Bug#386385: xkb-data: Wrong codes for Greek polytonic breathing signs
Denis Barbier wrote:
> But we cannot be sure that they will enter testing together, so adding a
> transition plan is desired. Here is mine:
> 1. Rewrite /usr/share/X11/locale/el_GR.UTF-8/Compose to include
> /usr/share/X11/locale/en_US.UTF-8/Compose (as in pt_BR.UTF-8) and
> only add Greek specific compose sequences.
> 2. Add compose sequences with U0313 and U0314 to en_US.UTF-8/Compose.
> Note that this file already contains such sequences, so maybe there
> is nothing to do here.
> 3. Push these changes upstream.
> 4. When en_US.UTF-8/Compose and el_GR.UTF-8/Compose are fixed in
> testing, modify xkb/symbols/gr as requested, and close #386385.
> 5. Drop el_GR.UTF-8/Compose after etch if this file becomes useless.
> If this roadmap makes sense to you, can you please file bug
> reports with patches against libx11-data for steps 1 and 2?
> This is much easier to perform by people reading Greek ;-)
I am neither a Greek, nor a real expert in Ancient Greek, so I
want to be very careful. I am just interested in making the
'breathing signs' accessible to *international* users (using a
non-Greek UTF-8 locale).
Perhaps a safe way to do this, is to use the following transition
plan:
1. Add U0313 and U0314 sequences to the Greek UTF-8 Compose file.
Because U0313 and U0314 never occur together on the same
line, this can simply be done by running the file through the
following filter:
#!/usr/bin/perl
while (<>) {
print $_;
if (/dead_horn/) {
s/dead_horn/U0313/;
print $_;
}
elsif (/dead_ogonek/) {
s/dead_ogonek/U0314/;
print $_;
}
}
This would keep all the original definitions, and would also
not damage any Greek-speficic tricks that might be present in
the Greek Compose file (this is guaranteed because the "real"
dead_horn and dead_ogonek do not occur in Greek), but it would
make an "international" keyboard (with U0313 and U0314) usable
for Greek users with el_GR.UTF-8 locale.
2. Changing /usr/share/X11/locale/en_US.UTF-8/Compose is not
necessary for the moment. If there are any Greek-specific
tricks that must be added, they can be added later.
3. For the rest, the same as your transition plan.
Do you still want me to send a separate bug report to libx11-data,
or do you take it up with the libx11-data maintainer directly? I
noticed before that Debian does not have a good mechanism for
reporting this kind of "double bug". Maybe we can find a way of
following this up by CC-ing.
Regards, Jan
Reply to: