[RULE] TinyX-4.3.0-6RULE (kdrive with xkb) Howto?

Richard Kweskin rkwesk at mail.ariadne-t.gr
Sun Aug 3 19:49:42 EEST 2003

On Sat, 2 Aug 2003 01:38:16 -0400
Michael Fratoni <mfratoni at tuxfan.homeip.net> wrote:

> Hash: SHA1
> On Friday 01 August 2003 03:22 pm, Richard Kweskin wrote:
> > Hello All
> >
> > Living in Greece, I want to try through xkb to get an additional layout
> > for Greek. Thus Michael recompiled the TinyX package with xkb.
> >
> > Slinky at version 0.3.96 is still using the smaller package,
> > TinyX-4.3.0-5RULE which hasn't got xkb.
> Correct, I'll update the package for the next release.
> Typo in your command, perhaps? You should have "setxkbmap" installed which 
> should do the trick. It's part of the XFree86 package.

Correct, as usual. Thank you

> You might also try running redhat-config-keyboard. The bonus here is that 
> when run from an xterm, redhat-config-keyboard prints out the commands it 
> issues. You can then copy them to a script and cause the script to be 
> executed when X is started to make the keyboard load automatically. Here 
> Good luck, hope that helps,
> - -- 
> - -Michael

Yes, here is the final outcome:

A script /usr/X11R6/bin/kb with the following:
	exec /usr/bin/X11/setxkbmap -rules xfree86 -model pc105 -layout us,el -option -option grp:alt_shift_toggle

when invoked in an xterm in icewm running on top of kdrive from TinyX-4.3.0-6RULE allows a toggle between the us and el (Greek) layouts, as the Greeks are already accustomed to doing in other setups.


The extra "-option" with no subparameters was necessary in order to run setxkbmap with *only* the options specified. While changing from one possibility to another it was a nightmare trying to figure out what was going on without it!

Attempts to place the exec line near or at the end of /usr/X11R6/bin/startx or (without "exec" but with a trailing "&") /home/rkwesk/.xinitrc did not work and complained "cannot open display "default display". Apparently the window manager has to be fully up first.

The "-layout us,el" part is valid with versions 4.3.0 or later. Earlier versions accepted only one alternative layout. Thus with Greek it was either "-layout el" or "-layout gr".

The "-option grp:alt_shift_toggle" enables the Alt+Shift, "-option grp:toggle" enables the AltGr-key not available on us keyboards but present on some others (e.g.a uk keyboard) "-option grp:shift_toggle" enables both shift keys simultaneously, "-option grp:ctrl_shift_toggle", "-option grp:ctrl_alt_toggle", "-option grp:caps_toggle" enable those combinations instead.

The "-option" parameter is singular and does not accept more than one option, but putting as many additional "-option" parameters is allowed. Thus "-option grp:shift_toggle,grp:led_scroll" was not accepted but "-option grp:shift_toggle" "-option grp:led_scroll" is.

On my setup "-option grp_led:scroll" or "-option grp_led:caps" or "-option grp_led:num" made no difference whether present or not, i.e. there is no other indication which layout is "active" other than trying it out :( - but it may just be the particular hardware.

The actual physical keyboard used only places *where* keys are. Anyone who wishes to type Greek (or any other) characters may do so using this procedure. :)

Small complaints: (certainly worth putting up with compared to what is gained)

Once invoked only the keys on the numberpad (numberlock off) allow arrow keys, etc. to be used and some of them have been altered, too.

The exact same script invoking setxkbmap worked without complaint while /etc/sysconfig/i18n (a configuration file affecting the locale both inside X and out) was as originally installed by RULE (i.e. Redhat's default.) But once i18n is altered to suit the use of Greek *the script works* but gives the standard error message "Couldn't interpret _XKB_RULES_NAMES property, use defaults rules -'xfree86' model -'pc101' -layout 'us'. It doesn't say it *is* using those defaults, only suggesting that they be used! Similarly, "small" errors don't prevent another character set from being typed but the script will exit giving the same error message! (e.g. counting carefully, my keyboard seems to have 104 keys. The "-model pc104" parameter always induces this while "-model pc 105" does not!

The credit for all of this revelation falls to Michael Fratoni (our list is continually grateful for his contributions) and to Athanasios Kyritsis (on non-kdrive setups) who is well known in the Greek linux community for his contributions.

Very gratefully, Richard

Rule Project HOME PAGE:  http://www.rule-project.org/en/
Rule Development Site:   http://savannah.gnu.org/projects/rule/
Rule-list at nongnu.org

More information about the Rule-list mailing list