- tz - lists.iana.org

Re: zic.c:mkdirs broken
by ado Feb. 10, 1994

Feb. 10, 1994

> The fix I sent before for this never made it into 94b. I don't know how > this fails to break everyone else. mkdirs is broken unless you compile > with -fwritable-strings (which is generally losing to do). > ... Thanks for the fix; it (and only it) is incorporated into tzcode94c.tar.gz on elsie.nci.nih.gov. The relevant diff is attached below. --ado SCCS/s.zic.c: 7.21 vs. 7.22 *** 7.21/zic.c Thu Feb 10 08:54:03 1994 --- 7.22/zic.c Thu Feb 10 08:54:04 1994 *************** *** 1,6 **** #ifndef lint #ifndef NOID ! static char elsieid[] = "@(#)zic.c 7.21"; #endif /* !defined NOID */ #endif /* !defined lint */ --- 1,6 ---- #ifndef lint #ifndef NOID ! static char elsieid[] = "@(#)zic.c 7.22"; #endif /* !defined NOID */ #endif /* !defined lint */ *************** *** 1878,1890 **** } static int ! mkdirs(name) ! char * const name; { register char * cp; ! if ((cp = name) == NULL || *cp == '\0') return 0; while ((cp = strchr(cp + 1, '/')) != 0) { *cp = '\0'; #ifndef unix --- 1878,1892 ---- } static int ! mkdirs(argname) ! char * const argname; { + register char * name; register char * cp; ! if (argname == NULL || *argname == '\0') return 0; + cp = name = ecpyalloc(argname); while ((cp = strchr(cp + 1, '/')) != 0) { *cp = '\0'; #ifndef unix *************** *** 1906,1916 **** --- 1908,1920 ---- "%s: Can't create directory ", progname); (void) perror(name); + ifree(name); return -1; } } *cp = '/'; } + ifree(name); return 0; }

1 0

Re: zic.c:mkdirs broken
by Bradley White Feb. 9, 1994

Feb. 9, 1994

+ /* We must make a copy in case the + passed pointer is a read-only string. */ + char *name = ecpyalloc (argname); + if ((cp = name) == NULL || *cp == '\0') { + free (name); + return 0; + } ecpyalloc(x) [a.k.a. memcheck(icpyalloc(x))] cannot return 0. But, if you just want to program defensively, it is also good practice not to pass 0 to free(), despite what any standards might say. With the ialloc package you can achieve this via ifree(). Bradley

1 0

zic.c:mkdirs broken
by roland＠gnu.ai.mit.edu Feb. 9, 1994

Feb. 9, 1994

The fix I sent before for this never made it into 94b. I don't know how this fails to break everyone else. mkdirs is broken unless you compile with -fwritable-strings (which is generally losing to do). diff -u /home/gd3/roland/libc/time/zic.c.\~4\~ /home/gd3/roland/libc/time/zic.c --- /home/gd3/roland/libc/time/zic.c.~4~ Mon Feb 7 14:30:53 1994 +++ /home/gd3/roland/libc/time/zic.c Wed Feb 9 14:09:11 1994 @@ -1877,14 +1877,20 @@ charcnt += eitol(i); } + static int -mkdirs(name) -char * const name; +mkdirs(argname) +char * const argname; { register char * cp; + /* We must make a copy in case the + passed pointer is a read-only string. */ + char *name = ecpyalloc (argname); - if ((cp = name) == NULL || *cp == '\0') - return 0; + if ((cp = name) == NULL || *cp == '\0') { + free (name); + return 0; + } while ((cp = strchr(cp + 1, '/')) != 0) { *cp = '\0'; #ifndef unix @@ -1906,11 +1912,13 @@ "%s: Can't create directory ", progname); (void) perror(name); + free (name); return -1; } } *cp = '/'; } + free (name); return 0; } Diff exited abnormally with code 1 at Wed Feb 9 14:09:45

1 0

leap seconds [forwarded with permission from wollman@uvm-gen.emba.uvm.edu]
by ado Feb. 7, 1994

