CLDR-14493 Rework CLDR-2698/PR-964 changes to make nb a child of no, not an alias to it

[pedberg-icu]

Checklist

• Issue filed: https://unicode-org.atlassian.net/browse/CLDR-14493
• Updated PR title and link in previous line to include Issue number

The changes in CLDR-2698 / PR-964 swapped the role of nb and no, making nb a deprecated alias to no, and removing the xml files for nb, while also making no a parent locale with child nn. The first part of this would have caused serious compatibility problems for some CLDR and ICU clients, including Windows.

This PR reworks those changes to make no a parent of both nb and no; nb is not deprecated or aliased. All of the xml files that used to be present for nb and nn are restored in this PR, though in some cases their content has been moved elsewhere (as in the previous PR) so that the nb/nn xml files restored in this PR are all empty stubs. This includes having empty stubs in collation/ for nb and nn; however the "actual" collation content for these is in no.

Some constraints/considerations/changes in this PR:

• In common/main, this restores nb.xml, nb_NO.xml, and nb_SJ.xml; nb is the default content for no, and nb_NO is the default content for nb. Both of the regional sublocales of nb (or no) are empty in any case.
• Since we can only have one default content locale for a parent, this means that we cannot have no_NO (and thus it makes no sense to have no_SJ). That is fine, we did not have them in earlier CLDR versions. So we have no (with no regional locales) as the parent of nb and nn, and then nb as the parent of nb_NO and nb_SJ (and nn as the parent of nn_NO).
• LocaleIDParser.getParent is fixed to not assume locales without '_' do not have a non-root parent (Mark I think this is the fix you were mentioning).
• TestInheritance.TestLikelyAndDefaultConsistency is fixed to use LocaleIDParser.getParent, not getSimpleParent.
• ... I may update this to mention other notable changes

[macchiati]

Don't think this is necessary because plurals do inheritance, but doesn't hurt.

[srl295]

LGTM…

[srl295]

i don't know how the keyboard files are used, but seems plausible

[pedberg-icu]

I was assuming that what was being referred to here was the locale ID used on the platform. But this actually goes back to the code that was used in CLDR 38, I had chnaged in my previous Norwegian PR since nb was no longer valid.

[srl295]

This is a generated file. Will have to watch it when its next generated.

[srl295]

This is a snapshot for review purposes (an output, not an input). The actual content is regenerated in the ST. So this may not work properly and may need some followon. Shouldn't hold this PR though. s

[pedberg-icu]

I had looked up what that Wikidata entity referred to and it is in fact nb, not no. I had changed it before when we aliasing so no really meant nb.

[srl295]

@pedberg-icu right, but nothing reads entityToCode.tsv, it's output to give a historical record of the process.

[jefgen]

Thank you Peter! 👍