- tz - lists.iana.org

Re: [tz] EU Public Consultation on summertime arrangements
by Pierpaolo Bernardi July 11, 2018

July 11, 2018

Not time related. Only for pedantry sake: at the border between San Marino and Italy there *are* controls, by the Italian police. This is due to different tax laws in the two countries, which would make shopping in San Marino very convenient for Italy residents, if they weren't controlled when exiting from RSM for any good bought in RSM, and forced to pay due taxes. FWIW Il giorno 11 luglio 2018, alle ore 09:08, John Wilcock <john(a)tradoc.fr> ha scritto: Le 11/07/2018 à 00:05, Brian Inglis a écrit : > Most people are probably not that aware of European geography [...] Taking the world as a whole, you are no doubt correct, but most of the people at the EU who will be reading Paul's submission will be perfectly aware of European geography! >> I don't see the relevance of "Schengen area" - all current and potential >> Schengen area countries are either EU or EEA/EFTA members. > The Schengen Area also includes non-EU/EEA/EFTA countries Monaco, San Marino, > and the Vatican, which have not signed the Schengen agreement, but have no > borders, so are de facto included. While these three countries are indeed de facto within the Schengen area due to the lack of border *controls* — they do still have borders ;-) — it still seems (at least to me as a European) very odd to mention Schengen purely in order to include its *non*-signatories, especially as the next item in the list is in any case the all-inclusive "other countries not in the EU". -- John

1 0

(no subject)
by Ashish Ahmed July 10, 2018

July 10, 2018

Confirm

1 0

tz-art suggestion
by Steve Jones July 6, 2018

July 6, 2018

Not sure how this was missed, having passed its third birthday, but probably deserves a mention in tz-art: "Daylight Saving Time: How is this still a thing?" https://www.youtube.com/watch?v=br0NW9ufUUw Regards, Steve Jones [image: Emacs!]

3 2

strftime and nonnormalized dates/times
by scs＠eskimo.com July 5, 2018

July 5, 2018

I have a question about the strftime function. This is not related to time zones per se, but there are a lot of date/time programming experts here, and of course there's a reference implementation of strftime in tzcode. [Full disclosure: I am posting the same question to StackOverflow: https://stackoverflow.com/questions/51145727/ .] Most definitions of the %H, %M, and %S specifiers (including those in the ANSI/ISO C Standards) contain language like %H is replaced by the hour (24-hour clock) as a decimal number (00-23). Now, that little notation "(00-23)" at the end tells you the range of numbers you're going to see *if the struct tm you're printing was just created by localtime or gmtime*. My question concerns the presence or absence of a requirement that the struct tm being referenced by strftime *must* have been one just generated by localtime or gmtime. I don't think there's such a requirement, but if there isn't, those notations "(00-23)" and "(00-59)" don't really have any normative implications for strftime -- they're more of a misplaced requirement on the values generated by localtime and gmtime. But if strftime doesn't restrict the ranges of the various struct tm fields to their "normal" values, it makes me wonder how much latitude a push-the-boundaries programmer actually has when calling strftime. Are the values as unlimited as they are when calling, say, mktime? To make the question more concrete, it came to me when I found myself writing this code: time_t dt = t2 - t1; struct tm *tmp = gmtime(&dt); if(dt > 86400) tmp->tm_hour += dt / 86400 * 24; strftime(etbuf, sizeof(etbuf), "elapsed: %H:%M:%S", tmp); That is, I'm subtracting two times, and letting gmtime and strftime convert the difference to HH:MM:SS format for me. But if the time delta was more than a day, I'm not printing it as days; I'm just lumping it in with the hours (which, it's true, might end up being a very large number if there were months or years between t1 and t2). So, language lawyer question: Is this legal and portable? When calling strftime, are tm_hour, tm_min, and the rest limited to their normal ranges, or not? (Yes, I've tried it, and nonnormalized values work -- not at all surprisingly -- as expected under the popular implementations, but that's not the question.)

4 6

African Timezones
by Robin Roos July 4, 2018

July 4, 2018

Hi Folks, I am using a Java platform (hosted on Linux and on Windows) in which I find some expected timezones to be missing. I write to enquire whether they are defined at IANA and, if not, whether they can be added. The timezones in question are as follows: Africa/Praia (capital city of Cabo Verde, also known also as Cape Verde) = UTC-1 CVT (Cabo Verde Time) = UTC-1 (one hour West of Greenwich) WAT (West Africa Time) = UTC+1 (one hour East of Greenwich) CAT (Central Africa Time) = UTC+2 (two hours East of Greenwich) EAT (East Africa Time) = UTC+3 (three hours East of Greenwich) My source for this information is: https://www.worldtimezone.com/time-africa24.php Many thanks, Robin.

2 2

"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 27, 2018

June 27, 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 23, 2018

June 23, 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

US daylight saving proposals in Texas, California, and Louisiana
by Paul Eggert June 19, 2018

June 19, 2018

This week the Republican Party of Texas is writing its official platform, which contains new planks that call for "abolishing daylight saving time and Texas Child Protective Services". (This quote is from McGaughy's report cited below.) The California State Senate approved a proposal Thursday to move California toward Mountain Standard Time all year (or permanent PDT, if you prefer). The measure would require voter approval in November, plus a second bill and federal approval. The basic idea is similar to the proposal in Florida that Arthur noted in March. Last month the Louisiana legislature overwhelmingly approved a resolution to study a proposal that presumably would also be similar to Florida's, in that Louisiana would move to Eastern Standard Time all year (or permanent CDT, if you prefer). Sources: Bacon-Blood L. Should Louisiana use Daylight Saving Time? State task force will evaluate the options. Times-Picayune. 2018-05-24. https://www.nola.com/politics/index.ssf/2018/05/la_task_force_will_study_me… Luna T. End of daylight saving time inches closer in California. Sacramento Bee. 2018-06-14. https://www.sacbee.com/news/politics-government/capitol-alert/article213187… McGaughy L. New Texas GOP platform calls for abolishing daylight saving, protecting Confederate monuments. Dallas Morning News. 2018-06-18. https://www.dallasnews.com/news/texas-politics/2018/06/16/vote-count-underw…

1 0

[PROPOSED] Merge some changes from man-pages project
by Paul Eggert June 19, 2018