Feb. 7, 1994

> From wollman(a)uvm-gen.EMBA.UVM.EDU Sun Feb 6 17:45:52 1994 > Return-Path: <wollman(a)uvm-gen.EMBA.UVM.EDU> > Received: from bajoran.emba.uvm.edu by elsie.nci.nih.gov (4.1/SMI-4.1) > id AA18437; Sun, 6 Feb 94 17:45:51 EST > Received: by bajoran.emba.uvm.edu id AA18471 > (5.65/6.05); Sun, 6 Feb 94 17:45:50 -0500 > Date: Sun, 6 Feb 94 17:45:50 -0500 > From: wollman(a)uvm-gen.EMBA.UVM.EDU > Message-Id: <9402062245.AA18471(a)bajoran.emba.uvm.edu> > To: ado(a)elsie.nci.nih.gov > Subject: Re: timezone servers > Newsgroups: comp.protocols.time.ntp > In-Reply-To: <1994Feb4.234532.26832(a)alw.nih.gov> > References: <199402020109.AA01671(a)ni.umd.edu> > Organization: University of Vermont, EMBA Computer Facility > > ...According to the bulletins that I've seen, |TAI-UTC| is about > 28 seconds now. However, the leapseconds file only lists 19 steps. > This is probably due to the fact that the first step was 10 seconds, > not just 1. Is there an easy way to fix this, so my users who use NTP > don't complain about their clocks being nine seconds off? > > -GAWollman > -- > Garrett A. Wollman | Shashish is simple, it's discreet, it's brief. ... > wollman(a)emba.uvm.edu | Shashish is the bonding of hearts in spite of distance. > NB: Address soon to | It is a bond more powerful than absence. We like people > be changed. | who like Shashish. - Claude McKenzie + Florent Vollant

4 3

tzcode94b.tar.gz and tzdata94b.tar.gz
by ado Feb. 7, 1994

Feb. 7, 1994

Updated time zone code and data is available by anonymous ftp from elsie.nci.nih.gov (128.231.16.1) in the "pub" directory; the culprits are -rw-r--r-- 1 ado 62145 Feb 7 10:04 tzcode94b.tar.gz -rw-r--r-- 1 ado 63993 Feb 7 10:04 tzdata94b.tar.gz The updates are the work by Paul Eggert (eggert(a)twinsun.com) who notes: > I found another book of time zone histories by E W Whitman; it's not as > extensive as Shanks but has a few goodies of its own. I used it to > update the tables. I also fixed some more as a result of > correspondence with Adam David and Peter Ilieve, and move some stray > links from `europe' to `backward'. I corrected some scanning errors in > usno1989. > > As far as the code goes, I fixed zic to allow years in the range > INT_MIN to INT_MAX; this fixed a few boundary conditions around 1900. > And I cleaned up the zic documentation a little bit. --ado

1 0

tzdata94a.tar.gz
by ado Feb. 3, 1994

Feb. 3, 1994

tzdata94a.tar.gz is now available on elsie.nci.nih.gov; it simply incorporates the recently announced leap second into the "leapseconds" file. For those who want to forego a trip to the ftp well, a diff from the previous version is attached. --ado SCCS/s.leapseconds: 7.6 vs. 7.7 *** 7.6/leapseconds Thu Feb 3 09:02:22 1994 --- 7.7/leapseconds Thu Feb 3 09:02:22 1994 *************** *** 1,4 **** ! # @(#)leapseconds 7.6 # Allowance for leapseconds added to each timezone file. --- 1,4 ---- ! # @(#)leapseconds 7.7 # Allowance for leapseconds added to each timezone file. *************** *** 38,40 **** --- 38,41 ---- Leap 1990 Dec 31 23:59:60 + S Leap 1992 Jun 30 23:59:60 + S Leap 1993 Jun 30 23:59:60 + S + Leap 1994 Jun 30 23:59:60 + S

