`.getTimezones()` seems problematic about proposal-intl-locale-info

[anba]

1. First off which timezone would be picked as this information isn't necessarily available (definitely not in TZDB).

ICU generates this info from https://github.com/eggert/tz/blob/main/zone.tab (plus additional overrides from https://github.com/unicode-org/icu/blob/main/icu4c/source/tools/tzcode/icuregions.)

[FrankYFTang]

1. the canonical timezone requirement was suggested by @anba in #46 changed in #61 and reach TC39 consensus in
   Dec 14, 2021 meeting (see
   https://github.com/tc39/notes/blob/main/meetings/2021-12/dec-14.md#conclusionresolution-7 and https://docs.google.com/presentation/d/1PZ0_WiE9PNInY2bgyHGJH0DbKd0PKL9RApXxVPKJjUY/edit#slide=id.ge3738e58f4_0_135 )

2. "What are the values for this? ... which timezone would be picked as this information isn't necessarily available (definitely not in TZDB)."
   See the First and third column in the file zone.tab and zone1970.tab in tzdb

3. "destroy an intentional order provided by underlying data."
   What "intentional order provided by underlying data." are you talking about? If you believe " which timezone would be picked as this information isn't necessarily available (definitely not in TZDB)." then how could you also believe there is an "intentional order provided by underlying data." which you claim that should not even exist? I do not quite understand your logic. This is a double-bind argument. Either you believe there are no such list so there are no way there is a order for a non existing list or you believe there is a list which the order is important then you have to believe such list exist. You cannot believe both in the same time.

The order requirement was
A. discussed in #28
B. changed in #37
C. Discussed in TG2 https://github.com/tc39/ecma402/blob/master/meetings/notes-2021-06-03.md#normative-sort-time-zone-in-ordered-as--arrayprototypesort- and
D. Presented to TC39 in the July 2021 meeting

From all the document I can read from tzdb, there are no information suggest there is an underline order between records (rows) inside zone*.tab

4. "Presume Locale 'is_IS' then the canonical timezone according to IANA is Africa/Abidjan which isn't even in the country."
   I have no idea how do you get the Africa/Abidjan from 'is_IS'

The algorithm in https://tc39.es/proposal-intl-locale-info/#sec-time-zones-of-locale
said in step 3

Let region be the substring of locale corresponding to the unicode_region_subtag production of the unicode_language_id.

so in this case, the region is "IS"

What I got is Atlantic/Reykjavik .
Simply in tzdb

~/tz$ egrep ^IS zone*
zone.tab:IS	[phone]	Atlantic/Reykjavik

[FrankYFTang]

All the issues. you asked about is either already discussed and reached consensus in TG2 as well as in TC39 the last 3 years or already addressed by TZDB data files. So I will close this issue for now. Feel free to reopen if you still have issue about this.