK_CFLAGS): New macro. --- Makefile | 240 +++++++++++++++++++++++++++++++++++++-------------------------- 1 file changed, 142 insertions(+), 98 deletions(-) diff --git a/Makefile b/Makefile index 2289dff..29058e5 100644 --- a/Makefile +++ b/Makefile @@ -106,8 +106,10 @@ MANDIR = $(TOPDIR)/$(USRSHAREDIR)/man LIBDIR = $(TOPDIR)/$(USRDIR)/lib -# Types to try, as an alternative to time_t. int64_t should be first. -TIME_T_ALTERNATIVES = int64_t int32_t uint32_t uint64_t +# Types to try, as an alternative to time_t. +TIME_T_ALTERNATIVES = $(TIME_T_ALTERNATIVES_HEAD) $(TIME_T_ALTERNATIVES_TAIL) +TIME_T_ALTERNATIVES_HEAD = int64_t +TIME_T_ALTERNATIVES_TAIL = int32_t uint32_t uint64_t # What kind of TZif data files to generate. # (TZif is the binary time zone data format that zic generates.) @@ -493,11 +495,14 @@ AWK_SCRIPTS= checklinks.awk checktab.awk leapseconds.awk \ ziguard.awk zishrink.awk MISC= $(AWK_SCRIPTS) zoneinfo2tdf.pl TZS_YEAR= 2050 +TZS_CUTOFF_FLAG= -c $(TZS_YEAR) TZS= to$(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) tzdata.zi +# EIGHT_YARDS is just a yard short of the whole ENCHILADA. +EIGHT_YARDS = $(COMMON) $(DOCS) $(SOURCES) $(DATA) $(MISC) tzdata.zi +ENCHILADA = $(EIGHT_YARDS) $(TZS) # Consult these files when deciding whether to rebuild the 'version' file. # This list is not the same as the output of 'git ls-files', since @@ -606,14 +611,16 @@ INSTALLARGS = \ YEARISTYPE='$(YEARISTYPE)' \ ZIC='$(ZIC)' +INSTALL_DATA_DEPS = zic leapseconds yearistype tzdata.zi + # 'make install_data' installs one set of TZif files. -install_data: zic leapseconds yearistype tzdata.zi +install_data: $(INSTALL_DATA_DEPS) $(ZIC_INSTALL) tzdata.zi -posix_only: +posix_only: $(INSTALL_DATA_DEPS) $(MAKE) $(INSTALLARGS) LEAPSECONDS= install_data -right_only: +right_only: $(INSTALL_DATA_DEPS) $(MAKE) $(INSTALLARGS) LEAPSECONDS='-L leapseconds' \ install_data @@ -640,7 +647,7 @@ posix_right: posix_only # This obsolescent rule is present for backwards compatibility with # tz releases 2014g through 2015g. It should go away eventually. -posix_packrat: +posix_packrat: $(INSTALL_DATA_DEPS) $(MAKE) $(INSTALLARGS) PACKRATDATA=backzone posix_only zones: $(REDO) @@ -651,29 +658,33 @@ ZDS = dummy.zd # Rule used only by submakes invoked by the $(TZS_NEW) rule. # It is separate so that GNU 'make -j' can run instances in parallel. $(ZDS): zdump - ./zdump -i -c $(TZS_YEAR) '$(wd)/'$$(expr $@ : '$.*$.zd') >$@ - -$(TZS_NEW): tzdata.zi zdump zic - rm -fr tzs.dir - mkdir tzs.dir - $(zic) -d tzs.dir tzdata.zi + ./zdump -i $(TZS_CUTOFF_FLAG) '$(wd)/'$$(expr $@ : '$.