On Wed, Jul 03, 2024 at 06:27:32AM +0200, Christoph Hellwig wrote: > On Tue, Jul 02, 2024 at 06:59:56PM -0700, Darrick J. Wong wrote: > > > > $ wget https://www.unicode.org/Public/UCD/latest/ucd/UnicodeData.txt > > > > $ grep -E '(zero width|invisible|joiner|application)' -i UnicodeData.txt > > > > > > Should this be automated? > > > > That will require a bit more thought -- many distro build systems these > > days operate in a sealed box with no network access, so you can't really > > automate this. libicu (the last time I looked) didn't have a predicate > > to tell you if a particular code point was one of the invisible ones. > > Oh, I absolutely do not suggest to run the wget from a normal build! > > But if you look at the kernel unicode CI support, it allows you to > place the downloaded file into the kernel tree, and then case make file > rules to re-generate the tables from it (see fs/unicode/Makefile). Ah ok got it, will take a closer look tomorrow. --D
