New Apple time zone data for Canada?
Gary Ng reported[1] that on Tuesday he got a notification on his Apple TV that updated time zone information was available, and that this was because Apple pushed out an over-the-air time zone update for Canadian users due to recent provincial legislation in BC and Alberta. I'm skeptical at least about the "Alberta" part, since the Alberta changes are not yet official. Does anyone have definitive information about this? The only public repository I found for Apple[2] is currently based on TZDB 2025b, which is so old that it doesn't have the official changes for British Columbia, much less the not-yet-official changes for Alberta. I would guess the updates Ng reported are newer than 2025b. Another possibility is that Apple's changes for Canada are about interfacing to cell phone service providers, which are presumably having their own issues with time zones. [1]: Ng G. Apple Sends Overnight Time Zone Fix to Canadian iPhones. Here’s Why. iPhone in Canada. 2026-05-27. <https://www.iphoneincanada.ca/2026/05/27/apple-iphone-time-zone-update-canad...> [2]: https://github.com/apple-oss-distributions/TimeZoneData/commits/main/
On Sun, 31 May 2026 at 04:16, Paul Eggert via tz <tz@iana.org> wrote:
Gary Ng reported[1] that on Tuesday he got a notification on his Apple TV that updated time zone information was available, and that this was because Apple pushed out an over-the-air time zone update for Canadian users due to recent provincial legislation in BC and Alberta.
I'm skeptical at least about the "Alberta" part, since the Alberta changes are not yet official. Does anyone have definitive information about this?
I just checked my machine running macOS 26.5 Tahoe, which seems to have received tz 2026b as an update early Wednesday, though I have not yet done the "manual restart" to apply it. As such, I can compare them directly and confirm that, for Canada, this only contains the BC changes we've published: /private/var/db/timezone/tz % diff -r {2025c.1.0,2026b.1.0}/zoneinfo/America Binary files 2025c.1.0/zoneinfo/America/Vancouver and 2026b.1.0/zoneinfo/America/Vancouver differ Ng's comment implying that the update includes or is prompted by Alberta's more recent actions was an incorrect (but well-intentioned) assumption or simplification, given it took ~5 weeks for Apple to propagate BC's changes. -- Tim Parenti
Tim Parenti via tz <tz@iana.org> writes:
I just checked my machine running macOS 26.5 Tahoe, which seems to have received tz 2026b as an update early Wednesday, though I have not yet done the "manual restart" to apply it. As such, I can compare them directly and confirm that, for Canada, this only contains the BC changes we've published:
I see the same. Interestingly, while the directories in the zoneinfo tree are all dated May 27, the individual files are all dated May 11: $ ls -l 2026b.1.0/zoneinfo/ total 656 -rw-r--r-- 1 root wheel 6 May 11 19:10 +VERSION drwxr-xr-x 56 root wheel 1792 May 27 04:25 Africa/ drwxr-xr-x 149 root wheel 4768 May 27 04:25 America/ drwxr-xr-x 14 root wheel 448 May 27 04:25 Antarctica/ drwxr-xr-x 3 root wheel 96 May 27 04:25 Arctic/ drwxr-xr-x 101 root wheel 3232 May 27 04:25 Asia/ drwxr-xr-x 14 root wheel 448 May 27 04:25 Atlantic/ drwxr-xr-x 25 root wheel 800 May 27 04:25 Australia/ drwxr-xr-x 6 root wheel 192 May 27 04:25 Brazil/ -rw-r--r-- 1 root wheel 2933 May 11 19:10 CET -rw-r--r-- 1 root wheel 3592 May 11 19:10 CST6CDT drwxr-xr-x 10 root wheel 320 May 27 04:25 Canada/ drwxr-xr-x 4 root wheel 128 May 27 04:25 Chile/ -rw-r--r-- 1 root wheel 2416 May 11 19:10 Cuba -rw-r--r-- 1 root wheel 2262 May 11 19:10 EET -rw-r--r-- 1 root wheel 182 May 11 19:10 EST -rw-r--r-- 1 root wheel 3552 May 11 19:10 EST5EDT ... Evidently the package was built May 11 but not pushed out till last week. I wonder why... regards, tom lane
On Sun, 31 May 2026 at 11:47, Tom Lane <tgl@sss.pgh.pa.us> wrote:
I see the same. Interestingly, while the directories in the zoneinfo tree are all dated May 27, the individual files are all dated May 11:
Yes, I checked with a friend and the file timestamps appear to be consistent across devices, while the directory timestamps seem to be whenever your machine received the update and created those directories. Evidently the package was built May 11 but not pushed out till
last week. I wonder why...
