Remove or correct non-NFD confusables

[roozbehp]

See https://github.com/unicode-org/properties/issues/486

[roozbehp]

Sorry, this is a little hard to review, but be assured that I triple-checked it myself very thoroughly. confusablesSummary.txt is relatively easier to check compared to other files, but it's still pretty long.

Anyways, please take a look.

[roozbehp]

Converting to draft to fix Java style issues.

[roozbehp]

Java style fixed. Ready for review.

[josh-hadley]

@roozbehp, apologies for taking a long time to get this reviewed. It was tricky as advertised, but all looks good. The process did make me want to see if a one-time revamp of confusables-source.txt to force most or all entries to our new preferred format universally would be worthwhile (presumably making reviews/maintenance like this a little less taxing).

[macchiati]

Note that we need to add confusable changes to the Migration section of each release. That is, if the skeleton(X) changes from Y to Z, that requires implementations that have mappings to skeletons need to update.

(We had some breakages of indexes in production softward with the U17 integrations; luckily caught by unittests.)

[roozbehp]

Note that we need to add confusable changes to the Migration section of each release. That is, if the skeleton(X) changes from Y to Z, that requires implementations that have mappings to skeletons need to update.

We're going to have a lot of confusable changes in Unicode 18.0, but this specific pull request should not affect the skeleton of any string, since a conformant implementation would not use any of the data removed in this pull request: They simply don't occur in NFD form that the algorithm applies before looking for the prototypes.

[macchiati]

I'm not worried about dropping all the non-nfd forms. I just want to make sure that we alert people of changes, and that anytime we had X -> Y before, we have nfd(X) -> nfd(Y) after (when nfd(X) is a single code point -- and you include a onetime test that that is true.

…

On Mon, Nov 24, 2025, 18:33 Roozbeh Pournader ***@***.***> wrote: *roozbehp* left a comment (unicode-org/unicodetools#1238) <#1238 (comment)> Note that we need to add confusable changes to the Migration section of each release. That is, if the skeleton(X) changes from Y to Z, that requires implementations that have mappings to skeletons need to update. We're going to have a lot of confusable changes in Unicode 18.0, but this specific pull request should not affect the skeleton of any string, since a conformant implementation would not use any of the data removed in this pull request: They simply don't occur in NFD form that the algorithm applies before looking for the prototypes. — Reply to this email directly, view it on GitHub <#1238 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACJLEMCGVFUS42MPA5ORERD36O5YDAVCNFSM6AAAAACL6VXZVWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTKNZTGU2DIMJWGQ> . You are receiving this because you commented.Message ID: ***@***.***>