June 19, 2018

Make some minor formatting changes to be more like man-pages. * tzfile.5: Break a line the same place that man-pages does. * zdump.8, zic.8: Use bold, not italics, for command and function names. Do not capitalize command names. Add a section header OPTIONS. Document --help. Fix some problems with fonts. Don’t quote "SEE ALSO", and don’t mention newctime(3). Put FILES before NOTES, and format a la man-pages. * zic.8: Also, do some fixes of our own, noticed while merging the above. Use real life US Rule, not an invented one. Write “UTOFF”, not “GMTOFF” (man-pages used “UTCOFF” but it’s not necessarily UTC either). Quote more carefully. --- tzfile.5 | 3 +- zdump.8 | 22 ++++++++------ zic.8 | 88 ++++++++++++++++++++++++++++++++------------------------ 3 files changed, 66 insertions(+), 47 deletions(-) diff --git a/tzfile.5 b/tzfile.5 index 13213d2..2e6c3aa 100644 --- a/tzfile.5 +++ b/tzfile.5 @@ -181,7 +181,8 @@ use two minor extensions to the POSIX TZ format, as described in .BR newtzset (3). First, the hours part of its transition times may be signed and range from \*-167 through 167 instead of the POSIX-required unsigned values -from 0 through 24. Second, DST is in effect all year if it starts +from 0 through 24. +Second, DST is in effect all year if it starts January 1 at 00:00 and ends December 31 at 24:00 plus the difference between daylight saving and standard time. .PP diff --git a/zdump.8 b/zdump.8 index 833dd54..92e9e62 100644 --- a/zdump.8 +++ b/zdump.8 @@ -18,16 +18,19 @@ zdump \- time zone dumper .. .ie \n(.g .ds - \f(CW-\fP .el ds - \- -.I Zdump -prints the current time in each +The +.B zdump +program prints the current time in each .I zonename named on the command line. -.PP -These options are available: +.SH OPTIONS .TP -.BI "\*-\*-version" +.B \*-\*-version Output version information and exit. .TP +.B \*-\*-help +Output short usage message and exit. +.TP .B \*-i .I "(This option is experimental: its behavior may change in future versions.)" Output a description of time intervals. For each @@ -68,7 +71,7 @@ except omit the times relative to the extreme time values. This generates output that is easier to compare to that of implementations with different time representations. .TP -.BI "\*-c " [loyear,]hiyear +.BI "\*-c " \fR[\fIloyear , \fR]\fIhiyear Cut off interval output at the given year(s). Cutoff times are computed using the proleptic Gregorian calendar with year 0 and with Universal Time (UT) ignoring leap seconds. @@ -80,7 +83,7 @@ of 1970 includes the transition. The default cutoff is .BR \*-500,2500 . .TP -.BI "\*-t " [lotime,]hitime +.BI "\*-t " \fR[\fIlotime , \fR]\fIhitime Cut off interval output at the given time(s), given in decimal seconds since 1970-01-01 00:00:00 Coordinated Universal Time (UTC). @@ -220,7 +223,8 @@ for newer and .q "UT" for older timestamps, partly because the exact date of the introduction of UTC is problematic. -.SH "SEE ALSO" -newctime(3), tzfile(5), zic(8) +.SH SEE ALSO +.BR tzfile (5), +.BR zic (8) .\" This file is in the public domain, so clarified as of .\" 2009-05-17 by Arthur David Olson. diff --git a/zic.8 b/zic.8 index 00d4267..3e33618 100644 --- a/zic.8 +++ b/zic.8 @@ -28,27 +28,30 @@ zic \- time zone compiler . ds : . ds - \- .\} -.I Zic -reads text from the file(s) named on the command line +The +.B zic +program reads text from the file(s) named on the command line and creates the time conversion information files specified in this input. If a .I filename is .q "\*-" , -the standard input is read. -.PP -These options are available: +standard input is read. +.SH OPTIONS .TP -.BI "\*-\*-version" +.B "\*-\*-version" Output version information and exit. .TP +.B \*-\*-help +Output short usage message and exit. +.TP .BI "\*-d " directory Create time conversion information files in the named directory rather than in the standard directory named below. .TP .BI "\*-l " timezone Use the given time zone as local time. -.I Zic +.B zic will act as if the input contained a link line of the form .sp .ti +.5i @@ -57,7 +60,7 @@ Link \fItimezone\fP localtime .BI "\*-p " timezone Use the given time zone's rules when handling POSIX-format time zone environment variables. -.I Zic +.B zic will act as if the input contained a link line of the form .sp .ti +.5i @@ -80,17 +83,17 @@ The input specifies a link to a link. .PP A year that appears in a data file is outside the range of years representable by -.IR time (2) +.BR time (2) values. .PP A time of 24:00 or more appears in the input. Pre-1998 versions of -.I zic +.B zic prohibit 24:00, and pre-2007 versions prohibit times greater than 24:00. .PP A rule goes past the start or end of the month. Pre-2004 versions of -.I zic +.B zic prohibit this. .PP The output file does not contain all the information about the @@ -102,7 +105,7 @@ represented. .PP The output contains data that may not be handled properly by client code designed for older -.I zic +.B zic output formats. These compatibility issues affect only timestamps before 1970 or after the start of 2038. .PP @@ -148,7 +151,7 @@ to the end of the line the sharp character appears on. White space characters and sharp characters may be enclosed in double quotes (") if they're to be used as part of a field. Any line that is blank (after comment stripping) is ignored. -Non-blank lines are expected to be of one of three types: +Nonblank lines are expected to be of one of three types: rule lines, zone lines, and link lines. .PP Names must be in English and are case insensitive. @@ -165,14 +168,14 @@ abbreviation must be unambiguous in context. A rule line has the form .nf .ti +.5i -.ta \w'Rule\0\0'u +\w'NAME\0\0'u +\w'FROM\0\0'u +\w'1973\0\0'u +\w'TYPE\0\0'u +\w'Apr\0\0'u +\w'lastSun\0\0'u +\w'2:00s\0\0'u +\w'1:00d\0\0'u +.ta \w'Rule\0\0'u +\w'NAME\0\0'u +\w'FROM\0\0'u +\w'1973\0\0'u +\w'TYPE\0\0'u +\w'Apr\0\0'u +\w'lastSun\0\0'u +\w'2:00w\0\0'u +\w'1:00d\0\0'u .sp Rule NAME FROM TO TYPE IN ON AT SAVE LETTER/S .sp For example: .ti +.5i .sp -Rule US 1967 1973 \*- Apr lastSun 2:00s 1:00d D +Rule US 1967 1973 \*- Apr lastSun 2:00w 1:00d D .sp .fi The fields that make up a rule line are: @@ -220,7 +223,7 @@ field. should be .q \*- and is present for compatibility with older versions of -.I zic +.B zic in which it could contain year types. .TP .B IN @@ -274,7 +277,7 @@ Recognized forms include: where hour 0 is midnight at the start of the day, and hour 24 is midnight at the end of the day. Although -.I zic +.B zic rounds times to the nearest integer second (breaking ties to the even integer), the fractions may be useful to other applications requiring greater precision. @@ -321,7 +324,7 @@ Negative offsets are allowed; in Ireland, for example, daylight saving time is observed in winter and has a negative offset relative to Irish Standard Time. The offset is merely added to standard time; for example, -.I zic +.B zic does not distinguish a 10:30 standard time plus an 0:30 .B SAVE from a 10:00 standard time plus a 1:00 @@ -347,8 +350,8 @@ A zone line has the form .sp .nf .ti +.5i -.ta \w'Zone\0\0'u +\w'Asia/Amman\0\0'u +\w'GMTOFF\0\0'u +\w'Jordan\0\0'u +\w'FORMAT\0\0'u -Zone NAME GMTOFF RULES FORMAT [UNTIL] +.ta \w'Zone\0\0'u +\w'Asia/Amman\0\0'u +\w'UTOFF\0\0'u +\w'Jordan\0\0'u +\w'FORMAT\0\0'u +Zone NAME UTOFF RULES FORMAT [UNTIL] .sp For example: .sp @@ -357,7 +360,7 @@ Zone Asia/Amman 2:00 Jordan EE%sT 2017 Oct 27 01:00 .sp .fi The fields that make up a zone line are: -.TP "\w'GMTOFF'u" +.TP "\w'UTOFF'u" .B NAME The name of the time zone. This is the name used in creating the time conversion information file for the @@ -369,7 +372,7 @@ or a file name component is a maximal substring that does not contain .q "/" . .TP -.B GMTOFF +.B UTOFF The amount of time to add to UT to get standard time in this zone. This field has the same format as the .B AT @@ -413,7 +416,10 @@ Alternatively, a slash (/) separates standard and daylight abbreviations. To conform to POSIX, a time zone abbreviation should contain only -alphanumeric ASCII characters, "+" and "\*-". +alphanumeric ASCII characters, +.q "+" +and +.q "\*-". .TP .B UNTIL The time at which the UT offset or the rule(s) change for a location. @@ -518,7 +524,7 @@ if the leap second time given by the other fields should be interpreted as local wall clock time. .SH "EXTENDED EXAMPLE" Here is an extended example of -.I zic +.B zic input, intended to illustrate many of its features. In this example, the EU rules are for the European Union and for its predecessor organization, the European Communities. @@ -539,8 +545,8 @@ Rule EU 1979 1995 \*- Sep lastSun 1:00u 0 \*- Rule EU 1981 max \*- Mar lastSun 1:00u 1:00 S Rule EU 1996 max \*- Oct lastSun 1:00u 0 \*- .sp -.ta \w'# Zone\0\0'u +\w'Europe/Zurich\0\0'u +\w'GMTOFF\0\0'u +\w'0:34:08\0\0'u +\w'FORMAT\0\0'u -# Zone NAME GMTOFF RULES FORMAT [UNTIL] +.ta \w'# Zone\0\0'u +\w'Europe/Zurich\0\0'u +\w'0:34:08\0\0'u +\w'RULES\0\0'u +\w'FORMAT\0\0'u +# Zone NAME UTOFF RULES FORMAT [UNTIL] Zone Europe/Zurich 0:34:08 \*- LMT 1853 Jul 16 0:29:46 \*- BMT 1894 Jun 1:00 Swiss CE%sT 1981 @@ -556,8 +562,9 @@ seconds east of UT until 1853-07-16 at 00:00, when the legal offset was changed to 7\(de\|26\(fm\|22.50\(sd; although this works out to 0:29:45.50, the input format cannot represent fractional seconds so it is rounded here. After 1894-06-01 at 00:00 the UT offset became one hour -and Swiss daylight saving rules (defined with lines beginning with "Rule -Swiss") apply. From 1981 to the present, EU daylight saving rules have +and Swiss daylight saving rules (defined with lines beginning with +.q "Rule Swiss") +apply. From 1981 to the present, EU daylight saving rules have applied, and the UTC offset has remained at one hour. .PP In 1941 and 1942, daylight saving time applied from the first Monday @@ -568,11 +575,21 @@ saving has begun on the last Sunday in March at 01:00 UTC. Until 1995 it ended the last Sunday in September at 01:00 UTC, but this changed to the last Sunday in October starting in 1996. .PP -For purposes of -display, "LMT" and "BMT" were initially used, respectively. Since +For purposes of display, +.q "LMT" +and +.q "BMT" +were initially used, respectively. Since Swiss rules and later EU rules were applied, the display name for the time zone has been CET for standard time and CEST for daylight saving time. +.SH FILES +.TP +.I /etc/localtime +Default local time zone file. +.TP +.I /usr/share/zoneinfo +Default time zone information directory. .SH NOTES For areas with more than two types of local time, you may need to use local standard time in the @@ -585,17 +602,14 @@ for a particular zone, a clock advance caused by the start of daylight saving coincides with and is equal to a clock retreat caused by a change in UT offset, -.IR zic +.B zic produces a single transition to daylight saving at the new UT offset (without any change in wall clock time). To get separate transitions use multiple zone continuation lines specifying transition instants using universal time. -.SH FILES -.ta \w'/usr/share/zoneinfo\0\0'u -/etc/localtime default local time zone file -/usr/share/zoneinfo default time zone information directory -.SH "SEE ALSO" -newctime(3), tzfile(5), zdump(8) +.SH SEE ALSO +.BR tzfile (5), +.BR zdump (8) .\" This file is in the public domain, so clarified as of .\" 2009-05-17 by Arthur David Olson. -- 2.17.1

