ICU-23364 Add support of the new dayOfMonths CLDR resources (and the "ddd" pattern symbol) to ICU

[richgillam]

This PR adds support in ICU for the new daysOfMonths resources introduced to CLDR by https://unicode-org.atlassian.net/browse/CLDR-7408. It includes:

• Temporary local ICU-data additions for the new resources added in CLDR (they're temporary until we get the "real" ones copied over in a future CLDR integration).
• An update to the CLDR-to-ICU tool to bring the new CLDR resources over into ICU.
• New API on DateFormatSymbols for the new resources (API proposal forthcoming).
• Code to properly handle the new ddd pattern/skeleton sequence in SimpleDateFormat (both formatting and parsing) and DateTimePatternGenerator.
• Unit tests for all of the above.

This PR is large; I split it into 10 separate commits to make it easier to review.

A few issues worth calling out:

• As I mentioned above, I manually brought over the changes to the ICU data that correspond to the new changes in CLDR. The changes were generated by the CLDR-to-ICU tool with my change in place. I'm expecting these changes to be stomped over by the next official CLDR integration (which hopefully won't actually change the dayOfMonths stuff), but this lets us test the implementation in advance of the next CLDR integration landing.
• The temporary ICU-data changes on the Java side required regenerating ALL of the Java .res files, and I think I may have accidentally picked up some stray changes we don't want. I think we lost a few files we want to keep, and there are a couple stray unit-test changes unrelated to the work of this PR that I think happened because we picked up extraneous changes in the Java ICU data. I'm assuming (hoping?) this will all get straightened out in the next CLDR integration.
• There is no testing included for the <dayOfMonth type="x"> resources right now because nothing in the CLDR change uses them. I have tested these resources in Apple ICU on changes that aren't in CLDR-7408. I made two CLDR changes in Apple CLDR that we don't have in open-source CLDR: 1) Support for ordinals (i.e., just "1er") in French, and 2) replacement of the "hanidays" numbering system with equivalent dayOfMonths resources. I can file tickets to add these things to CLDR and add appropriate unit tests for them to ICU. (As I said, I have all that in Apple ICU, so the relevant code here has been tested.)

Checklist

• Required: Issue filed: ICU-23364
• Required: The PR title must be prefixed with a JIRA Issue number. Example: "ICU-NNNNN Fix xyz"
• Required: Each commit message must be prefixed with a JIRA Issue number. Example: "ICU-NNNNN Fix xyz"
• Issue accepted (done by Technical Committee after discussion)
• Tests included, if applicable
• API docs and/or User Guide docs changed or added, if applicable
• Approver: Feel free to merge on my behalf

[richgillam]

NOTE TO SELF: In addition to the current failing tests in Java (which seem to be because I screwed up the Java data generation process and accidentally deleted some key files), I'm also getting reports that the Java code fails on skeletons that include dd (such as yMMMMdd and yMMMdd). The C++ code apparently works. Claude claimed I didn't need to make any DateTimePatternGenerator changes on the Java side, but this is apparently wrong and needs a closer look, plus we need to add the dd tests in both C++ and Java.

[richgillam]

NOTE TO SELF: In addition to the current failing tests in Java (which seem to be because I screwed up the Java data generation process and accidentally deleted some key files), I'm also getting reports that the Java code fails on skeletons that include dd (such as yMMMMdd and yMMMdd). The C++ code apparently works. Claude claimed I didn't need to make any DateTimePatternGenerator changes on the Java side, but this is apparently wrong and needs a closer look, plus we need to add the dd tests in both C++ and Java.

I think I've fixed this now...

[macchiati]

Java looks good; my C++ skills are too rusty for that part.