Unicode character collation properties and ordering

[ericpruitt]

I'm investigating implementing the Unicode Collation Algorithm using utf8proc for normalization and character property information. I know different locales use different collation rules, but the standard supplies DUCET, the Default Unicode Collation Element Table which generally produces much more palatable results than something like a naive wcscmp(3). Would the y'all consider accepting a patch implementing a collation comparison function and adding DUCET data to the property tables?

[stevengj]

That seems reasonable.

[vadz]

Just curious, has there been work on this since then or any plans to add collation support to utf8proc? It would be great to have it, as right now using ICU is required to sort things correctly.