1 0

[PROPOSED] Remove Big Bang hack
by Paul Eggert June 13, 2018

June 13, 2018

The Big Bang hack introduced in 2013d should no longer be needed, now that type 0 is the default type and types are reordered to make this happen. I.e., since we now have a better fix for the Macquarie Island problem reported by Alan Gutierrez in <https://mm.icann.org/pipermail/tz/2013-May/019368.html>, we can revert much of the 2013-05-23 12:19:13 -0700 change “Rework to avoid problem with Australia/Macquarie on 32-bit hosts” and the 2014-05-01 16:32:16 -0700 followup change “zic: improve handling of low-valued time stamps” which caused zic to insert a dummy transition. As a result, zic now generates TZif files that are typically a few bytes shorter, as they were in 2013c and earlier. * Makefile, zic.8: Remove mention of Big Bang. * NEWS: Mention. * zic.c (BIG_BANG, WORK_AROUND_CLIENT_OVERFLOW_BUGS, early_time): Remove. All uses removed. (writezone): New arg DEFAULTTYPE. All callers changed. Simplify by not inserting a dummy transition before the Big Bang. (outzone): Compute default type for use by writezone. (addtt): Don’t second-guess type 0. --- Makefile | 1 - NEWS | 6 +++ zic.8 | 6 --- zic.c | 151 ++++++++++++++++----------------------------------------------- 4 files changed, 44 insertions(+), 120 deletions(-) diff --git a/Makefile b/Makefile index 3705fc0..2289dff 100644 --- a/Makefile +++ b/Makefile @@ -173,7 +173,6 @@ LDLIBS= # Add the following to the end of the "CFLAGS=" line as needed to override # defaults specified in the source code. "-DFOO" is equivalent to "-DFOO=1". -# -DBIG_BANG=-9999999LL if the Big Bang occurred at time -9999999 (see zic.c) # -DDEPRECATE_TWO_DIGIT_YEARS for optional runtime warnings about strftime # formats that generate only the last two digits of year numbers # -DEPOCH_LOCAL if the 'time' function returns local time not UT diff --git a/NEWS b/NEWS index b90e68c..87f675e 100644 --- a/NEWS +++ b/NEWS @@ -27,6 +27,12 @@ Unreleased, experimental changes legacy zones EST5EDT, CST6CDT, MST7MDT, PST8PDT, CET, MET, and EET, which previously used nonzero types for these timestamps. + Because of the type 0 change, zic no longer outputs a dummy + transition at time -2**59 (before the Big Bang), as clients should + no longer need this to handle historical timestamps correctly. + This reverts a change introduced in 2013d and shrinks most TZif + files by a few bytes. + localtime.c no longer requires at least one time type in TZif files that lack transitions or have a POSIX-style TZ string. This future-proofs the code against possible future extensions to the diff --git a/zic.8 b/zic.8 index ba4796e..00d4267 100644 --- a/zic.8 +++ b/zic.8 @@ -591,12 +591,6 @@ produces a single transition to daylight saving at the new UT offset To get separate transitions use multiple zone continuation lines specifying transition instants using universal time. -.PP -timestamps well before the Big Bang are silently omitted from the output. -This works around bugs in software that mishandles large negative time -stamps. Call it sour grapes, but pre-Big-Bang timestamps are -physically suspect anyway. The pre-Big-Bang cutoff time is -approximate and may change in future versions. .SH FILES .ta \w'/usr/share/zoneinfo\0\0'u /etc/localtime default local time zone file diff --git a/zic.c b/zic.c index fd73f66..88153db 100644 --- a/zic.c +++ b/zic.c @@ -948,48 +948,6 @@ dolink(char const *fromfield, char const *tofield, bool staysymlink) static zic_t const min_time = MINVAL(zic_t, TIME_T_BITS_IN_FILE); static zic_t const max_time = MAXVAL(zic_t, TIME_T_BITS_IN_FILE); -/* Estimated time of the Big Bang, in seconds since the POSIX epoch. - rounded downward to the negation of a power of two that is - comfortably outside the error bounds. - - For the time of the Big Bang, see: - - Ade PAR, Aghanim N, Armitage-Caplan C et al. Planck 2013 results. - I. Overview of products and scientific results. - arXiv:1303.5062 2013-03-20 20:10:01 UTC - <https://arxiv.org/pdf/1303.5062v1> [PDF] - - Page 36, Table 9, row Age/Gyr, column Planck+WP+highL+BAO 68% limits - gives the value 13.798 plus-or-minus 0.037 billion years. - Multiplying this by 1000000000 and then by 31557600 (the number of - seconds in an astronomical year) gives a value that is comfortably - less than 2**59, so BIG_BANG is - 2**59. - - BIG_BANG is approximate, and may change in future versions. - Please do not rely on its exact value. */ - -#ifndef BIG_BANG -#define BIG_BANG (- (1LL << 59)) -#endif - -/* If true, work around GNOME glib bug 878 - <https://gitlab.gnome.org/GNOME/glib/issues/878> - by refusing to output timestamps before BIG_BANG. - Such timestamps are physically suspect anyway. - The glib bug is scheduled to be fixed in glib 2.58. - There is a similar bug in 32-bit glibc 2.27 and earlier - <https://sourceware.org/bugzilla/show_bug.cgi?id=19738> - that is scheduled to be fixed in glibc 2.28; - unfortunately a similar bug might persist in 64-bit glibc - though this has not been verified. With all this in mind, - let's hope that the workaround will no longer be needed - in the year 2025. */ -enum { WORK_AROUND_CLIENT_OVERFLOW_BUGS = true }; - -static const zic_t early_time = (WORK_AROUND_CLIENT_OVERFLOW_BUGS - ? BIG_BANG - : MINVAL(zic_t, TIME_T_BITS_IN_FILE)); - /* Return true if NAME is a directory. */ static bool itsdir(char const *name) @@ -1775,7 +1733,8 @@ swaptypes(int i, int j) } static void -writezone(const char *const name, const char *const string, char version) +writezone(const char *const name, const char *const string, char version, + int defaulttype) { register FILE * fp; register ptrdiff_t i, j; @@ -1805,13 +1764,11 @@ writezone(const char *const name, const char *const string, char version) toi = 0; fromi = 0; - while (fromi < timecnt && attypes[fromi].at < early_time) - ++fromi; for ( ; fromi < timecnt; ++fromi) { - if (toi > 1 && ((attypes[fromi].at + + if (toi != 0 && ((attypes[fromi].at + gmtoffs[attypes[toi - 1].type]) <= - (attypes[toi - 1].at + - gmtoffs[attypes[toi - 2].type]))) { + (attypes[toi - 1].at + gmtoffs[toi == 1 ? 