January 2009 - tz - lists.iana.org

Meaning of 'goforward' flag ?
by John Dlugosz Jan. 13, 2009

Jan. 13, 2009

I think I understand how the ats array is extrapolated to dates beyond its end. The Gregorian calendar repeats every 400 years, so any rules based on weekdays and day of month numbers will have results that repeat. So, if the date is beyond the end of the array, subtract the smallest multiple of 400 years that produces a date that is within the array, and use the ttis entry that points to. This assumes that the last set of rules in force were used to generate at least 400 years of entries in ats. Is that correct? But, the code to set goforward based on the ats array decides TRUE if there exists an entry that is equivalent to the last entry but exactly 400 years earlier. That does not guarantee that the last 400 years was meant to be repeated. I'm guessing that zic fills out a whole cycle of the last set of rules in force and then one more ats entry to serve as a cue for tzload, or that the POSIX string could generate more than one complete cycle (i.e. MAX_TIMES was large enough). But simply setting a flag in the header, or noting that the POSIX string existed and could fill out 400 years, would be clearer and avoid hidden meanings. As it is, zic would have to be careful to avoid letting that cue occur by happenstance. If you don't worry about that happening, why worry about 'goback', which never occurs in the actual data? Any newly minted time zone would inherit the past from the region is split off from, so it never will. --John

1 0

code review issue re t outside of ats array's range
by John Dlugosz Jan. 13, 2009

Jan. 13, 2009

In 'localsub', If t is before the first entry in ats, and goback is not set, then it uses the first non-DST entry in ats. But, if t is after the last entry in ats and goforward is not net, then it uses the last entry in ats, regardless of whether it is a DST or standard time entry. Is that intentional? ... thinking ... time passes ... I think it's an error either way - if the ats table ends with a seasonal time change and it isn't flagged as goforward, it should be an error that the table was declared too small. It stays in DST forever and if it's flagged as DST that means it was supposed to be a temporary adjustment to the Zone definition. --John

1 0

Timezone code question
by John Dlugosz Jan. 13, 2009

Jan. 13, 2009

The zic program goes to the trouble of generating a POSIX string for "the future" that is the last thing in the data file. The library code turns that into two more ttis records. Why not just store two more ttis records in the file in the first place? What does 'ttis' stand for? What about 'ats'?

1 0

Leap Second Data
by John Dlugosz Jan. 13, 2009

Jan. 13, 2009

Why is the leap second data associated with each data file? Shouldn't it be the same everywhere? --John

1 0

FW: time zone library
by Olson, Arthur David (NIH/NCI) [E] Jan. 13, 2009

Jan. 13, 2009

I'm forwarding this message from John Dlugosz, who is not on the time zone mailing list. Those of you who are on the list, please direct replies appropriately. --ado -----Original Message----- From: John Dlugosz [mailto:JDlugosz@TradeStation.com] Sent: Tuesday, January 06, 2009 7:10 To: tz(a)lecserver.nci.nih.gov Subject: time zone library I downloaded the timezone library, and in adapting it to my needs I've been figuring out how it works. I have a few questions, and also (hopefully) a contribution. First, my contribution. Regarding /* ** Adapted from code provided by Robert Elz, who writes: ** The "best" way to do mktime I think is based on an idea of Bob ** Kridle's (so its said...) from a long time ago. ** It does a binary search of the time_t space. Since time_t's are ** just 32 bits, its a max of 32 iterations (even at 64 bits it ** would still be very reasonable). */ I'm using 64-bit time_t values, and find that it usually takes 62 to 64 iterations. Basically, it is always worst case. I tried using a rough guess for the initial hi and low: lo = ((yourtm.tm_year-70) * 365i64 + yourtm.tm_mon * 28) * SecondsPerDay; hi = ((yourtm.tm_year-70) * 366i64 + (yourtm.tm_mon+1)*31) * SecondsPerDay; when the resulting time_t would be positive, and it completes in 20 to 23 iterations. I punted on the negative case, but a very similar calculation would work. When tm_year < 70, I just used the smallest value I care about for lo, and 0 for hi. That gives me worst case 35 iterations, because I'm only supporting values back to 1752 CE, not the full range of a 64-bit number, which would be 21 times longer than the age of the universe. Now for my first question. TZ_MAX_TIMES is 1200, and that is exactly the length of the arrays for state::ats and state::types. I suppose it is noting every time change, e.g. DST adjustments twice a year for 600 years? There is no need to store anything before the first reported time in the configuration files, as it simply uses the first entry for anything older (no DST in pre-history); it seems to have code for dealing with dates beyond the end of the table. Does it extrapolate the current DST rule in the future? I'd like to cut off the table at a more reasonable date for my application. Where is that table populated? Is it when the data file is loaded, or in zic when it is created? My second question concerns leap seconds. It appears that the leap second data is associated with every zone file. Why isn't it simply the same for everything, and global? --John