1 0

time zones [forwarded with permission]
by ado Jan. 5, 1994

Jan. 5, 1994

> From ncm(a)netcom.com Wed Jan 5 03:20:09 1994 > Return-Path: <ncm(a)netcom.com> > Received: from mail.netcom.com (netcom3.netcom.com) by elsie.nci.nih.gov (4.1/SMI-4.1) > id AA17468; Wed, 5 Jan 94 03:20:07 EST > Received: from localhost by mail.netcom.com (8.6.4/SMI-4.1/Netcom) > id AAA02279; Wed, 5 Jan 1994 00:20:40 -0800 > Date: Wed, 5 Jan 1994 00:20:40 -0800 > From: ncm(a)netcom.com (Nathan Myers) > Message-Id: <199401050820.AAA02279(a)mail.netcom.com> > To: ado(a)elsie.nci.nih.gov > Subject: time zones > Status: RO > > Hello, > > I have been working on time zones at Rogue Wave Software, and > on locale support for the ISO/ANSI C++ Standard Library committee. > It seems to me that the C++ Library effort offers an excellent > place to standardize on a time zone library, as the contents > of a class timezone. We are not bound to C Library facilities > or conventions, so this is an opportunity to correct old mistakes. > (e.g. hidden static variables implied by the interface) > > Please forward this to anyone you think might be interested in > helping put together a proposal. > > Thank you > Nathan Myersn > myersn(a)roguewave.com

1 0

Re: Iceland's timezone
by Adam David Dec. 9, 1993

Dec. 9, 1993

> # time the norsemen first settled Iceland. The first day of winter is always > # Saturday, but is not dependent on the Julian or Gregorian calendars. I will try and find a general rule for this, but it may take awhile. > # We go with the Almanak, except for two claims from Shanks: > # 1. In 1968 standard time wasn't switched to GMT until Oct 27; > # this affects only the time zone name and dst flag. I wonder if this one is such a good idea. Here is the actual quote from the almanak (paraphrased): In the year 1968 "summer time" was finally made standard time in Iceland by law number 6 from April 5th that year. The change was made on April 7th at 2:00 GMT and the clock was adjusted forwards by one hour. Clock time in Iceland has since then been on GMT the whole year round. Of course, both Shanks and the Almanak have the same actual time offset for summer 1968, but technically it is the Almanak that is correct. Shanks chooses October 27th as the transition time because that is when the end of the summer time would otherwise have been (by the previous general rule). -- Adam David <adam(a)veda.is>

1 0

Iceland's timezone
by yata!eggert＠twinsun.com Dec. 9, 1993

Dec. 9, 1993

Thanks for looking up the U of Iceland Almanak. Here's how I translated what you wrote into tzfile format; I'll include this in the next set of revisions that I propose to ado. # Iceland # # From Adam David <adam(a)veda.is> (November 6, 1993): # The name of the timezone in Iceland for system / mail / news purposes is GMT. # # (December 5, 1993): # This material is paraphrased from the 1988 edition of the University of # Iceland Almanak. # # From January 1st, 1908 the whole of Iceland was standardised at 1 hour # behind GMT. Previously, local mean solar time was used in different parts # of Iceland, the almanak had been based on Reykjavik mean solar time which # was 1 hour and 28 minutes behind GMT. # # "first day of winter" referred to [below] means the first day of the 26 weeks # of winter, according to the old icelandic calendar that dates back to the # time the norsemen first settled Iceland. The first day of winter is always # Saturday, but is not dependent on the Julian or Gregorian calendars. # # From Paul Eggert <eggert(a)twinsun.com> (December 9, 1993): # The Iceland Almanak, Shanks and Whitman disagree on many points. # We go with the Almanak, except for two claims from Shanks: # 1. In 1968 standard time wasn't switched to GMT until Oct 27; # this affects only the time zone name and dst flag. # 2. Reykavik was -1:28 from 1837 to 1908, local mean time before that. # # Rule NAME FROM TO TYPE IN ON AT SAVE LETTER/S Rule Iceland 1908 only - Jan 1 0:00 0 S Rule Iceland 1917 1918 - Feb 19 23:00 1:00 D Rule Iceland 1917 only - Oct 21 1:00 0 S Rule Iceland 1918 only - Nov 16 1:00 0 S Rule Iceland 1939 only - Apr 29 23:00 1:00 D Rule Iceland 1939 only - Nov 29 2:00 0 S Rule Iceland 1940 only - Feb 25 2:00 1:00 D Rule Iceland 1940 only - Nov 3 2:00 0 S Rule Iceland 1941 only - Mar 2 1:00s 1:00 D Rule Iceland 1941 only - Nov 2 1:00s 0 S Rule Iceland 1942 only - Mar 8 1:00s 1:00 D Rule Iceland 1942 only - Oct 25 1:00s 0 S # 1943-1946 - first Sunday in March until first Sunday in winter Rule Iceland 1943 1946 - Mar Sun>=1 1:00s 1:00 D Rule Iceland 1943 1948 - Oct Sun>=22 1:00s 0 S # 1947-1967 - first Sunday in April until first Sunday in winter Rule Iceland 1947 1968 - Apr Sun>=1 1:00s 1:00 D # 1949 Oct transition delayed by 1 week Rule Iceland 1949 only - Oct 30 1:00s 0 S Rule Iceland 1950 1966 - Oct Sun>=22 1:00s 0 S Rule Iceland 1967 only - Oct 29 1:00s 0 S # Zone NAME GMTOFF RULES FORMAT [UNTIL] Zone Atlantic/Reykjavik -1:27:24 - LMT 1837 -1:28 - RMT 1908 # Reykjavik Mean Time -1:00 Iceland I%sT 1968 Oct 27 1:00s 0:00 - GMT