0 + : attypes[toi - 2].type]))) { attypes[toi - 1].type = attypes[fromi].type; continue; @@ -1919,8 +1876,8 @@ writezone(const char *const name, const char *const string, char version) for (pass = 1; pass <= 2; ++pass) { register ptrdiff_t thistimei, thistimecnt, thistimelim; register int thisleapi, thisleapcnt, thisleaplim; - int old0, new0; - char omittype[TZ_MAX_TYPES + 1]; + int old0; + char omittype[TZ_MAX_TYPES]; int typemap[TZ_MAX_TYPES]; register int thistypecnt; char thischars[TZ_MAX_CHARS]; @@ -1945,45 +1902,17 @@ writezone(const char *const name, const char *const string, char version) error(_("too many transition times")); thistimelim = thistimei + thistimecnt; thisleaplim = thisleapi + thisleapcnt; - memset(omittype, thistimecnt != timecnt, typecnt); - if (thistimecnt == 0) { - /* - ** No transition times fall in the current - ** (32- or 64-bit) window. - */ - if (typecnt != 0) - omittype[typecnt - 1] = false; - } else { - for (i = thistimei - 1; i < thistimelim; ++i) - if (i >= 0) - omittype[types[i]] = false; - /* - ** For America/Godthab and Antarctica/Palmer - */ - if (thistimei == 0) - omittype[0] = false; - } - - /* Reorder types to make type 0 the first-used standard type - if one is used, otherwise the first-used DST type if one - is used, otherwise no reordering. Use TYPEMAP to swap - OLD0 and NEW0 so that NEW0 appears as type 0 in the output - instead of OLD0. TYPEMAP also omits unused types. */ - omittype[typecnt] = false; /* strlen sentinel */ + memset(omittype, true, typecnt); + omittype[defaulttype] = false; + for (i = thistimei; i < thistimelim; i++) + omittype[types[i]] = false; + + /* Reorder types to make DEFAULTTYPE type 0. + Use TYPEMAP to swap OLD0 and DEFAULTTYPE so that + DEFAULTTYPE appears as type 0 in the output instead + of OLD0. TYPEMAP also omits unused types. */ old0 = strlen(omittype); - new0 = -1; - for (i = 0; i < timecnt; i++) - if (!omittype[types[i]]) { - if (! isdsts[types[i]]) { - new0 = types[i]; - break; - } - if (new0 < 0) - new0 = types[i]; - } - if (new0 < 0) - new0 = old0; - swaptypes(old0, new0); + swaptypes(old0, defaulttype); #ifndef LEAVE_SOME_PRE_2011_SYSTEMS_IN_THE_LURCH /* @@ -2035,7 +1964,8 @@ writezone(const char *const name, const char *const string, char version) thistypecnt = 0; for (i = old0; i < typecnt; i++) if (!omittype[i]) - typemap[i == old0 ? new0 : i == new0 ? old0 : i] + typemap[i == old0 ? defaulttype + : i == defaulttype ? old0 : i] = thistypecnt++; for (i = 0; i < sizeof indmap / sizeof indmap[0]; ++i) @@ -2133,7 +2063,7 @@ writezone(const char *const name, const char *const string, char version) for (i = old0; i < typecnt; i++) if (!omittype[i]) putc(ttisgmts[i], fp); - swaptypes(old0, new0); + swaptypes(old0, defaulttype); } fprintf(fp, "\n%s\n", string); close_file(fp, directory, name); @@ -2487,6 +2417,7 @@ outzone(const struct zone *zpfirst, ptrdiff_t zonecount) zic_t one = 1; zic_t y2038_boundary = one << 31; zic_t max_year0; + int defaulttype = -1; max_abbr_len = 2 + max_format_len + max_abbrvar_len; max_envvar_len = 2 * max_abbr_len + 5 * 9; @@ -2595,9 +2526,9 @@ outzone(const struct zone *zpfirst, ptrdiff_t zonecount) */ stdoff = 0; zp = &zpfirst[i]; - usestart = i > 0 && (zp - 1)->z_untiltime > early_time; + usestart = i > 0 && (zp - 1)->z_untiltime > min_time; useuntil = i < (zonecount - 1); - if (useuntil && zp->z_untiltime <= early_time) + if (useuntil && zp->z_untiltime <= min_time) continue; gmtoff = zp->z_gmtoff; eat(zp->z_filename, zp->z_linenum); @@ -2612,7 +2543,8 @@ outzone(const struct zone *zpfirst, ptrdiff_t zonecount) if (usestart) { addtt(starttime, type); usestart = false; - } else addtt(early_time, type); + } else + defaulttype = type; } else for (year = min_year; year <= max_year; ++year) { if (useuntil && year > zp->z_untilrule.r_hiyear) break; @@ -2726,6 +2658,8 @@ outzone(const struct zone *zpfirst, ptrdiff_t zonecount) offset = oadd(zp->z_gmtoff, rp->r_stdoff); type = addtype(offset, ab, rp->r_isdst, rp->r_todisstd, rp->r_todisgmt); + if (defaulttype < 0 && !rp->r_isdst) + defaulttype = type; if (rp->r_hiyear == ZIC_MAX && ! (0 <= lastatmax && ktime < attypes[lastatmax].at)) @@ -2742,11 +2676,14 @@ outzone(const struct zone *zpfirst, ptrdiff_t zonecount) eat(zp->z_filename, zp->z_linenum); if (*startbuf == '\0') error(_("can't determine time zone abbreviation to use just after until time")); - else addtt(starttime, - addtype(startoff, startbuf, - startoff != zp->z_gmtoff, - startttisstd, - startttisgmt)); + else { + bool isdst = startoff != zp->z_gmtoff; + type = addtype(startoff, startbuf, isdst, + startttisstd, startttisgmt); + if (defaulttype < 0 && !isdst) + defaulttype = type; + addtt(starttime, type); + } } /* ** Now we may get to set starttime for the next zone line. @@ -2761,6 +2698,8 @@ error(_("can't determine time zone abbreviation to use just after until time")); starttime = tadd(starttime, -gmtoff); } } + if (defaulttype < 0) + defaulttype = 0; if (0 <= lastatmax) attypes[lastatmax].dontmerge = true; if (do_extend) { @@ -2788,7 +2727,7 @@ error(_("can't determine time zone abbreviation to use just after until time")); attypes[timecnt - 1].dontmerge = true; } } - writezone(zpfirst->z_name, envvar, version); + writezone(zpfirst->z_name, envvar, version, defaulttype); free(startbuf); free(ab); free(envvar); @@ -2797,20 +2736,6 @@ error(_("can't determine time zone abbreviation to use just after until time")); static void addtt(zic_t starttime, int type) { - if (starttime <= early_time - || (timecnt == 1 && attypes[0].at < early_time)) { - gmtoffs[0] = gmtoffs[type]; - isdsts[0] = isdsts[type]; - ttisstds[0] = ttisstds[type]; - ttisgmts[0] = ttisgmts[type]; - if (abbrinds[type] != 0) - strcpy(chars, &chars[abbrinds[type]]); - abbrinds[0] = 0; - charcnt = strlen(chars) + 1; - typecnt = 1; - timecnt = 0; - type = 0; - } attypes = growalloc(attypes, sizeof *attypes, timecnt, &timecnt_alloc); attypes[timecnt].at = starttime; attypes[timecnt].dontmerge = false; -- 2.7.4