4 5

Australian timezone abbreviations
by Ted Percival Jan. 12, 2009

Jan. 12, 2009

Paul Eggert wrote: > I'd like to hear more from Australian correspondents on this before > thinking about specific changes, though. +1 me for changing to AEST, AEDT and so on. I encourage you to read through a relatively recent discussion about this. I straw-polled our Unix users' group's mailing list: http://lists.humbug.org.au/archives/general/2008-April/thread.html#26929 . If you'll excuse the deviation to a discussion of whether various places should have daylight saving time, you'll see a preference for AEST: I count 6 for and 1 possibly against. Eric Ulevik wrote: > ABC NewsRadio have been referring to 'Eastern Standard Time' and > 'Eastern Summer Time'. "Australian" is presumed for an Australian radio broadcast; for computer use there is no regional context. At any rate, the discussion should focus on the *abbreviations* rather than the full names because only the abbreviations are specified by the timezone database. I'm not convinced that searching the web for "EST" is a valid survey, particularly as it is a common abbreviation for "estimate" or "established". There are also websites such as Flickr that use EST but there is evidence that this is simply a result of using the timezone database: a feedback loop. I wonder to what extent the timezone database is seen as an authority for what people use. -- Ted Percival http://tedp.id.au/tzdata-au/

1 0

"Military" time zones
by Paul Eggert Jan. 10, 2009

Jan. 10, 2009

msb(a)vex.net (Mark Brader; please reply accordingly) writes: > As you would know, there is a system of "military" time zone > designations where zones are identified by a single letter: > http://www.timeanddate.com/library/abbreviations/timezones/military/ > gives letter equivalents for zones from -12 = Y to +12 = M. > > I was having an email conversation with Russ Rowlett, who maintains > the units-of-measurement site at <http://www.unc.edu/~rowlett/units/>, > and we were wondering whether these single-letter designations are > actually still used in practice -- and, if so, how time zones above > +12 or not on integer hours are represented. Is there an official > reference for current practice on this anywhere? Yes, please see the Combined Communications-Electronics Board general communication instructions, dated 2007 <http://www.jcs.mil/j6/cceb/acps/acp121/ACP121H.pdf>. These are authoritative for the armed forces of Australia, Canada, NZ, the UK, and the US. Briefly: "N" is also used for UTC+13, "to provide for a ship in zone -12 keeping Daylight Saving Time". There's no official provision for UTC+14, but I presume that'd be "O". For DST, use the corresponding next letter (e.g., the eastern-US "R" becomes "Q" during DST.) As for non-integer hours, the instructions simply say: "For time midway between zones, both letters are used."

2 1

Time Zone y Chihuahua Mexico
by Marco Antonio Bravo Grajeda Jan. 8, 2009

Jan. 8, 2009

Chihuahua has a different time zone than the center of Mexico, please correct this.

1 0

timezone code - re-entrancy issue
by John Dlugosz Jan. 7, 2009

Jan. 7, 2009

At the bottom of localsub, tzname[tmp->tm_isdst] = &sp->chars[ttisp->tt_abbrind]; I think this is to update the tzname string to the last time zone abbr. actually used in to produce the tm values. It is not re-entrant. This is called even by localtime_r, which may be changing the values even as another thread (who called plain localtime) is reading it. It also changes all the time as the "binary search" is being done by mktime. That violates the "behaves as if never called" rule, and putting it in a subroutine doesn't help there. Am I missing something? --John

1 0

time zone library
by John Dlugosz Jan. 7, 2009

Jan. 7, 2009

I downloaded the timezone library, and in adapting it to my needs I've been figuring out how it works. I have a few questions, and also (hopefully) a contribution. First, my contribution. Regarding /* ** Adapted from code provided by Robert Elz, who writes: ** The "best" way to do mktime I think is based on an idea of Bob ** Kridle's (so its said...) from a long time ago. ** It does a binary search of the time_t space. Since time_t's are ** just 32 bits, its a max of 32 iterations (even at 64 bits it ** would still be very reasonable). */ I'm using 64-bit time_t values, and find that it usually takes 62 to 64 iterations. Basically, it is always worst case. I tried using a rough guess for the initial hi and low: lo = ((yourtm.tm_year-70) * 365i64 + yourtm.tm_mon * 28) * SecondsPerDay; hi = ((yourtm.tm_year-70) * 366i64 + (yourtm.tm_mon+1)*31) * SecondsPerDay; when the resulting time_t would be positive, and it completes in 20 to 23 iterations. I punted on the negative case, but a very similar calculation would work. When tm_year < 70, I just used the smallest value I care about for lo, and 0 for hi. That gives me worst case 35 iterations, because I'm only supporting values back to 1752 CE, not the full range of a 64-bit number, which would be 21 times longer than the age of the universe. Now for my first question. TZ_MAX_TIMES is 1200, and that is exactly the length of the arrays for state::ats and state::types. I suppose it is noting every time change, e.g. DST adjustments twice a year for 600 years? There is no need to store anything before the first reported time in the configuration files, as it simply uses the first entry for anything older (no DST in pre-history); it seems to have code for dealing with dates beyond the end of the table. Does it extrapolate the current DST rule in the future? I'd like to cut off the table at a more reasonable date for my application. Where is that table populated? Is it when the data file is loaded, or in zic when it is created? My second question concerns leap seconds. It appears that the leap second data is associated with every zone file. Why isn't it simply the same for everything, and global? --John

1 0