offtopic: implicit weights in DUCET. can I ask for some advice?

[gpawru]

Hello! I'm really, REALLY sorry for going off-topic. I understand that this might not be entirely relevant, but perhaps you could help me figure it out?

I'm writing a collator, and for the Tangut language, the weights are implicit.

In this case, I'm considering the DUCET, without delving into the CLDR territory.

In the specification (TR#10), it's stated:

Siniform ideographic scripts: Tangut = Assigned code points in Block=Tangut OR Block=Tangut_Components OR Block=Tangut_Supplement

So, as I understand it, it includes U+17000 ..= U+18AFF and Tangut_Supplement.

If we trust the UCD Blocks.txt, Tangut_Supplement ends at U+18D7F. If we trust UCA DUCET allkeys.txt, Tangut_Supplement ends at U+18D8F.

So is it U+18D7F or U+18D8F?
Or should one calculate implicit weights according to the scheme for Tangut_Supplement for U+18D00 ..= U+18D08, and consider U+18D09 as Unassigned (which seems logical, but the confusion about UCD/UCA remains)?

I understand that the issue is minor, and perhaps not an issue at all, but if there's a mistake somewhere here, it would be good to correct it. If not, and it's just me being a bit stupid, I would really appreciate comments from smart and knowledgeable people in this matter.

Thank you!

[Manishearth]

I believe 7F is correct. Not sure what's going on with DUCET.

But since the area in question is unassigned anyway, I don't think it should matter for your use case.

[gpawru]

Sometimes I myself start to think that my meticulousness is excessive. But in fact, the inaccuracies made now can lead to problems that will arise later :) Thank you for your response, have a good day!

[Manishearth]

Got a proper answer, Unicode 14.0 decided to size down the block, and it wasn't propagated to all files. Hopefully will be fixed in the Unicode 16 files.

[gpawru]

Huge thank you!