1 0

[PROPOSED] GNOME glib bug fix installed
by Paul Eggert June 13, 2018

June 13, 2018

The bug fix was merged into glib master today. Since it’s four years after the fix was proposed (!), push back an estimate of when we can remove the workaround. Also, adjust to GNOME bug renumbering. * NEWS: Adjust bug number in history. * zic.c (WORK_AROUND_GLIB_BUG_878): Rename from WORK_AROUND_GNOME_BUG_730332. All uses changed. --- NEWS | 4 ++-- zic.c | 14 +++++++------- 2 files changed, 9 insertions(+), 9 deletions(-) diff --git a/NEWS b/NEWS index 34f7810..b90e68c 100644 --- a/NEWS +++ b/NEWS @@ -2106,8 +2106,8 @@ Release 2014d - 2014-05-27 21:34:40 -0700 Changes affecting code zic no longer generates files containing timestamps before the Big Bang. - This works around GNOME bug 730332 - <https://bugzilla.gnome.org/show_bug.cgi?id=730332>. + This works around GNOME glib bug 878 + <https://gitlab.gnome.org/GNOME/glib/issues/878> (Thanks to Leonardo Chiquitto for reporting the bug, and to Arthur David Olson and James Cloos for suggesting improvements to the fix.) diff --git a/zic.c b/zic.c index ce9ccaf..9e25b02 100644 --- a/zic.c +++ b/zic.c @@ -972,17 +972,17 @@ static zic_t const max_time = MAXVAL(zic_t, TIME_T_BITS_IN_FILE); #define BIG_BANG (- (1LL << 59)) #endif -/* If true, work around GNOME bug 730332 - <https://bugzilla.gnome.org/show_bug.cgi?id=730332> +/* If true, work around GNOME glib bug 878 + <https://gitlab.gnome.org/GNOME/glib/issues/878> by refusing to output timestamps before BIG_BANG. Such timestamps are physically suspect anyway. - The GNOME bug is scheduled to be fixed in GNOME 3.22, and if so - this workaround will no longer be needed when GNOME 3.21 and - earlier are obsolete, say in the year 2021. */ -enum { WORK_AROUND_GNOME_BUG_730332 = true }; + The glib bug is scheduled to be fixed in glib 2.58, and if so + this workaround will no longer be needed when glib 2.57 and + earlier are obsolete, say in the year 2025. */ +enum { WORK_AROUND_GLIB_BUG_878 = true }; -static const zic_t early_time = (WORK_AROUND_GNOME_BUG_730332 +static const zic_t early_time = (WORK_AROUND_GLIB_BUG_878 ? BIG_BANG : MINVAL(zic_t, TIME_T_BITS_IN_FILE)); -- 2.17.1

2 2

Clarification on LMT entries
by Matt Johnson June 8, 2018

June 8, 2018

Quick question. The entry for Europe/Kiev has two solar mean time entries: Zone Europe/Kiev 2:02:04 - LMT 1880 2:02:04 - KMT 1924 May 2 # Kiev Mean Time While I know how to read this, I'm not sure why the abbreviation KMT exists. Is this another invented abbreviation, or is there a source? Also, why 1880 on the LMT line? What is special about 1880? Thanks, Matt

