Mailman 3 November 2016 - Datetime-SIG

Re: [Datetime-SIG] Add a strftime/strptime code for timezone ID
by Random832 Nov. 3, 2016

Nov. 3, 2016

On Thu, Nov 3, 2016, at 10:20, Alexander Belopolsky wrote: > My proposal was indeed prompted by the recent discussion on the IANA > mailing list about human readable abbreviations being changed to numeric > codes. It is not uncommon for people to display timezones with '%Z %z' > format which with recent changes would result in "+10 +1000" for say > Asia/Vladivostok. With the proposed %/ code, "%/ %z" would print > "Asia/Vladivostok +1000" which looks like an improvement even over the > original "VLAT +1000". Surely it'd be better to support the CLDR mechanisms for timezone names and abbreviations, and then %Z %z could print "Vladivostok Time +1000" or "Владивосток +1000" (or perhaps "Vladivostok Standard Time +1000", "Владивосток, стандартное время +1000", or perhaps abbreviations could be added to the CLDR Russian locales.) Done right, this would mean people don't even need to change their format strings. Supporting it in strptime is more dicey, but possible - maybe loading a locale* in LC_TIME builds a trie with all the possible timezone names. This would give you the metazone, and CLDR does define a notion of the "primary" tzid for a metazone. It's even locale-sensitive - America_Eastern maps to America/New_York by default, but America/Toronto in canada, etc. > The results of strftime() are not necessarily intended for displaying > directly to end users. An application can have an extra layer that > translates tzids for display and the raw tzids may be useful for saving > times in text files or a database. It would probably be better to have a generic mechanism which could include the tzid for tzid-based timezones, or some other identifier for timezones that work a different way. And maybe a mechanism for timezone-providing modules to be registered so that strptime or some other mechanism knows how to reconstitute a timezone from such a string... something that's missing from your proposal. Remember, Python doesn't actually support IANA timezones out of the box. *where? should setlocale do this even if datetime isn't loaded? do we need a general mechanism for modules to install "set locale hooks"? Is there a more appropriate place to discuss this? i18n-sig seems dead.

1 0

Re: [Datetime-SIG] Add a strftime/strptime code for timezone ID
by Alexander Belopolsky Nov. 3, 2016

Nov. 3, 2016

On Thu, Nov 3, 2016 at 12:38 AM, Random832 <random832(a)fastmail.com> wrote: > On the IANA side of things, the tzid is not intended to be displayed as > a human-readable way to identify the timezone (and this is used to > rebuff demands by users to create tzids for cities like Brasilia and > Beijing that people want to see represented over Sao Paulo and Shanghai) > > What is the reason for wanting this? > My proposal was indeed prompted by the recent discussion on the IANA mailing list about human readable abbreviations being changed to numeric codes. It is not uncommon for people to display timezones with '%Z %z' format which with recent changes would result in "+10 +1000" for say Asia/Vladivostok. With the proposed %/ code, "%/ %z" would print "Asia/Vladivostok +1000" which looks like an improvement even over the original "VLAT +1000". The results of strftime() are not necessarily intended for displaying directly to end users. An application can have an extra layer that translates tzids for display and the raw tzids may be useful for saving times in text files or a database.

1 0

Re: [Datetime-SIG] Add a strftime/strptime code for timezone ID
by Random832 Nov. 3, 2016

Nov. 3, 2016

On Wed, Nov 2, 2016, at 14:33, Alexander Belopolsky wrote: > I propose to add a code %/ to strftime/strptime to print/parse the > timezone name (tzid) such as Australia/Sydney or America/New_York. > > While the choice of code character can be debated, I like "/" because > it is visually similar to "Z", can be associated with the Area-Region > separator in the IANA timezone names and is unlikely to conflict with > any other code. On the IANA side of things, the tzid is not intended to be displayed as a human-readable way to identify the timezone (and this is used to rebuff demands by users to create tzids for cities like Brasilia and Beijing that people want to see represented over Sao Paulo and Shanghai) What is the reason for wanting this?

