XPost: linux.debian.maint.boot
Hi/やあ,
It took me much longer than planned before getting back to the buggy
Japanese support we've been having in the graphical installer. Long
story short, Unicode is definitely *not* helping in this particular
area[1], and we end up having weird characters[2] when the Japanese
language is picked up.
1.
https://en.wikipedia.org/wiki/Han_unification
2.
https://heistak.github.io/your-code-displays-japanese-wrong/
Kentaro-san worked on this a lot, teaching us about the issue, proposing solutions, even exploring ways to keep the size increase to a minimum.
You'll find a number of discussions in the bug report and on the main
merge requests (a number of others have been opened while exploring
ideas):
https://bugs.debian.org/1037256
https://salsa.debian.org/installer-team/rootskel-gtk/-/merge_requests/5
I've come back to this topic lately and it looks like sticking to the MotoyaLCedar font is the best compromise (not as nicely-looking as BIZ UDPGothic, but much smaller!).
At the moment, I'm using a trick on the debian-installer side to get
this font embedded without the usual “font ships a udeb, d-i lists it,
and boom”:
https://salsa.debian.org/installer-team/debian-installer/-/commit/ee21b57e30138b044245fe2f7394a942dfec6945
https://salsa.debian.org/installer-team/debian-installer/-/commit/7aedbaf6cee3abb3333fe146536febea624f21ec
If the addition of a fonts-motoya-l-cedar-udeb would be acceptable, this
would be much slicker:
https://salsa.debian.org/installer-team/debian-installer/-/commit/c3773974bbd99935f3ed7a9c0f8cb6eb05e2184a
Kentaro-san adjusted the proposed udeb addition following some remarks
of mine, and I've confirmed that the clean approach would work as well
as the dirty one (runtime-wise); as a nice bonus, we get the package
registered in Built-Using just like other font udebs.
If Hideki-san agrees, this could look like this:
https://salsa.debian.org/fonts-team/fonts-motoya-l-cedar/-/merge_requests/1
The package was last uploaded in 2019, has been in sync between
unstable and testing ever since, and a udeb addition shouldn't
interfere with anything else (that I can think of).
Release team, do you prefer enforcing the no new package rule (which
has been active for many weeks already), and our sticking to the dirty approach? Or would you be open to having the udeb addition reach trixie
so that things are a little cleaner?
In hindsight, we could have requested an extra udeb for each candidate
font, but I usually prefer coming up with a plan instead of brute
forcing my way through…
I'd totally understand either answer, I know we're way past due
already… I'd just hate not asking after all the time spend on
preliminary work and testing.
Thanks for your time.
Cheers,
--
Cyril Brulebois (
[email protected]) <
https://debamax.com/>
D-I release manager -- Release team member -- Freelance Consultant
-----BEGIN PGP SIGNATURE-----
iQIzBAABCgAdFiEEtg6/KYRFPHDXTPR4/5FK8MKzVSAFAmgyYRgACgkQ/5FK8MKz VSDAlQ//UIV26EusGJTkgkHkXmZZfBGLl+v/FIm3SAdNMCm/Pjuhco1/g384NPzv krur5i2ggLdV2ieeg4lcTY7zAyqGDCO1vPs/JGGKl1ErrESARX6YTd88gpp4IB0v KqGCqFYpD8Sv7XPAXR1RBwNOngIeT9X+J6EBAZ1iC0ZOe/Ek+ebdZgGydqubq6D2 DC3z79CypApb5jGYmWqFBnS/0JQ/uXog53mzTbqbKkC3W5gQKnmFI27MduVL6PlE 1xr2wZz3zzAEWRhAGnlA8KnxIsN9TUAAbXUm1ipJKoslgPASCjKUV23F78+UYABB 6FNPVNNH08LJap4jqByAQWJYCfJ0pYdqLE2H+C4NOCh0b4w5W6/SnBZXyUnHn4wk D1BBuMmL0AjTej+ZCGtJC2aVnr/cfmIRD4vYzi+ctyo5HS2+l8sHP5wLfuMqZ10i /EGs65LYsLYKvBiOdXLeYYbsRgKf0R3olHbSxoI5WY43nfobvi92LJDj7w58v3aD DNE2ab+MprtbZqxJuxHhK2MFaToDmPknl/CFglxD4Tpj8+SN0OL92PDRAwlVwq/8 0vq46ZewgSgYlVXosUUhgpgvjQU2Izlj5QOzm+YRoi1tDHJWy3zJ9gRdlh2+Fucq 6w/6fuU/nQPoFl065ec4U/ST2XgnuTAi97vXo0suVvmH3XEZCd4=
=GTeM
-----END PGP SIGNATURE-----
--- SoupGate-Win32 v1.05
*