*$.zd') \ + >$@ + +TZS_NEW_DEPS = tzdata.zi zdump zic +$(TZS_NEW): $(TZS_NEW_DEPS) + rm -fr tzs$(TZS_YEAR).dir + mkdir tzs$(TZS_YEAR).dir + $(zic) -d tzs$(TZS_YEAR).dir tzdata.zi $(AWK) '/^L/{print "Link\t" $$2 "\t" $$3}' \ tzdata.zi | LC_ALL=C sort >$@.out wd=`pwd` && \ - set x `$(AWK) '/^Z/{print "tzs.dir/" $$2 ".zd"}' tzdata.zi \ + x=`$(AWK) '/^Z/{print "tzs$(TZS_YEAR).dir/" $$2 ".zd"}' \ + tzdata.zi \ | LC_ALL=C sort -t . -k 2,2` && \ + set x $$x && \ shift && \ ZDS=$$* && \ - $(MAKE) wd="$$wd" TZS_YEAR=$(TZS_YEAR) ZDS="$$ZDS" $$ZDS && \ - sed 's,^TZ=".*tzs\.dir/,TZ=",' $$ZDS >>$@.out - rm -fr tzs.dir + $(MAKE) wd="$$wd" TZS_CUTOFF_FLAG="$(TZS_CUTOFF_FLAG)" \ + ZDS="$$ZDS" $$ZDS && \ + sed 's,^TZ=".*\.dir/,TZ=",' $$ZDS >>$@.out + rm -fr tzs$(TZS_YEAR).dir mv $@.out $@ -# If $(TZS) does not already exist (e.g., old-format tarballs), create it. -# If it exists but 'make check_tzs' fails, a maintainer should inspect the +# If $(TZS) exists but 'make check_tzs' fails, a maintainer should inspect the # failed output and fix the inconsistency, perhaps by running 'make force_tzs'. $(TZS): - $(MAKE) force_tzs + touch $@ force_tzs: $(TZS_NEW) cp $(TZS_NEW) $(TZS) @@ -718,12 +729,14 @@ check_character_set: $(ENCHILADA) leapseconds yearistype.sh zone.tab && \ ! grep -Env $(OK_LINE) $(ENCHILADA); \ } + touch $@ check_white_space: $(ENCHILADA) patfmt=' \t|[\f\r\v]' && pat=`printf "$$patfmt\\n"` && \ ! grep -En "$$pat" $(ENCHILADA) ! grep -n '[[:space:]]$$' \ $$(ls $(ENCHILADA) | grep -Fvx leap-seconds.list) + touch $@ PRECEDES_FILE_NAME = ^(Zone|Link[[:space:]]+[^[:space:]]+)[[:space:]]+ FILE_NAME_COMPONENT_TOO_LONG = \ @@ -732,6 +745,7 @@ FILE_NAME_COMPONENT_TOO_LONG = \ check_name_lengths: $(TDATA_TO_CHECK) backzone ! grep -En '$(FILE_NAME_COMPONENT_TOO_LONG)' \ $(TDATA_TO_CHECK) backzone + touch $@ CHECK_CC_LIST = { n = split($$1,a,/,/); for (i=2; i<=n; i++) print a[1], a[i]; } @@ -744,52 +758,61 @@ check_sorted: backward backzone iso3166.tab zone.tab zone1970.tab LC_ALL=C sort -c $(AWK) '/^[^#]/ $(CHECK_CC_LIST)' zone1970.tab | \ LC_ALL=C sort -cu + touch $@ check_links: checklinks.awk $(TDATA_TO_CHECK) tzdata.zi $(AWK) -f checklinks.awk $(TDATA_TO_CHECK) $(AWK) -f checklinks.awk tzdata.zi + touch $@ check_tables: checktab.awk $(PRIMARY_YDATA) $(ZONETABLES) for tab in $(ZONETABLES); do \ $(AWK) -f checktab.awk -v zone_table=$$tab $(PRIMARY_YDATA) \ || exit; \ done + touch $@ check_tzs: $(TZS) $(TZS_NEW) - diff -u $(TZS) $(TZS_NEW) + if test -s $(TZS); then \ + diff -u $(TZS) $(TZS_NEW); \ + else \ + cp $(TZS_NEW) $(TZS); \ + fi + touch $@ # This checks only the HTML 4.01 strict page. # To check the the other pages, use <https://validator.w3.org/>. check_web: tz-how-to.html $(VALIDATE_ENV) $(VALIDATE) $(VALIDATE_FLAGS) tz-how-to.html + touch $@ # Check that zishrink.awk does not alter the data, and that ziguard.awk # preserves main-format data. -check_zishrink: zic leapseconds $(PACKRATDATA) $(TDATA) \ - $(DATAFORM).zi tzdata.zi - for type in posix right; do \ - mkdir -p time_t.dir/$$type time_t.dir/$$type-t \ - time_t.dir/$$type-shrunk && \ - case $$type in \ - right) leap='-L leapseconds';; \ - *) leap=;; \ - esac && \ - $(ZIC) $$leap -d time_t.dir/$$type $(DATAFORM).zi && \ +check_zishrink: check_zishrink_posix check_zishrink_right +check_zishrink_posix check_zishrink_right: \ + zic leapseconds $(PACKRATDATA) $(TDATA) $(DATAFORM).zi tzdata.zi + rm -fr $@.dir $(a)-t.dir $(a)-shrunk.dir + mkdir $@.dir $(a)-t.dir $(a)-shrunk.dir + case $@ in \ + *_right) leap='-L leapseconds';; \ + *) leap=;; \ + esac && \ + $(ZIC) $$leap -d $@.dir $(DATAFORM).zi && \ + $(ZIC) $$leap -d $(a)-shrunk.dir tzdata.zi && \ case $(DATAFORM) in \ main) \ - $(ZIC) $$leap -d time_t.dir/$$type-t $(TDATA) && \ + $(ZIC) $$leap -d $(a)-t.dir $(TDATA) && \ $(AWK) '/^Rule/' $(TDATA) | \ - $(ZIC) $$leap -d time_t.dir/$$type-t - \ - $(PACKRATDATA) && \ - diff -r time_t.dir/$$type time_t.dir/$$type-t;; \ - esac && \ - $(ZIC) $$leap -d time_t.dir/$$type-shrunk tzdata.zi && \ - diff -r time_t.dir/$$type time_t.dir/$$type-shrunk || exit; \ - done - rm -fr time_t.dir + $(ZIC) $$leap -d $(a)-t.dir - $(PACKRATDATA) && \ + diff -r $@.dir $(a)-t.dir;; \ + esac + diff -r $@.dir $(a)-shrunk.dir + rm -fr $@.dir $(a)-t.dir $(a)-shrunk.dir + touch $@ clean_misc: - rm -f core *.o *.out \ + rm -f *.o *.out $(TIME_T_ALTERNATIVES) \ + check_* core typecheck_* \ date tzselect version.h zdump zic yearistype libtz.a clean: clean_misc rm -fr *.dir *.zi tzdb-*/ $(TZS_NEW) @@ -826,10 +849,10 @@ $(MANTXTS): workman.sh # If git or GNU 'touch' is absent, don't bother to sync with git timestamps. # Also, set the timestamp of each prebuilt file like 'leapseconds' # to be the maximum of the files it depends on. -set-timestamps.out: $(ENCHILADA) +set-timestamps.out: $(EIGHT_YARDS) rm -f $@ if (type git) >/dev/null 2>&1 && \ - files=`git ls-files $(ENCHILADA)` && \ + files=`git ls-files $(EIGHT_YARDS)` && \ touch -md @1 test.out; then \ rm -f test.out && \ for file in $$files; do \ @@ -847,62 +870,74 @@ set-timestamps.out: $(ENCHILADA) exit; \ done touch -cmr `ls -t $(TZDATA_ZI_DEPS) | sed 1q` tzdata.zi - touch -cmr `ls -t $(TZS_DEPS) | sed 1q` $(TZS) touch -cmr `ls -t $(VERSION_DEPS) | sed 1q` version touch $@ +set-tzs-timestamp.out: $(TZS) + touch -cmr `ls -t $(TZS_DEPS) | sed 1q` $(TZS) + touch $@ # The zics below ensure that each data