Bug#436923: psili/dasia problem (Aarghhh!)

[Jan Willem Stumpel]

I have forwarded this bug on the upstream bugzilla at the URL
above. Feel free to add any comments there if you think it
could help. I know nothing about Greek accents and I don't have
a Greek polytonic keyboard, so I won't be able to help much :)

At the moment in Sid, Greek breathing signs *again* do not work.

1- First some Greek developers called the breathing signs
(falsely) "dead_horn" and "dead_ogonek" -- this was an ugly
hack (and acknowledged as such by said developers) which worked
fine for Greeks, but for nobody else on the planet (i.e. not
for anyone working in a non-Greek locale).

2- Then "dead_horn" and "dead_ogonek" (which are entirely
non-Greek keysyms, just borrowed by said Greek developers for
the occasion) were replaced by U0313 and U0314 respectively. We
enjoyed a short period in which Greek breathing signs could
actually be typed by anyone in the whole wide world.

3- Then somebody thought it was a bright idea to again change the
definitions of the breathing signs. In the Compose file U0313
was changed into U10000313. What could the poor user do? Make
the same change in /usr/share/X11/xkb/symbols/gr, of course.
Debian did not do it, so the user had to do it by hand.

4- But in the latest Sid, the Compose file has reverted to
definitions like U0313, while /usr/share/X11/xkb/symbols/gr has
reverted to the use of "dead_horn" and "dead_ogonek". So we are
literally back to #1. Sigh..

TIMETE DANAOS ET DONA FERENTES!

When will we ever get a stable system for entering the Greek
breathing signs? It is not rocket science. It only requires that
the people maintaining the keyboard files agree with the people
maintaining the Compose file. I suspect that there are two
different groups of Greeks working on those files, independently
of one another. This has to stop. Somebody has to knock some heads
together.

Regards, Jan

[William J Poser]

Jan Willem Stumpel wrote:
--

I suspect that there are two different groups of
Greeks working on those files, independently
of one another. This has to stop. Somebody has
to knock some heads together.

Judging from the lessons of history, the most effective
approach is a Persian threat. Perhaps President Ahmadinejad's
help could be enlisted. :)

[Αλέξανδρος Διαμαντίδης]

Post by Jan Willem Stumpel
When will we ever get a stable system for entering the Greek
breathing signs? It is not rocket science. It only requires that
the people maintaining the keyboard files agree with the people
maintaining the Compose file.

Fortunately, 6 weeks ago, definitions for new keysyms for psili and
dasia were commited to X: (thanks to James Cloos)

http://bugs.freedesktop.org/show_bug.cgi?id=9306

http://gitweb.freedesktop.org/?p=xorg/proto/x11proto.git;a=commitdiff;h=40ed4eef92e31fcf7ea0a436e1a00cdf49484c1b

#define XK_dead_abovecomma 0xfe64
#define XK_dead_psili 0xfe64 /* alias for dead_abovecomma */
#define XK_dead_abovereversedcomma 0xfe65
#define XK_dead_dasia 0xfe66 /* alias for dead_abovereversedcomma */

xproto-7.0.11 which contains this change hasn't been released yet, but
after it is and the new keysyms start being used, this will be fixed
once and for all...

Alexandros

[Jan Willem Stumpel]

Post by Αλέξανδρος Διαμαντίδης
Fortunately, 6 weeks ago, definitions for new keysyms for psili and
dasia were commited to X: (thanks to James Cloos)
[..]

Hi Alexandros,

This is of course very good news. But does this also guarantee
that the psili/dasia definitions in /usr/share/X11/xkb/symbols/gr
and in /usr/share/X11/locale/en_US.UTF-8/Compose will, from now
on, always agree?

Regards, Jan

[Alexandros Diamantidis]

Post by Jan Willem Stumpel
This is of course very good news. But does this also guarantee
that the psili/dasia definitions in /usr/share/X11/xkb/symbols/gr
and in /usr/share/X11/locale/en_US.UTF-8/Compose will, from now
on, always agree?

Well, I believe that this disagreement was caused by the fact that no
keysyms were obviously the correct ones for dead psili and dasia. People
saw the original dead_horn and dead_ogonek, which as you said was an
ugly hack, and decided to fix them by using U0313 and U0314 instead. But
this wasn't correct either, since these are not *function* keysyms,
which should be used for dead keys, but correspond to printable Unicode
characters, albeit combining ones. Also this didn't work if they didn't
agree with the Compose file, or with the internal composition tables in
GTK+. So, others saw this problem, and tried to revert the change or fix
it another way (the U10000313 stuff is probably confusion created by the
fact that UXXXX = 0x01000000 + 0x0000XXXX).

Hopefully, all places where the old hacks appeared will now converge to
the new definitions, but of course somebody has to submit the patches
and make sure that everything's consistent. The only problem I can think
of is that the new keysyms can't be used before systems are upgraded to
xproto-7.0.11 - I see that Debian Sid already has it.

I'll post a notice about the new definitions to the Greek I18N list,
to raise awareness of them.

Regards,
Alexandros

[James Cloos]

Αλέξανδρος> Fortunately, 6 weeks ago, definitions for new keysyms for
Αλέξανδρος> psili and dasia were commited to X: (thanks to James Cloos)

Αλέξανδρος> xproto-7.0.11 which contains this change hasn't been
Αλέξανδρος> released yet,

I pushed out the 7.0.11 release back on September 8:

http://xorg.freedesktop.org/archive/individual/proto/xproto-7.0.11.tar.gz
http://xorg.freedesktop.org/archive/individual/proto/xproto-7.0.11.tar.bz2

But it was not in time to be part of the katamari release. Hopefully
distributions have picked it up, though.