4 3

Secure FTP support for IANA downloads
by pbassey＠ups.com June 6, 2018

June 6, 2018

Hello, I am writing to inquire about your site's use of FTP and the possibility of migrating over to Secure FTP in order to comply with our company's security compliance requirements. Does your site offer Secure FTP for your IANA downloads and if not, what would it take to enable this protocol so our company could securely download your files? Currently, we use this URL to download the IANA time zone files which is not secure: ftp://ftp.iana.org/tz/ Thank you in advance for your help. Regards, Peter Bassey Peter A. Bassey Enterprise Business Intelligence UPS 340 MacArthur Blvd Mahwah, NJ 07430 Phone: 201-828-7847 Mobile: 845-662-3743 pbassey(a)ups.com<mailto:pbassey@ups.com>

6 6

On solving the tzdata changes problem
by Aldrin Martoq Ahumada June 5, 2018

June 5, 2018

Hello, I think we can semi-automate when a tzdata change occurs in software, given the following: 1. Your software stores all dates as unix timestamps (seconds since the epoch, “ignoring” leap seconds). 2. Your system (or users) choose a local representation using an oficial timezone, like ‘America/Santiago’. 3. Your system stores somewhere what version of tzdata is currently being used, like ‘2018e’. So, when a new version of tzdata is released (say ‘2018f’), the sysadmin updates their tzdata packages, and then runs a tool that checks for every timezone used (‘America/Santiago’ for ex) if there are changes that could affect users between 2018e and 2018f. The typical example: my dentist appointment is not correct anymore. So, your system may ask the user: “There have been changes in the timezone where you live, would you like to review this 3 appointments and check if they are still correct?" The problem is that there is no historical data of tzdata: when a sysadmin updates the system, it gets a new version 2018f, but there is no automatic way of knowing the differences. I didn’t find one, so I wrote it. A web app that demo this concept is here: https://a0.github.io/a0-tzmigration/demo/ <https://a0.github.io/a0-tzmigration/demo/> You can even share the url, for example here is the diff between America/Santiago 2016j and America/Punta_Arenas 2017a: https://a0.github.io/a0-tzmigration/demo/?ta=America%2FSantiago&va=2016j&tb… <https://a0.github.io/a0-tzmigration/demo/?ta=America/Santiago&va=2016j&tb=A…> This can automate many issues we all have when updating tzdata. For the link above, a whole region of my country switched so the new America/Punta_Arenas timezone was created. You can easily spot errors, too: the America/Punta_Arenas should be equal to America/Santiago up to 2017, but there are differences in the years 1890, 1946. And you can do this, in your own systems, using the same data, which is here: https://a0.github.io/a0-tzmigration-ruby/data/ <https://a0.github.io/a0-tzmigration-ruby/data/>. It’s not perfect. It probably has errors. I generated it from the ruby tzinfo-data gem, which has info only from 2013c onwards, and transitions up to ≈ 2068 (50 years from now). Use with caution, happy to accept corrections. I’m writing a ruby gem and an npm package, would love to see other language implementations: https://github.com/a0/a0-tzmigration-ruby <https://github.com/a0/a0-tzmigration-ruby> https://github.com/a0/a0-tzmigration-js <https://github.com/a0/a0-tzmigration-js> API is not really stable, will check and release documentation next week, but the core idea is there: grab the JSON files from our github page and get the changes that MAY have to be applied. I’m writing a medium blog post to cover all of this. Hope that this help all of you, and happy to listen any suggestions or corrections. Cheers, — Aldrin.

5 6

[PROPOSED 1/3] Don’t reject TZif TZ strings that lack DST
by Paul Eggert June 5, 2018

June 5, 2018

* NEWS: Mention this. * localtime.c (tzloadbody): Don’t require a TZif file’s POSIX TZ string to specify DST. That way, the TZ string consistently overrides the type of the TZif file’s last transition, instead of overriding only when specifying DST. Although this currently matters only for invalid TZif files where the TZ string disagrees with the last transition’s time type, it will become more important if we stop requiring at least one time type in the binary data, as the TZ string will then be the only source for time types. --- NEWS | 6 ++++++ localtime.c | 11 +++++------ 2 files changed, 11 insertions(+), 6 deletions(-) diff --git a/NEWS b/NEWS index ce009e7..2e9bf99 100644 --- a/NEWS +++ b/NEWS @@ -27,6 +27,12 @@ Unreleased, experimental changes legacy zones EST5EDT, CST6CDT, MST7MDT, PST8PDT, CET, MET, and EET, which previously used nonzero types for these timestamps. + localtime.c no longer ignores TZif POSIX-style TZ strings that + specify only standard time. Instead, these TZ strings now + override the default time type for timestamps after the last + transition (or for all time stamps if there are no transitions), + just as DST strings specifying DST have always done. + Changes to documentation New restrictions: A Rule name must start with a character that diff --git a/localtime.c b/localtime.c index 21160c0..5a3fe91 100644 --- a/localtime.c +++ b/localtime.c @@ -587,8 +587,7 @@ tzloadbody(char const *name, struct state *sp, bool doextend, struct state *ts = &lsp->u.st; up->buf[nread - 1] = '\0'; - if (tzparse(&up->buf[1], ts, false) - && ts->typecnt == 2) { + if (tzparse(&up->buf[1], ts, false)) { /* Attempt to reuse existing abbreviations. Without this, America/Anchorage would be right on @@ -599,7 +598,7 @@ tzloadbody(char const *name, struct state *sp, bool doextend, stay 40 in this example. */ int gotabbr = 0; int charcnt = sp->charcnt; - for (i = 0; i < 2; i++) { + for (i = 0; i < ts->typecnt; i++) { char *tsabbr = ts->chars + ts->ttis[i].tt_abbrind; int j; for (j = 0; j < charcnt; j++) @@ -618,7 +617,7 @@ tzloadbody(char const *name, struct state *sp, bool doextend, } } } - if (gotabbr == 2) { + if (gotabbr == ts->typecnt) { sp->charcnt = charcnt; /* Ignore any trailing, no-op transitions generated @@ -640,8 +639,8 @@ tzloadbody(char const *name, struct state *sp, bool doextend, sp->timecnt++; i++; } - sp->ttis[sp->typecnt++] = ts->ttis[0]; - sp->ttis[sp->typecnt++] = ts->ttis[1]; + for (i = 0; i < ts->typecnt; i++) + sp->ttis[sp->typecnt++] = ts->ttis[i]; } } } -- 2.17.1

1 2

UTC is enough for everyone, right?
by Philip Paeps June 4, 2018

June 4, 2018

I've just read this amusing introduction to the vagaries of dealing with time. A fun read. https://zachholman.com/talk/utc-is-enough-for-everyone-right I've not watched the accompanying presentations but the article is quite insightful. May merit inclusion in tz-link.html? Philip -- Philip Paeps Senior Reality Engineer Ministry of Information

1 0

[PROPOSED] Use type 0 for timestamps before first transition
by Paul Eggert June 3, 2018

