I can open another issue to keep track of the permanent fix for this (and for unicodedata), either pulling these aliases into a separate little crate or eventually adding them to unicode_names.

is this doing the same thing? 49adb5c

does that compile? that's basically what I was wanted to go with but there's some characters that need to be escaped or else phf can't really generate the map correctly for them.

Yes, that does compile and working well at least for BACKSPACE.
include_str! obtains the whole file content as a valid str. And I didn't put them in source code but handled only with string operations. I think something similar like this version can be filed as patch to unicode_names2

Yes, that does compile and working well at least for BACKSPACE.

Ah then push it through to here, I have no issue.

I think something similar like this version can be filed as patch to unicode_names2

I still am not sure for this since they already have the data in UnicodeData.txt and they use their own implementation of phf (as far as I can tell). But if you wanna close this PR and go with the patch to unicode_names2 I'd be fine with that too.

wow, i didn't know their phf is their own's.

I tried to add alias, but I found unicode_names2 cannot regenerate master branch's output using the same version of generator 😂
directly putting your version seems more realistic

Close in favor of #4582

We still need the tests from here, right?