1 0

Re: [Datetime-SIG] Add a strftime/strptime code for timezone ID
by Paul G Nov. 2, 2016

Nov. 2, 2016

Honestly, I would be very worried about doing something like this in the standard library if you're talking about IANA support in particular. It's a very specific type of time zone data which isn't usually present on Windows systems, for example, and which has to have a very flexible release cycle that I doubt would be compatible with Python's release cycle if Python were to try to supply it. I would say there's two good ways to do something like this (on the format side): 1. Extend the tzinfo API such that there's a version of "tzname()" that represents any kind of canonical "zone string" that doesn't change throughout the year, then allow that to be printed in a formatstr. 2. Offer format string option that inserts str(tzinfo) somewhere, and encourage tzinfo implementations to make str(tzinfo) resolve to some human-readable zone identification. I think in both of these cases it's important to make it clear it's not necessarily the tzid from an IANA string, but just whatever the string representation is, whether that's a GNU-style tz string, an iCal zone spec or whatever. All that said, it also seems a bit unfortunate to deviate from the cross-platform strftime standard. I guess there are already some deviations in the python strftime spec, but that should be at least a minor weight against doing something like this. On the parse side, I think there are a lot of issues. For one thing, the open ended nature of "all potential time zone strings" will probably yield a lot of edge case headaches that will take quite a while to work out (though this may not be too bad as long as multiple open-ended format options are not added). Another problem is that if you have any random string that could be your time zone ID, what kind of tzinfo object is produced by "America/New_York" or something similar? Again we come back to Python favoring IANA and shipping with an IANA zone parser using who knows what data. I'd think the most natural way to implement something like this would be to define an interface for "zone parser" which would take a subset of the string to be parsed and return a tzinfo object, which is passed as an optional parameter to strptime. That way, third parties could provide an implementation that could take a string like "EST5EDT" and produce an appropriate tzinfo object. On November 2, 2016 3:27:18 PM EDT, Alexander Belopolsky <alexander.belopolsky(a)gmail.com> wrote: >On Wed, Nov 2, 2016 at 3:15 PM, Paul G <paul(a)ganssle.io> wrote: > >> Where would the tzid come from? > > >Hmm, I was actually thinking of the time module strftime where %/ can >simply print the value of the TZ environment variable or the system >local >zone setting (which may be tricky to obtain, which is another reason to >do >it in stdlib.) > > >> It's not a standard feature of tzinfo, so wouldn't that involve >changing >> the tzinfo API as well? > > >Yes, to add this to datetime.strftime(), we will need to extend the >tzinfo >API. I have to think some more before making a specific proposal. Any >suggestions? I think this was discussed a few years ago. I'll try to >find >the relevant threads.

1 0

Re: [Datetime-SIG] Add a strftime/strptime code for timezone ID
by Paul G Nov. 2, 2016

Nov. 2, 2016

Where would the tzid come from? It's not a standard feature of tzinfo, so wouldn't that involve changing the tzinfo API as well? On November 2, 2016 2:33:50 PM EDT, Alexander Belopolsky <alexander.belopolsky(a)gmail.com> wrote: >I propose to add a code %/ to strftime/strptime to print/parse the >timezone >name (tzid) such as Australia/Sydney or America/New_York. > >While the choice of code character can be debated, I like "/" because >it is >visually similar to "Z", can be associated with the Area-Region >separator >in the IANA timezone names and is unlikely to conflict with any other >code.

2 1

Add a strftime/strptime code for timezone ID
by Alexander Belopolsky Nov. 2, 2016

Nov. 2, 2016

I propose to add a code %/ to strftime/strptime to print/parse the timezone name (tzid) such as Australia/Sydney or America/New_York. While the choice of code character can be debated, I like "/" because it is visually similar to "Z", can be associated with the Area-Region separator in the IANA timezone names and is unlikely to conflict with any other code.

1 0