Re: [9front] Re: removal of /lib/unicode

[Jacob Moody <moody@mail.posixcafe.org>]

On 3/25/23 21:05, sl@stanleylieber.com wrote:
>> /lib/unicode was removed in favor of just using the upstream data file
>> it was ripped from. I have updated the references in the man pages
>> to reflect this. I apologize for that mistake.
> 
> ah. no problem, i was just curious what was going on. i'll never
> remember the new file path off the top of my head, but on the other
> hand i guess i'm not manually checking /lib/unicode all that often
> anyway.
> 
> why didn't we just have the mkfile put the information in the original
> place?

Yes a couple reasons. /lib/unicode only included those in the 16bit BMP domain.
When extending this to include all the modern code points you need more hex digits.
Unless you went with using two different zero padded first columns. I could not
find a good way of updating these for things past the BMP and not breaking someone's
script.

Likewise /lib/unicode while derived from UnicodeData.txt does not include the data fields
that are needed to do the generation. So either way I was going to need to update each line
of the file anyway. I figured to instead just say 'we use the official txt file' is bit nicer
in my opinion.

Likewise other files were desired for the generation, so moving things to a directory seemed
to make more sense.