ICU-22575 Change AvailableFormatsSink to allow locales to inherit ava…

[richgillam]

…ilableFormats items from the root locale.

Checklist

• Required: Issue filed: https://unicode-org.atlassian.net/browse/ICU-22575
• Required: The PR title must be prefixed with a JIRA Issue number.
• Required: The PR description must include the link to the Jira Issue, for example by completing the URL in the first checklist item
• Required: Each commit message must be prefixed with a JIRA Issue number.
• Issue accepted (done by Technical Committee after discussion)
• Tests included, if applicable
• API docs and/or User Guide docs changed or added, if applicable

[richgillam]

NOTE: I think this is the same problem we've been getting reports from various Web users about. A few quick comments:

• This is the fix I made on the Apple side. There may be a better, cleaner way to do this, although I thought about it a little and this seemed right.
• If there are new test cases I should add that would correspond better to the reports were getting from the browser people, I'll add them. Can somebody suggest some?
• I have no idea why we were doing !isRoot before; there didn't seem to be any bad side effects from taking it out.
• I haven't yet vetted the test changes in this PR; I vetted some of them on the Apple side and confirmed that the new results were the right ones, but I need to go back and do that.
• I haven't yet ported this change to Java. I'll do that before I check in (but feel free to remind me).
• I just realized to put this change up against the main branch, but that it's also needed against the maint-74 branch. I can cherry-pick over there and post another PR, but I'd like some validation of this approach first.

Anyway, sorry this is a little half-baked, but I ran out of time and wanted this up before the Thanksgiving break (I'm off all next week) to get feedback and reassure people that the problem was being worked on.

[FrankYFTang]

There should be a Java counter part, right? Could we change the Java one in this PR together?

[richgillam]

There should be a Java counter part, right? Could we change the Java one in this PR together?

Of course. I just ran out of time before the break to do that, and I kind of wanted some feedback on what I'm doing and whether it's a good idea or there's a better way to do it. If people are behind this (or don't care), I'll try to put up a new commit that extends the fix to the ICU4J side.

[FrankYFTang]

please fix the failing tests

[richgillam]

Okay, I think this is finally ready to be reviewed. I've fixed the build failures, verified that the changed unit test results make sense, added a new unit test that specifically replicates the problem described in the Mozilla bug report, and ported all of those things to Java. @markusicu @FrankYFTang @pedberg-icu , PTAL.

[macchiati]

I looked this over, and it all looks good.

One could also write a little tool that generates patterns from skeletons for different languages, where the skeletons are unions of those of the root and child locales, and captures that in a file. Then run it with and without the code change, and spot-check the differences.

But I'm maybe a bit paranoid..., so glad to approve if you don't have the time to do that.

[pedberg-icu]

The code change and the expected data changes in tests all seem good to me. The expected data changes seem to be places where the CLDR data had ↑↑↑ assuming inheritance from root, so making the ICU code align with that seems good.

[richgillam]

One could also write a little tool that generates patterns from skeletons for different languages, where the skeletons are unions of those of the root and child locales, and captures that in a file. Then run it with and without the code change, and spot-check the differences.

But I'm maybe a bit paranoid..., so glad to approve if you don't have the time to do that.

It definitely seems worth doing, but I really don't have the time. The affected cases should pretty much all be situations where CLDR changed a literal value to ↑↑↑ expecting to inherit from root. Although as you can see, making that change uncovered a few unit tests that were expecting something different (e.g., deriving the pattern algorithmically from one of its sibling patterns).

[jira-pull-request-webhook]

Hooray! The files in the branch are the same across the force-push. 😃

~ Your Friendly Jira-GitHub PR Checker Bot

[richgillam]

Thank you for your reviews, @pedberg-icu and @macchiati ! I just squashed my commits-- would somebody mind re-approving?

[FrankYFTang]

LGTM. Thanks for working on this.