Re: Bug#228486: Apology to german users required in the release notes

[Michael Stone <mstone@debian.org>]

On Tue, Aug 31, 2004 at 02:14:48PM -0700, Thomas Bushnell BSG wrote:
> Actually that's not a "real technical reason".  If you could point to
> a package with a bug that would be a serious problem if people used
> UTF-8, then that would be a real technical reason.  But saying "it
> isn't required for sarge" doesn't mean that it actually doesn't
> *work*.

I didn't say it did. I pointed to a potential problem. For all I know,
in an exclusively german context UTF-8 won't cause problems at all. What
I did say is that full multibyte compliance isn't guaranteed. That's a
fact. Now, those suggesting that a bunch of people who aren't currently
using UTF-8 suddenly start using it should look at the current state of
affairs and decide whether that's a good idea. IOW, is the quote issue a
big enough problem that we should suggest that all german users by
default use a less tested environment? I would suggest that those
proposing the change bear some responsibility for testing that
configuration from scratch to see what kind of impact it might have on a
new user's experience, compared to the current mode of operation. If
someone were to do that in a comprehensive fashion then I'd have no
question at all about the proposed change--but I doubt there's really
time for that given the number of packages in sarge. 

> FWIW, I use a UTF-8 locale on sarge and sid exclusively, and I have
> seen no multibyte-related problems.

I also use a UTF-8 locale, and I can assure you that there are. Most are
surmountable, but will raise the complexity of an installation. (It's
not just "use LANG=de_DE.UTF-8 and everything will magically work" in an
upgrade scenario. Maybe not even in a new install--I don't know. Think
things like terminal emulator settings, fonts, remote sessions, etc.)
There are definately, e.g., issues with multibyte support in coreutils.
How much impact the issues have in the current context, I don't know.
Several weeks into the freeze might not be the right time to start
quantifying this.

Mike Stone