1 0

Re: Iceland's timezone
by Adam David Dec. 6, 1993

Dec. 6, 1993

Sorry about the delay getting back with this information. > I'm working on a rewrite of `europe' with a large set of historical > data. My current draft (see below) claims that Iceland changed its > timezone in 1968. I've taken this claim from Thomas G. Shanks, The > international atlas (3rd edition), San Diego: ACS Publications, Inc. > (1991). If you have some better data, please let me know. In a way, there was a timezone change in 1968, where the summertime change was applied during the spring, and became the all-year-round time, and there have been no changes since then, not for DST or any other reason. When looking back to previous years when there was DST, icelanders tend to refer to the concept of "winter time", as the summer time had the same offset as the present all-year-round time. This is why I was led to believe that no timezone change had taken place in 1968. > It would be nice to know what time zone abbreviations were used before > 1968, since Shanks doesn't say. In the following table I've made up > the names `IST', `IDT', `RMT', and `LMT' for Iceland Standard Time, > Iceland Daylight Time, Reykjavik Mean Time, and Local Mean Time > respectively. I have no details of any abbreviations, sorry. > # From Shanks (1991): > # Rule NAME FROM TO TYPE IN ON AT SAVE LETTER/S > Rule Iceland 1908 only - Jan 1 0:00 0 S > Rule Iceland 1917 1918 - Feb 20 2:00 1:00 D > Rule Iceland 1917 only - Oct 25 2:00 0 S > Rule Iceland 1918 1919 - Nov 15 2:00 0 S > Rule Iceland 1919 only - Feb 19 2:00 1:00 D > Rule Iceland 1921 only - Mar 19 2:00 1:00 D > Rule Iceland 1921 only - Jun 22 2:00 0 S > Rule Iceland 1941 1942 - Mar Sat>=1 2:00 1:00 D > Rule Iceland 1941 1942 - Jul 2 2:00 0 S > Rule Iceland 1943 1946 - Mar Sun>=1 2:00 1:00 D > Rule Iceland 1943 1948 - Oct Sun>=22 2:00 0 S > Rule Iceland 1947 1962 - Apr Sun>=1 2:00 1:00 D > Rule Iceland 1949 only - Oct 30 2:00 0 S > Rule Iceland 1950 only - Oct 22 2:00 0 S > # From Paul Eggert <eggert(a)twinsun.com> (November 10, 1993): > # Shanks's ``1953 Oct 26'' is probably a typo for Oct 25. > # Likewise, correct Shanks's ``1961 Oct 22'' to Oct 29. > Rule Iceland 1951 1967 - Oct Sun>=23 1:00s 0 S > Rule Iceland 1963 1968 - Apr Sun>=1 1:00s 1:00 D > # Zone NAME GMTOFF RULES/SAVE FORMAT [UNTIL] > Zone Atlantic/Iceland/Reykjavik -1:27:24 - LMT 1837 Jan 1 > -1:28 - RMT 1908 Jan 1 > -1:00 Iceland I%sT 1968 Oct 27 1:00s > 0:00 - GMT > # This link is for backward compatibility with the old naming convention. > Link Atlantic/Iceland/Reykjavik Iceland I don't feel up