file can stand on its own. # We also do an all-files run to catch links to links. -check_public: - $(MAKE) maintainer-clean - $(MAKE) CFLAGS='$(GCC_DEBUG_FLAGS)' ALL - mkdir -p public.dir - for i in $(TDATA_TO_CHECK) tzdata.zi; do \ - $(zic) -v -d public.dir $$i 2>&1 || exit; \ +check_public: $(VERSION_DEPS) + rm -fr public.dir + mkdir public.dir + ln $(VERSION_DEPS) public.dir + cd public.dir && $(MAKE) CFLAGS='$(GCC_DEBUG_FLAGS)' ALL + for i in $(TDATA_TO_CHECK) public.dir/tzdata.zi; do \ + public.dir/zic -v -d public.dir/zoneinfo $$i 2>&1 || exit; \ done - $(zic) -v -d public.dir $(TDATA_TO_CHECK) + public.dir/zic -v -d public.dir/zoneinfo-all $(TDATA_TO_CHECK) rm -fr public.dir + touch $@ # Check that the code works under various alternative # implementations of time_t. -check_time_t_alternatives: - if diff -q Makefile Makefile 2>/dev/null; then \ - quiet_option='-q'; \ - else \ - quiet_option=''; \ - fi && \ +check_time_t_alternatives: $(TIME_T_ALTERNATIVES) +$(TIME_T_ALTERNATIVES_TAIL): $(TIME_T_ALTERNATIVES_HEAD) +$(TIME_T_ALTERNATIVES): $(VERSION_DEPS) + rm -fr $@.dir + mkdir $@.dir + ln $(VERSION_DEPS) $@.dir + case $@ in \ + int32_t) range=-2147483648,2147483648;; \ + u*) range=0,4294967296;; \ + *) range=-4294967296,4294967296;; \ + esac && \ wd=`pwd` && \ zones=`$(AWK) '/^[^#]/ { print $$3 }' <zone1970.tab` && \ - for type in $(TIME_T_ALTERNATIVES); do \ - mkdir -p time_t.dir/$$type && \ - $(MAKE) clean_misc && \ - $(MAKE) TOPDIR="$$wd/time_t.dir/$$type" \ - CFLAGS='$(CFLAGS) -Dtime_tz='"'$$type'" \ + if test $@ = $(TIME_T_ALTERNATIVES_HEAD); then \ + range_target=; \ + else \ + range_target=to$$range.tzs; \ + fi && \ + (cd $@.dir && \ + $(MAKE) TOPDIR="$$wd/$@.dir" \ + CFLAGS='$(CFLAGS) -Dtime_tz='"'$@'" \ REDO='$(REDO)' \ - install && \ - diff $$quiet_option -r \ - time_t.dir/int64_t/etc \ - time_t.dir/$$type/etc && \ - diff $$quiet_option -r \ - time_t.dir/int64_t/usr/share \ - time_t.dir/$$type/usr/share && \ - case $$type in \ - int32_t) range=-2147483648,2147483647;; \ - uint32_t) range=0,4294967296;; \ - int64_t) continue;; \ - *u*) range=0,10000000000;; \ - *) range=-10000000000,10000000000;; \ - esac && \ - echo checking $$type zones ... && \ - time_t.dir/int64_t/usr/bin/zdump -V -t $$range $$zones \ - >time_t.dir/int64_t.out && \ - time_t.dir/$$type/usr/bin/zdump -V -t $$range $$zones \ - >time_t.dir/$$type.out && \ - diff -u time_t.dir/int64_t.out time_t.dir/$$type.out \ - || exit; \ - done - rm -fr time_t.dir + D=$$wd/$@.dir \ + TZS_YEAR="$$range" TZS_CUTOFF_FLAG="-t $$range" \ + install $$range_target) && \ + test $@ = $(TIME_T_ALTERNATIVES_HEAD) || { \ + (cd $(TIME_T_ALTERNATIVES_HEAD).dir && \ + $(MAKE) TOPDIR="$$wd/$@.dir" \ + TZS_YEAR="$$range" TZS_CUTOFF_FLAG="-t $$range" \ + D=$$wd/$@.dir \ + to$$range.tzs) && \ + diff -u $(TIME_T_ALTERNATIVES_HEAD).dir/to$$range.tzs \ + $@.dir/to$$range.tzs && \ + if diff -q Makefile Makefile 2>/dev/null; then \ + quiet_option='-q'; \ + else \ + quiet_option=''; \ + fi && \ + diff $$quiet_option -r $(TIME_T_ALTERNATIVES_HEAD).dir/etc \ + $@.dir/etc && \ + diff $$quiet_option -r \ + $(TIME_T_ALTERNATIVES_HEAD).dir/usr/share \ + $@.dir/usr/share; \ + } + touch $@ TRADITIONAL_ASC = \ tzcode$(VERSION).tar.gz.asc \ @@ -913,7 +948,8 @@ ALL_ASC = $(TRADITIONAL_ASC) $(REARGUARD_ASC) \ tzdb-$(VERSION).tar.lz.asc tarballs rearguard_tarballs traditional_tarballs \ -signatures rearguard_signatures traditional_signatures: version +signatures rearguard_signatures traditional_signatures: \ + version set-timestamps.out rearguard.zi VERSION=`cat version` && \ $(MAKE) VERSION="$$VERSION" $@_version @@ -963,7 +999,7 @@ tzdata$(VERSION)-rearguard.tar.gz: rearguard.zi set-timestamps.out gzip $(GZIPFLAGS)) >$@.out mv $@.out $@ -tzdb-$(VERSION).tar.lz: set-timestamps.out +tzdb-$(VERSION).tar.lz: set-timestamps.out set-tzs-timestamp.out rm -fr tzdb-$(VERSION) mkdir tzdb-$(VERSION) ln $(ENCHILADA) tzdb-$(VERSION) @@ -979,14 +1015,24 @@ tzdb-$(VERSION).tar.lz.asc: tzdb-$(VERSION).tar.lz $(ALL_ASC): gpg2 --armor --detach-sign $? -typecheck: - $(MAKE) clean - for i in "long long" unsigned; \ - do \ - $(MAKE) CFLAGS="-DTYPECHECK -D__time_t_defined -D_TIME_T \"-Dtime_t=$$i\"" ; \ - ./zdump -v Europe/Rome ; \ - $(MAKE) clean ; \ - done +TYPECHECK_CFLAGS = $(CFLAGS) -DTYPECHECK -D__time_t_defined -D_TIME_T +typecheck: typecheck_long_long typecheck_unsigned +typecheck_long_long typecheck_unsigned: $(VERSION_DEPS) + rm -fr $@.dir + mkdir $@.dir + ln $(VERSION_DEPS) $@.dir + cd $@.dir && \ + case $@ in \ + *_long_long) i="long long";; \ + *_unsigned ) i="unsigned" ;; \ + esac && \ + typecheck_cflags='' && \ + $(MAKE) \ + CFLAGS="$(TYPECHECK_CFLAGS) \"-Dtime_t=$$i\"" \ + TOPDIR="`pwd`" \ + install + $@.dir/zdump -i -c 1970,1971 Europe/Rome + touch $@ zonenames: tzdata.zi @$(AWK) '/^Z/ { print $$2 } /^L/ { print $$3 }' tzdata.zi @@ -1002,9 +1048,7 @@ zic.o: private.h tzfile.h version.h .KEEP_STATE: .PHONY: ALL INSTALL all -.PHONY: check check_character_set check_links check_name_lengths -.PHONY: check_public check_sorted check_tables -.PHONY: check_time_t_alternatives check_tzs check_web check_white_space +.PHONY: check check_time_t_alternatives .PHONY: check_zishrink .PHONY: clean clean_misc dummy.zd force_tzs .PHONY: install install_data maintainer-clean names -- 2.7.4