The previous version Apple pushed was 2025c, as it seems to have skipped 2026a's minor changes for Moldova. We released it on 10 December 2025, it appears to have been built by Apple on 30 January 2026, and my machine received it on 18 February 2026. So this does speak to the general magnitude of the lags inherent in downstream vendors testing and preparing updated tzdata for widespread release. FWIW, based on my testing, Google Calendar still hasn't seemed to pick up on the BC change yet, nor has my Samsung phone. -- Tim Parenti
On 2026-05-31 09:51, Tim Parenti wrote:
FWIW, based on my testing, Google Calendar still hasn't seemed to pick up on the BC change yet, nor has my Samsung phone.
Yes, and many other platforms are also behind. For example, although Fedora 44 tzdata was updated to 2026b last week, Ubuntu 26.04 tzdata is still based on 2026a. An amusing example is Google's ChromeOS which has two copies of tzdata, one updated and the other not. The current ChromeOS version 148.0.7778.214 still has an old America/Vancouver (tzdata.zi says "version 2024a-dirty"), whereas its supported GNU/Linux guest VM Crostini, based on Debian GNU/Linux 12 (bookworm), has an updated America/Vancouver (tzdata.zi reports "version 2026b"). I expect native ChromeOS's long lag time is partly because it gets updates from Gentoo only fitfully. And as ChromeOS is gradually being phased out in favor of Android, ChromeOS's /usr/share/zoneinfo is to some extent vestigial. Even with its longer lag time, though, I'd be surprised if ChromeOS doesn't update its America/Vancouver data before November rolls around.
Ng's comment implying that the update includes or is prompted by Alberta's more recent actions was an incorrect (but well-intentioned) assumption or simplification, given it took ~5 weeks for Apple to propagate BC's changes.
The Alberta changes haven’t been released yet (with good reason), which is the main reason it hasn’t been included in an Apple release. Apple’s OTA time zone updates also contain updated information for ICU, not just the zoneinfo data. The ICU team took some time to decide how to handle an issue raised by the BC change; Apple’s release had to wait for that discussion to conclude. It also took a few days to adjust to a change ICU made in the build process as a result of that discussion. Typically we try to turn releases around in less time than 2026b took. Debbie
On May 31, 2026, at 1:51 AM, Tim Parenti via tz <tz@iana.org> wrote:
On Sun, 31 May 2026 at 04:16, Paul Eggert via tz <tz@iana.org <mailto:tz@iana.org>> wrote:
Gary Ng reported[1] that on Tuesday he got a notification on his Apple TV that updated time zone information was available, and that this was because Apple pushed out an over-the-air time zone update for Canadian users due to recent provincial legislation in BC and Alberta.
I'm skeptical at least about the "Alberta" part, since the Alberta changes are not yet official. Does anyone have definitive information about this?
I just checked my machine running macOS 26.5 Tahoe, which seems to have received tz 2026b as an update early Wednesday, though I have not yet done the "manual restart" to apply it. As such, I can compare them directly and confirm that, for Canada, this only contains the BC changes we've published:
/private/var/db/timezone/tz % diff -r {2025c.1.0,2026b.1.0}/zoneinfo/America Binary files 2025c.1.0/zoneinfo/America/Vancouver and 2026b.1.0/zoneinfo/America/Vancouver differ
Ng's comment implying that the update includes or is prompted by Alberta's more recent actions was an incorrect (but well-intentioned) assumption or simplification, given it took ~5 weeks for Apple to propagate BC's changes.
-- Tim Parenti
On Sun, 31 May 2026 at 18:18, Debbie Goldsmith <goldsmit@apple.com> wrote:
The ICU team took some time to decide how to handle an issue raised by the BC change; Apple’s release had to wait for that discussion to conclude. It also took a few days to adjust to a change ICU made in the build process as a result of that discussion. Typically we try to turn releases around in less time than 2026b took.
Let no one think my comment about the ~5-week timeline was passing any form of judgement on Apple! These things genuinely do take time! Besides… it seems, even with the longer-than-usual timeframe for this one, you've still ended up quicker than most. -- Tim Parenti
participants (4)
-
Debbie Goldsmith -
Paul Eggert -
Tim Parenti -
Tom Lane