June 3, 2018

I ran into this problem when helping to draft: https://tools.ietf.org/html/draft-murchison-tzdist-tzif-04 Its section 3.2 gives a complex four-step procedure for determining the type of timestamps before the first transition, derived from 2018e localtime.c. However, I just now checked, and glibc does only the last two steps. Furthermore, in practice nowadays only the last step is needed in almost all zones, because type 0 is almost always the right type to use; the only exceptions are the legacy zones EST5EDT, CST6CDT, MST7MDT, PST8PDT, CET, MET, and EET. As the tzif-04 section 3.2 procedure is only for backward compatibility with older TZif files, now is a good time to clean up by specifying a simple way to deal with time stamps before the first transition, namely, to use time type 0. So, this patch changes zic so that the first-used standard time type (if there is one) is always time type 0. This is compatible with glibc and with any other program that uses the heuristic of using the lowest-numbered standard-time time type for timestamps before the first transition, since it merely changes zic to output a standard-time type 0. * NEWS, tzfile.5: Document this. * localtime.c (tzparse): Make standard time type 0, if standard time is in use. This should help future-proof this program if we ever get rid of the defaulttype member and just use 0. * zic.c (swaptypes): New function. (writezone): Reorder types to make type 0 the default type. Negate the local variable writetypes into omittypes, so that strlen can be used on it, and enlarge it by 1 to make room for a sentinel. Use memset where appropriate. --- NEWS | 13 +++++++++ localtime.c | 43 ++++++++++++++++++++++-------- tzfile.5 | 17 ++++++++---- zic.c | 77 +++++++++++++++++++++++++++++++++++++++-------------- 4 files changed, 114 insertions(+), 36 deletions(-) diff --git a/NEWS b/NEWS index 0ef70a1..ce009e7 100644 --- a/NEWS +++ b/NEWS @@ -17,6 +17,16 @@ Unreleased, experimental changes observed DST in 1942/79, not 1961/80, and there were several errors for transition times and dates. (Thanks to P Chan.) + Changes to code + + zic now always generates TZif files where time type 0 is used for + timestamps before the first transition. This simplifies the + reading of TZif files and should not affect behavior of existing + TZif readers because the same set of time types is used; only + their internal indexes may have changed. This affects only the + legacy zones EST5EDT, CST6CDT, MST7MDT, PST8PDT, CET, MET, and + EET, which previously used nonzero types for these timestamps. + Changes to documentation New restrictions: A Rule name must start with a character that @@ -25,6 +35,9 @@ Unreleased, experimental changes The latter restriction makes room for future extensions (a possibility noted by Tom Lane). + tzfile.5 now documents what time types apply before the first and + after the last transition, if any. + The name "TZif" is now used for the tz binary data format. tz-link.htm now mentions the A0 TimeZone Migration utilities. diff --git a/localtime.c b/localtime.c index 5b5a5b1..21160c0 100644 --- a/localtime.c +++ b/localtime.c @@ -131,7 +131,11 @@ struct state { char chars[BIGGEST(BIGGEST(TZ_MAX_CHARS + 1, sizeof gmt), (2 * (MY_TZNAME_MAX + 1)))]; struct lsinfo lsis[TZ_MAX_LEAPS]; - int defaulttype; /* for early times or if no transitions */ + + /* The time type to use for early times or if no transitions. + It is always zero for recent tzdb releases. + It might be nonzero for data from tzdb 2018e or earlier. */ + int defaulttype; }; enum r_type { @@ -657,6 +661,17 @@ tzloadbody(char const *name, struct state *sp, bool doextend, break; } } + + /* Infer sp->defaulttype from the data. Although this default + type is always zero for data from recent tzdb releases, + things are trickier for data from tzdb 2018e or earlier. + + The first set of heuristics work around bugs in 32-bit data + generated by tzdb 2013c or earlier. The workaround is for + zones like Australia/Macquarie where timestamps before the + first transition have a time type that is not the earliest + standard-time type. See: + https://mm.icann.org/pipermail/tz/2013-May/019368.html */ /* ** If type 0 is unused in transitions, ** it's the type to use for early times. @@ -678,6 +693,9 @@ tzloadbody(char const *name, struct state *sp, bool doextend, if (!sp->ttis[i].tt_isdst) break; } + /* The next heuristics are for data generated by tzdb 2018e or + earlier, for zones like EST5EDT where the first transition + is to DST. */ /* ** If no result yet, find the first standard type. ** If there is none, punt to type zero. @@ -690,7 +708,12 @@ tzloadbody(char const *name, struct state *sp, bool doextend, break; } } + /* A simple 'sp->defaulttype = 0;' would suffice here if we + didn't have to worry about 2018e-or-earlier data. Even + simpler would be to remove the defaulttype member and just + use 0 in its place. */ sp->defaulttype = i; + return 0; } @@ -1115,8 +1138,8 @@ tzparse(const char *name, struct state *sp, bool lastditch) /* ** Two transitions per year, from EPOCH_YEAR forward. */ - init_ttinfo(&sp->ttis[0], -dstoffset, true, stdlen + 1); - init_ttinfo(&sp->ttis[1], -stdoffset, false, 0); + init_ttinfo(&sp->ttis[0], -stdoffset, false, 0); + init_ttinfo(&sp->ttis[1], -dstoffset, true, stdlen + 1); sp->defaulttype = 0; timecnt = 0; janfirst = 0; @@ -1157,17 +1180,14 @@ tzparse(const char *name, struct state *sp, bool lastditch) if (! increment_overflow_time (&sp->ats[timecnt], janoffset + starttime)) - sp->types[timecnt++] = reversed; - else if (janoffset) - sp->defaulttype = reversed; + sp->types[timecnt++] = !reversed; sp->ats[timecnt] = janfirst; if (! increment_overflow_time (&sp->ats[timecnt], janoffset + endtime)) { - sp->types[timecnt++] = !reversed; + sp->types[timecnt++] = reversed; yearlim = year + YEARSPERREPEAT + 1; - } else if (janoffset) - sp->defaulttype = !reversed; + } } if (increment_overflow_time (&janfirst, janoffset + yearsecs)) @@ -1175,9 +1195,10 @@ tzparse(const char *name, struct state *sp, bool lastditch) janoffset = 0; } sp->timecnt = timecnt; - if (! timecnt) + if (! timecnt) { + sp->ttis[0] = sp->ttis[1]; sp->typecnt = 1; /* Perpetual DST. */ - else if (YEARSPERREPEAT < year - yearbeg) + } else if (YEARSPERREPEAT < year - yearbeg) sp->goback = sp->goahead = true; } else { register int_fast32_t theirstdoffset; diff --git a/tzfile.5 b/tzfile.5 index 530397f..13213d2 100644 --- a/tzfile.5 +++ b/tzfile.5 @@ -70,7 +70,9 @@ at which the rules for computing local time change. one-byte unsigned integer values; each