1 0

Europe/Zaporozhye
by Paul G June 26, 2018

June 26, 2018

Recently during a conversation about the tzdb and a colleague of mine mentioned that it seems to be a strange outlier that Europe/Zaporozhye has its own zone, since Zaporozh'ye is not a particularly large city (it's quite close to Donetsk, which he deemed more likely to be the major city in the area). I looked at the source code and saw that it mentions: # Zaporozh'ye and eastern Lugansk oblasts observed DST 1990/1991. (https://github.com/eggert/tz/blob/master/europe#L3850) He seemed to think that this was an absurd notion and that more or less any city in Ukraine would either follow Moscow time or Crimea time, but none of them would have yet a third history. It does seem a bit strange (though perfectly consistent with the weird nonsense everyone else gets up to) that only Zaporozh'ye and eastern Lugansk would follow a different time zone than their immediate neighbors - these are not even contiguous regions. In the timezone-boundary-builder project (https://github.com/evansiroky/timezone-boundary-builder) it seems that only Zaporozhia Oblast is included in the shapefile. We did a cursory search for references (including in the archives of this mailing list and the git history) in English and Russian and couldn't find any reference to Zaporozhia and Lugansk having a different time zone that wasn't derived from this project. So, questions about this: 1. Can anyone point to a reference to the fact that these regions used a different time zone than their neighbors? 2. If no reference is handy, does anyone know on what basis this zone was included in the database in the first place? 3. Is there any clarification as to the geographic extent of the zone? Does it cover Donetsk as well? What part of Lugansk counts as "eastern Lugansk"? Best, Paul

2 1

redundant Rule in proposal for Macao
by Michael H Deckers June 26, 2018

June 26, 2018

In Paul Eggert's proposal for Macao from 2018-05-10, we find +Rule Macau 1949 1951 - Apr Sat>=1 23:00s 1:00 D +Rule Macau 1951 only - Mar 31 23:00s 1:00 D with two successive switches to summer time at 1951-03-31 and 1951-04-07. The later switch has no effect, and can easily be omitted: +Rule Macau 1949 1950 - Apr Sat>=1 23:00s 1:00 D There are several similar cases in tzdb where the specification contains redundant data. For instance, Rule HK 1965 1976 - Apr Sun>=16 3:30 1:00 S Rule HK 1973 only - Dec 30 3:30 1:00 S with two successive switches to summer time at 1973-12-30 and 1974-04-21, the latter having no effect. In this case, removal of the redundancy would require an addtional Rule line, so concise notation could be construed as a reason for retaining the redundancy. Do we have a guideline for such cases? Michael Deckers.

2 1

Re: [tz] [PROPOSED] Use "PST/PDT" for Philippine time
by Paul Eggert June 25, 2018

June 25, 2018

Michael H Deckers wrote: > Fine if this were a generally applied guideline -- but it is not, > see eg America/Adak, Indonesia, and many others. When is this guideline > applied, and when not? The guideline is supposed to be applied reasonably consistently for clear-cut cases such as the Philippines. Alaska and Indonesia are more complicated, since they are not merely a question of what's the name of a time zone; they are also a question of time zone boundaries moving, and where it's not so clear what is being named. For example, a historian today would be unlikely to write "Sitka observed Alaska Standard Time in 1984" because in 1984 the Alaska time zone boundaries were quite different and Sitka was then in the geographically small part of Alaska that was in the same time zone as the Yukon. Unfortunately I don't see a good solution for these more-complicated cases, and am reluctant to change the data if it's likely this would introduce as much or more confusion than it would alleviate. > And all that does not justify the resurrection of an invented abbreviation. First, for the Phillippines it's no longer an abbreviation that exists only because we invented it. It's taken on a life of its own in reliable sources and in national law. And second, I doubt whether these sources got "PST" from tzdb, since tzdb has been using "PHST" for over two decades and used "PST" only briefly before that. Most likely it was independent reinvention.

3 2

Re: [tz] [IANA #985615] Sudan Time Zone change
by ahmed.atyya＠ntc.gov.sd June 25, 2018