Αλέξανδρος> but after it is and the new keysyms start being used,
Αλέξανδρος> this will be fixed once and for all...

With xproto-7.0.11 (for the key syms), current xlib (for the Compose file)
and current xkeyboard-symbols (for the XKB data) it should all work.

-JimC

[Julien Cristau]

Version: 2:1.1.4-1

forcemerge 436923 484062
kthxbye

Post by Jan Willem Stumpel

I have forwarded this bug on the upstream bugzilla at the URL
above. Feel free to add any comments there if you think it
could help. I know nothing about Greek accents and I don't have
a Greek polytonic keyboard, so I won't be able to help much :)

At the moment in Sid, Greek breathing signs *again* do not work.
1- First some Greek developers called the breathing signs
(falsely) "dead_horn" and "dead_ogonek" -- this was an ugly
hack (and acknowledged as such by said developers) which worked
fine for Greeks, but for nobody else on the planet (i.e. not
for anyone working in a non-Greek locale).
2- Then "dead_horn" and "dead_ogonek" (which are entirely
non-Greek keysyms, just borrowed by said Greek developers for
the occasion) were replaced by U0313 and U0314 respectively. We
enjoyed a short period in which Greek breathing signs could
actually be typed by anyone in the whole wide world.
3- Then somebody thought it was a bright idea to again change the
definitions of the breathing signs. In the Compose file U0313
was changed into U10000313. What could the poor user do? Make
the same change in /usr/share/X11/xkb/symbols/gr, of course.
Debian did not do it, so the user had to do it by hand.
4- But in the latest Sid, the Compose file has reverted to
definitions like U0313, while /usr/share/X11/xkb/symbols/gr has
reverted to the use of "dead_horn" and "dead_ogonek". So we are
literally back to #1. Sigh..
TIMETE DANAOS ET DONA FERENTES!
When will we ever get a stable system for entering the Greek
breathing signs? It is not rocket science. It only requires that
the people maintaining the keyboard files agree with the people
maintaining the Compose file. I suspect that there are two
different groups of Greeks working on those files, independently
of one another. This has to stop. Somebody has to knock some heads
together.

It seems that the Compose file was fixed upstream in libX11 1.1.4. Can
you install the libx11-data package from experimental to verify? I hope
to be able to upload an updated version to sid in the near future.

Cheers,
Julien

[George W Gerrity]

Post by Julien Cristau
Version: 2:1.1.4-1
forcemerge 436923 484062
kthxbye

Post by Jan Willem Stumpel

I have forwarded this bug on the upstream bugzilla at the URL
above. Feel free to add any comments there if you think it
could help. I know nothing about Greek accents and I don't have
a Greek polytonic keyboard, so I won't be able to help much :)

At the moment in Sid, Greek breathing signs *again* do not work.
1- First some Greek developers called the breathing signs
(falsely) "dead_horn" and "dead_ogonek" -- this was an ugly
hack (and acknowledged as such by said developers) which worked
fine for Greeks, but for nobody else on the planet (i.e. not
for anyone working in a non-Greek locale).
2- Then "dead_horn" and "dead_ogonek" (which are entirely
non-Greek keysyms, just borrowed by said Greek developers for
the occasion) were replaced by U0313 and U0314 respectively. We
enjoyed a short period in which Greek breathing signs could
actually be typed by anyone in the whole wide world.
3- Then somebody thought it was a bright idea to again change the
definitions of the breathing signs. In the Compose file U0313
was changed into U10000313. What could the poor user do? Make
the same change in /usr/share/X11/xkb/symbols/gr, of course.
Debian did not do it, so the user had to do it by hand.
4- But in the latest Sid, the Compose file has reverted to
definitions like U0313, while /usr/share/X11/xkb/symbols/gr has
reverted to the use of "dead_horn" and "dead_ogonek". So we are
literally back to #1. Sigh..
TIMETE DANAOS ET DONA FERENTES!

I believe the quote should be “Timeo Danaos et dona ferentes” (I fear
the Greeks, especially (or ‘even’) bearing gifts).

Post by Julien Cristau

Post by Jan Willem Stumpel
When will we ever get a stable system for entering the Greek
breathing signs? It is not rocket science. It only requires that
the people maintaining the keyboard files agree with the people
maintaining the Compose file. I suspect that there are two
different groups of Greeks working on those files, independently
of one another. This has to stop. Somebody has to knock some heads
together.

It seems that the Compose file was fixed upstream in libX11 1.1.4.
Can
you install the libx11-data package from experimental to verify? I hope
to be able to upload an updated version to sid in the near future.
Cheers,
Julien
--
Linux-UTF8: i18n of Linux on all levels
Archive: http://mail.nl.linux.org/linux-utf8/

[Werner LEMBERG]

Post by George W Gerrity

Post by Jan Willem Stumpel
TIMETE DANAOS ET DONA FERENTES!

I believe the quote should be “Timeo Danaos et dona ferentes” (I
fear the Greeks, especially (or ‘even’) bearing gifts).

However, Julien's modification to replace the first person singular
with the plural imperative is appropr