2

Currently I am writing a component for an ESP32 microcontroller. Therefore I need a list of all timezones.

I found many sources of a timezone file, but they use different formats for the timezone values. One example ist Europe/Ulyanovsk. Some files use <+04>-4 and others use UNK-4.

The ESP32 can handle UNK-4 correctly, but fails with a value like <+04>-4.

Can someone explain me, what the format with "<" and ">" means? Why don't they use an abbreviation? Unfortunately that format is also used by the iana Time Zone Database.

In this link: https://www.gnu.org/software/libc/manual/html_node/TZ-Variable.html you can read "The std string specifies the name of the time zone. It must be three or more characters long and must not contain a leading colon, embedded digits, commas, nor plus and minus signs."

So how can this be a valid TZ value?

In this link: https://man7.org/linux/man-pages/man5/tzfile.5.html you can read "Some readers mishandle POSIX-style TZ strings that contain “<” or “>”. As a partial workaround, a writer can avoid using “<” or “>” for time zone abbreviations containing only alphabetic characters."

Obviously the ESP32 is affected by this problem. So is there another official source of a timezone database that uses a format without "<" and ">"?

Thanks for any help!

Cheers Harald

HarryVienna
  • 21
  • 2
  • Why do you need to parse a timezone file in EPS32? I think you should do it on your hosting machine (and possibly using a library). Using timezone file is tricky: you should follow development of it (and e.g. recent changes which remove old data for various countries). – Giacomo Catenazzi Jan 21 '22 at 13:31
  • The ESP32 will have no internet connection. The UTC time comes from a GPS module and I have written a library to get the timezone out of the GPS coordinates (https://github.com/HarryVienna/ESP32-Timezone-Finder-Component) You are right, the timezone data are not static, but I can see no other solution for an offline component,,, – HarryVienna Jan 21 '22 at 23:58
  • Still I do not understand. You can parse timezone on your computer. Prepare the data in a better format for EPS32, possibly removing most of not useful data (and already in EPS friendly format): You define the list of timezones you provide to users, and the offsets, and few rules (summer times). And remember that timezones is not meant to be offline: there are many changes (and on last days). if you are not ready to update the firmware twice a year, you should check to have just UTC offsets (which are constant): better interface stupid then wrong data. – Giacomo Catenazzi Jan 24 '22 at 07:42

1 Answers1

2

Here is the Posix specification for time zone names: https://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap08.html#tag_08_03

For "UNK-4", the "UNK" is the time zone abbreviation (designation in Posix-speak) and -4 is the offset and means 4h east of the prime meridian. Note that the sign of the offset is opposite what everyone else uses, including other parts of Posix.

To differentiate between the abbreviation and the offset, the abbreviation must be all alphabetic characters.

However if you would like the abbreviation to contain non-alphabetic characters, the Posix spec says that you can quote the abbreviation with a leading '<' and a trailing '>'.

  • In the quoted form, the first character shall be the ( '<' ) character and the last character shall be the ( '>' ) character. All characters between these quoting characters shall be alphanumeric characters from the portable character set in the current locale, the ( '+' ) character, or the ( '-' ) character. The std and dst fields in this case shall not include the quoting characters.

So with "<+04>-4", the abbreviation is "+04", while the offset remains 4h east.

Howard Hinnant
  • 206,506
  • 52
  • 449
  • 577