June 25, 2018

Dear Sir/ Madam we would like to inform you that the Republic of Sudan is going to change the time zone from (GMT + 3:00) to (GMT+ 2:00) starting from Wednesday 1 November 2017. this is for your information and action if required. Best Regards Dear Mr. Paul We would like to inform you as TZ coordinator that the Republic of Sudan is planning to change its time zone from GMT+ 3:00 to GMT +2:00 ( The original time zone ) starting from 1 November 2017 Please if you can update time zone data base Or you you can tell us what action is needed from Sudan government Best Regards Ahmed Atyya Head of Numbering Section National Telecomm. Corporation – Sudan Phone: +249123499988 Email: ahmed.atyya@ ----- Original Message ----- From: Sabrina Tanamal via RT <iana-questions(a)iana.org> To: ahmed atyya <ahmed.atyya(a)ntc.gov.sd> Sent: Tue, 17 Oct 2017 01:33:06 +0300 (EAT) Subject: [IANA #985615] Sudan Time Zone change Dear Ahmed, Thank you for contacting us. We host the Time Zone Database, but updates are made by the Time Zone Coordinator in consultation with the Time Zone mailing list, in accordance with RFC 6557. Questions about the contents of the database should be sent to the mailing list. For mailing list archives and subscription information, please see http://www.iana.org/time-zones Best regards, Sabrina Tanamal IANA Services Specialist On Mon Oct 16 09:55:06 2017, ahmed.atyya(a)ntc.gov.sd wrote: > Dear Sir/ Madam > > > > > > we would like to inform you that the Republic of Sudan is going to > change the time zone from (GMT + 3:00) to (GMT+ 2:00) starting from > Wednesday 1 November 2017. > > > this is for your information and action if required. > > > > > > Best Regards > > > > > > Ahmed Atyya > > > Head of Numbering Section > > National Telecomm. Corporation – Sudan > > Phone: +249123499988 > > Email: ahmed.atyya(a)ntc.gov.sd

6 19

Re: [tz] [PROPOSED] Use "PST/PDT" for Philippine time
by Paul Eggert June 24, 2018

June 24, 2018

Michael H Deckers wrote: > My second point was about the abbreviation "PDT", as proposed > for Manila summer time (which you have called "invented"), > not the abbreviation "PST" (which you have not called > "invented" in your proposal). Ah, sorry, I misunderstood. With "PDT" I was attempting to follow this guideline in theory.html: > Use a consistent style in a tz region's history. > For example, if history tends to use numeric > abbreviations and a particular entry could go either way, use a > numeric abbreviation. Since most Philippine history used an alphabetic abbreviation, I thought it better to do so with daylight-saving time too. This is consistent with Hong Kong, Taiwan and Korea, all of which have "JST" for time during Japanese occupation because they use alphabetic abbreviations now, and it is in contrast to other locations that use "+09" for timestamps under Japanese occupation. We have competing guidelines here: "be consistent" and "don't invent abbreviations", and it's a judgment call as to which is more important. PS. Many of those "+09"s are sheer guesswork by Shanks, but that's a different story....

1 0

Philippine Standard Time
by Paul Goyette June 22, 2018

June 22, 2018

In the Philippines, there is a national law, Republic Act No. 10535 which declares the official time here as "Philippine Standard Time". The act [1] even specifies use of PST as the abbreviation, although the FAQ provided by PAGASA [2] uses the "acronym PhST to distinguish it from the Pacific Standard Time (PST)." (Interestingly, the act does nothing to define the actual value of Philippine Standard Time; there is no UTC offset or other reference. It simply declares what the time is called, and instructs government agencies to coordinate with PAGASA for synchronization. :) There are references to "Section 6 of Batas Pambansa Bilang 8" [3] (which directs the adoption of the Metric System), but Section 6 merely delegates the task of establishing the definition of "second" to PAGASA. I've been unable so far to find any specific implementing regulation that establishes an official 8-hour offset from GMT.) Not long ago, the TZ database clean-up of "invented" abbreviations removed the PST abbreviation for Asia/Manila, leaving us with no abbreviation. Based on the existence of national law declaring Philippine Standard Time, and formally establishing the abbreviation/acronym, could we please get "PST" (or, if you prefer, "PhST") re-included into the next update of the TZDB? Thanks! References: [1] http://www.officialgazette.gov.ph/2013/05/15/republic-act-no-10535/ [2] https://www1.pagasa.dost.gov.ph/index.php/astronomy/philippine-standard-tim… [3] http://www.chanrobles.com/bataspambansabilang8.htm +------------------+--------------------------+----------------------------+ | Paul Goyette | PGP Key fingerprint: | E-mail addresses: | | (Retired) | FA29 0E3B 35AF E8AE 6651 | paul at whooppee dot com | | Kernel Developer | 0786 F758 55DE 53BA 7731 | pgoyette at netbsd dot org | +------------------+--------------------------+----------------------------+

6 8