to messing with this table and submitting a corrected version. Here I copy a list of periods when summertime was active, and explanatory text. This material is paraphrased from the 1988 edition of the University of Iceland Almanak. >From January 1st, 1908 the whole of Iceland was standardised at 1 hour behind GMT. Previously, local mean solar time was used in different parts of Iceland, the almanak had been based on Reykjavik mean solar time which was 1 hour and 28 minutes behind GMT. Summertime when it was applied was always by putting the clock an hour forward during the summer months, which made the clock time equal to GMT. The following list shows when this summertime was in force. The time of the clock change is specified in GMT. year from to ---- ---- -- 1908-1916 no summertime 1917 20.Feb. 0:00 21.Oct. 1:00 1918 20.Feb. 0:00 16.Nov. 1:00 1919-1938 no summertime 1939 30.Apr. 0:00 29.Nov. 2:00 1940 25.Feb. 3:00 3.Nov. 2:00 1941 2.Mar. 2:00 2.Nov. 2:00 1942 8.Mar. 2:00 25.Oct. 2:00 1943-1946 rule (first Sunday in March until first Sunday in winter) 1943 7.Mar. 2:00 24.Oct. 2:00 1944 5.Mar. 2:00 22.Oct. 2:00 1945 4.Mar. 2:00 28.Oct. 2:00 1946 3.Mar. 2:00 27.Oct. 2:00 1947-1967 rule (first Sunday in April until first Sunday in winter) 1947 6.Apr. 2:00 26.Oct. 2:00 1948 4.Apr. 2:00 24.Oct. 2:00 1949 3.Apr. 2:00 ******* 30.Oct. 2:00 (delayed by 1 week) 1950 2.Apr. 2:00 22.Oct. 2:00 1951 1.Apr. 2:00 28.Oct. 2:00 1952 6.Apr. 2:00 26.Oct. 2:00 1953 5.Apr. 2:00 25.Oct. 2:00 1954 4.Apr. 2:00 24.Oct. 2:00 1955 3.Apr. 2:00 23.Oct. 2:00 1956 1.Apr. 2:00 28.Oct. 2:00 1957 7.Apr. 2:00 27.Oct. 2:00 1958 6.Apr. 2:00 26.Oct. 2:00 1959 5.Apr. 2:00 25.Oct. 2:00 1960 3.Apr. 2:00 23.Oct. 2:00 1961 2.Apr. 2:00 22.Oct. 2:00 1962 1.Apr. 2:00 28.Oct. 2:00 1963 7.Apr. 2:00 27.Oct. 2:00 1964 5.Apr. 2:00 25.Oct. 2:00 1965 4.Apr. 2:00 24.Oct. 2:00 1966 3.Apr. 2:00 23.Oct. 2:00 1967 2.Apr. 2:00 29.Oct. 2:00 1968 7.Apr. 2:00 (actually GMT after this) "first day of winter" referred to above, means the first day of the 26 weeks of winter, according to the old icelandic calendar that dates back to the time the norsemen first settled Iceland. The first day of winter is always Saturday, but is not dependent on the Julian or Gregorian calendars. This is the reason for the discrepancy with October 1961. I can find out more exactly how to determine when it falls, if necessary. It is a pity I don't have that information handy, otherwise I could include it here. There are more serious discrepancies before 1941 which are not obviously explainable. What to do... (?) -- Adam David <adam(a)veda.is>

1 0