one tells which of the different types of local time types described in the file is associated with the time period -starting with the same-indexed transition time. +starting with the same-indexed transition time +and continuing up to but not including the next transition time +(or continuing for one second, if this is the last transition). These values serve as indices into the next field. .IP * .I tzh_typecnt @@ -159,15 +161,20 @@ eight bytes are used for each transition time or leap second time. After the second header and data comes a newline-enclosed, POSIX-TZ-environment-variable-style string for use in handling instants after the last transition time stored in the file -(with nothing between the newlines if there is no POSIX representation for -such instants). -The POSIX-style string must agree with the local time type after -both data's last transition times; for example, given the string +or for all instants if the file has no transitions. +The POSIX-style TZ string is empty (i.e., nothing between the newlines) +if there is no POSIX representation for such instants. +If nonempty, the POSIX-style TZ string must agree with the local time +type after both data's last transition times if present; +for example, given the string .q "WET0WEST,M3.5.0,M10.5.0/3" then if a last transition time is in July, the transition's local time type must specify a daylight-saving time abbreviated .q "WEST" that is one hour east of UT. +Also, if there is at least one transition, time type 0 is associated +with the time period from the indefinite past up to but not including +the earliest transition time. .SS Version 3 format For version-3-format time zone files, the POSIX-TZ-style string may use two minor extensions to the POSIX TZ format, as described in diff --git a/zic.c b/zic.c index 679b3e2..ce9ccaf 100644 --- a/zic.c +++ b/zic.c @@ -1759,6 +1759,17 @@ is32(const zic_t x) return INT32_MIN <= x && x <= INT32_MAX; } +static void +swaptypes(int i, int j) +{ + { zic_t t = gmtoffs[i]; gmtoffs[i] = gmtoffs[j]; gmtoffs[j] = t; } + { char t = isdsts[i]; isdsts[i] = isdsts[j]; isdsts[j] = t; } + { unsigned char t = abbrinds[i]; abbrinds[i] = abbrinds[j]; + abbrinds[j] = t; } + { bool t = ttisstds[i]; ttisstds[i] = ttisstds[j]; ttisstds[j] = t; } + { bool t = ttisgmts[i]; ttisgmts[i] = ttisgmts[j]; ttisgmts[j] = t; } +} + static void writezone(const char *const name, const char *const string, char version) { @@ -1904,7 +1915,8 @@ writezone(const char *const name, const char *const string, char version) for (pass = 1; pass <= 2; ++pass) { register ptrdiff_t thistimei, thistimecnt, thistimelim; register int thisleapi, thisleapcnt, thisleaplim; - int writetype[TZ_MAX_TYPES]; + int old0, new0; + char omittype[TZ_MAX_TYPES + 1]; int typemap[TZ_MAX_TYPES]; register int thistypecnt; char thischars[TZ_MAX_CHARS]; @@ -1929,25 +1941,46 @@ writezone(const char *const name, const char *const string, char version) error(_("too many transition times")); thistimelim = thistimei + thistimecnt; thisleaplim = thisleapi + thisleapcnt; - for (i = 0; i < typecnt; ++i) - writetype[i] = thistimecnt == timecnt; + memset(omittype, thistimecnt != timecnt, typecnt); if (thistimecnt == 0) { /* ** No transition times fall in the current ** (32- or 64-bit) window. */ if (typecnt != 0) - writetype[typecnt - 1] = true; + omittype[typecnt - 1] = false; } else { for (i = thistimei - 1; i < thistimelim; ++i) if (i >= 0) - writetype[types[i]] = true; + omittype[types[i]] = false; /* ** For America/Godthab and Antarctica/Palmer */ if (thistimei == 0) - writetype[0] = true; + omittype[0] = false; } + + /* Reorder types to make type 0 the first-used standard type + if one is used, otherwise the first-used DST type if one + is used, otherwise no reordering. Use TYPEMAP to swap + OLD0 and NEW0 so that NEW0 appears as type 0 in the output + instead of OLD0. TYPEMAP also omits unused types. */ + omittype[typecnt] = false; /* strlen sentinel */ + old0 = strlen(omittype); + new0 = -1; + for (i = 0; i < timecnt; i++) + if (!omittype[types[i]]) { + if (! isdsts[types[i]]) { + new0 = types[i]; + break; + } + if (new0 < 0) + new0 = types[i]; + } + if (new0 < 0) + new0 = old0; + swaptypes(old0, new0); + #ifndef LEAVE_SOME_PRE_2011_SYSTEMS_IN_THE_LURCH /* ** For some pre-2011 systems: if the last-to-be-written @@ -1965,8 +1998,8 @@ writezone(const char *const name, const char *const string, char version) if (isdsts[types[i]]) mrudst = types[i]; else mrustd = types[i]; - for (i = 0; i < typecnt; ++i) - if (writetype[i]) { + for (i = old0; i < typecnt; i++) + if (!omittype[i]) { if (isdsts[i]) hidst = i; else histd = i; @@ -1980,7 +2013,7 @@ writezone(const char *const name, const char *const string, char version) ttisstds[mrudst], ttisgmts[mrudst]); isdsts[mrudst] = 1; - writetype[type] = true; + omittype[type] = false; } if (histd >= 0 && mrustd >= 0 && histd != mrustd && gmtoffs[histd] != gmtoffs[mrustd]) { @@ -1991,20 +2024,23 @@ writezone(const char *const name, const char *const string, char version) ttisstds[mrustd], ttisgmts[mrustd]); isdsts[mrustd] = 0; - writetype[type] = true; + omittype[type] = false; } } #endif /* !defined LEAVE_SOME_PRE_2011_SYSTEMS_IN_THE_LURCH */ thistypecnt = 0; - for (i = 0; i < typecnt; ++i) - typemap[i] = writetype[i] ? thistypecnt++ : -1; + for (i = old0; i < typecnt; i++) + if (!omittype[i]) + typemap[i == old0 ? new0 : i == new0 ? old0 : i] + = thistypecnt++; + for (i = 0; i < sizeof indmap / sizeof indmap[0]; ++i) indmap[i] = -1; thischarcnt = 0; - for (i = 0; i < typecnt; ++i) { + for (i = old0; i < typecnt; i++) { register char * thisabbr; - if (!writetype[i]) + if (omittype[i]) continue; if (indmap[abbrinds[i]] >= 0) continue; @@ -2053,8 +2089,8 @@ writezone(const char *const name, const char *const string, char version) uc = typemap[types[i]]; fwrite(&uc, sizeof uc, 1, fp); } - for (i = 0; i < typecnt; ++i) - if (writetype[i]) { + for (i = old0; i < typecnt; i++) + if (!omittype[i]) { puttzcode(gmtoffs[i], fp); putc(isdsts[i], fp); putc((unsigned char) indmap[abbrinds[i]], fp); @@ -2087,12 +2123,13 @@ writezone(const char *const name, const char *const string, char version) else puttzcode64(todo, fp); puttzcode(corr[i], fp); } - for (i = 0; i < typecnt; ++i) - if (writetype[i]) + for (i = old0; i < typecnt; i++) + if (!omittype[i]) putc(ttisstds[i], fp); - for (i = 0; i < typecnt; ++i) - if (writetype[i]) + for (i = old0; i < typecnt; i++) + if (!omittype[i]) putc(ttisgmts[i], fp); + swaptypes(old0, new0); } fprintf(fp, "\n%s\n", string); close_file(fp, directory, name); -- 2.17.1

1 0