Bug#258399: xlibs: dvorak keyboard layout is missing right alt key
On Tue, 2004-08-31 at 18:08, Denis Barbier wrote:
> On Mon, Aug 30, 2004 at 12:08:07PM -0400, Adam C Powell IV wrote:
> > Interestingly, I had thought that the absence of Alt_R in mod2 was the
> > problem with the Dvorak/switched layout. But it's also missing in the
> > default US/no switch layout, in which right Alt works just fine...
> > FYI, in the case of Dvorak/no switch, right Alt does nothing; with
> > US/switch, right Alt works fine. So Dvorak seems to be the problem.
> Great, thank you for your detailed report.
> Your analysis is right, swapcaps and mod2 are not causing this bug.
> Since xmodmap output is similar in both cases, one could believe that
> their modifiers are the same, but this is wrong. The difference is
> more subtle, it becomes visible when running
> $ xmodmap -pke | grep 113
> (because 113 is the keycode of right Alt key in your case) with both
> layouts (do not swap caps/ctrl to make less changes) and compare their
> keycode 113 = Alt_R Meta_R
> keycode 113 = ISO_Level3_Shift Multi_key
> With pc/us, right Alt key is bound to "Alt_R Meta_R". Unfortunately
> xmodmap does not tell why Alt_R is bound to mod1, this is because XKB
> is much more powerful and xmodmap is unable to display all details.
> The reason can be found by running
> $ xkbcomp :0
> and searching for Alt_R in the generated server-0.xkb file.
> OTOH right Alt key is bound to ISO_Level3_Shift (and thus mod5) with
> dvorak layout, so pressing this key grabs the 3rd column found in
> /etc/X11/xkb/symbols/pc/dvorak (ie. dead keys). This event is
> intercepted by XKB and not sent to your window manager.
> So in fact, your right Alt key works as expected, but not as you want.
I see, so dvorak has different right alt behavior from pc/us. Thank you
for the detailed explanation.
> This bug should either be fixed by providing a pure ASCII dvorak
> variant, or you have to bind your right Alt key to mod1. I believe
> that selecting altwin:meta_alt option does the trick. In such a
> case, can this bugreport be closed, or do you really need another
Hmm... I am satisfied, but the next guy to come along using dvorak
might not be, so it's "good enough for me", but not necessarily "good
enough for Debian".
Let's leave it open, I don't want to take any more of your time, so I'll
try to create another variant myself and send it to this bug as a
patch. I should be able to get to it in the next day or two.
GPG fingerprint: D54D 1AEE B11C CE9B A02B C5DD 526F 01E8 564E E4B6
Welcome to the best software in the world today cafe!