Last year, I've contacted Xorg people to update the community driven french Dvorak layout, bépo (more info available at http://www.bepo.fr, in french), using bug reports #17821 and 17822. During this update, maintainers from the libX11 project, managing dead key definitions, added new dead keys used by our layout, 'dead_currency' and 'dead_belowcomma', but didn't add the third we need, the 'dead_greek' key, and I tried 2 possible workarounds (one using ISO_Level5_Latch and one with groups), but the side effects was too important and I finally proposed a layout without this key.
Since then, our layout project has matured, the layout is stabilized and available on multiple platforms : Windows, OS X, and Xorg (as Linux and BSD), and quite a lot of our users/contributors are always asking for a dead_greek support from Xorg, as on other platforms.
Indeed, it's a valuable dead key that let you naturally typing Greek letters among a text (useful for mathematical purposes for example), and is very different from a solution with a Greek layout as a second group :
- you don't have to configure a second group, not an easy task for
- with dead_greek letter, you don't need to type a combination to go to the second group, type your letter, and then not forget to go back to your main layout.
- and moreover, we find the greek letters *in the same place* as
the latin ones in the layout, an important point as our layout is very
different from a standard qwerty or azerty one (which the Greek
layout is issued), and has been built carefully with ergonomic
Can you actually add this dead_greek definition in libX11, to let us having our full 'bépo' layout in Xorg ?
I'm seconding this request as, after trying it, I think that a lot of scientifics will enjoy it.
This sounds like a great idea, and I'd certainly welcome a patch. Send your
changes to xorg-devel for review.
I posted on xorg-devel about dead_greek when the rest of that keyboard
was added, looking consensus on whether it was a good addition. It felt
controversial and I didn’t want to push it w/o discussion.
I received no replies from devs, and so let it drop.
Perhaps this time some discussion will ensue.
(In reply to comment #3)
> I posted on xorg-devel about dead_greek when the rest of that keyboard
> was added, looking consensus on whether it was a good addition. It felt
> controversial and I didn’t want to push it w/o discussion.
> I received no replies from devs, and so let it drop.
> Perhaps this time some discussion will ensue.
Hello Jeremy, James,
I'm happy this proposition is examined again :-) Even if the patch should be quite trivial, I have to update it with respect with xorg sources : I'll try to do it in the week or next !
Created attachment 52443 [details] [review]
Definition of dead_greek symbol
first part of dead_greek definitions
Created attachment 52444 [details] [review]
Compositions with the dead_greek symbol
Created attachment 52445 [details] [review]
use of dead_greek symbol in Bépo's keyboard layout
last part of proposed patches.
Comment on attachment 52443 [details] [review]
Definition of dead_greek symbol
Review of attachment 52443 [details] [review]:
The value 0xfe70 is already taken by AccessX_Enable.
(In reply to comment #8)
> Comment on attachment 52443 [details] [review] [review]
> Definition of dead_greek symbol
> Review of attachment 52443 [details] [review] [review]:
> The value 0xfe70 is already taken by AccessX_Enable.
Yes, you're right, I was fooled by the not-incremental order of defines (0xfe6f is followed by 0xfe80...0xfe8b, 0xfed0..0xfed5, then 0xfe70 !
I suggest to use 0xfe8c instead (see new patch attached).
Created attachment 53599 [details] [review]
[new] definition of dead greek symbol
new version of patch, following Andreas remark.
You should send your patch to xorg-devel for review and get it pushed.
How is the push of these patches going? We still have no news.
The patch in attachment #53599 [details] [review] was applied to x11proto in November (commit 67dd3bfe1ea2). The current tar release, xproto-7.0.23, was the first such release with it.
The other patches are for xkeyboard-config; I’m therefore reassigning.
xkeyboard-config 's part of this bug report has been commited in xkeyboard-config's git repository, and should be in 2.6 version : I think this bug could be closed now.