[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#25689: xbase: Delete mapping in Xresources wrong if connecting from remote



In article <199808170829.KAA11797@patty.verwaltung.Uni-Mainz.DE> you write:
>Mr Brian May writes:
> > In article <13778.61988.699473.476467@patty.verwaltung.Uni-Mainz.DE> you write:
> > >Package: xbase
> > >Version: 3.3.2.2-4
> > >
> > >If you connect from a non Debian Xserver via xdm to a Debian box, the
> > >following lines in /etc/X11/Xresources brake the usage of Delete.
> > 
> > I assume you start X on the non-Debian server using something like:
> > X -query debian-server-name
> > ?
> >
>
>Similar, yes: it is a local Xserver (e.G. HCL-EXceed on WinNT)
>listening for xdmcp broadcasts.

> > Does the same problem occur when logging directly into the Debian
> > server?
>No. 

> > echo $TERM
>
>is xterm-debian
>
> > stty -a
>
>erase = ^?
>
> > cat
> > (push delete several times; what happens?)
>#cat
>^[[3~^[[3~^[[3~
>
> > 
> > What happens if you change the given lines to:
> > 
> > *VT100.Translations: #override <Key>Home: string("\033OH")\n\
> >                                <Key>End: string("\033OF")
>
>This was my fix. Than it is working. But I haven't checked this if I
>log in directly int the Debian server. (Will report on this later.)

I think I know what your problem is. I think some X-Servers (including
this Ultrix 4.4 server I am currently using ;-) ), generate the
"Delete" keycode instead "Backspace" (I think). You can verify this
using a program like xkeycaps. 

I think the "Delete" keycode is meant to be delete to the right, and
the "Backspace" keycode is meant to be delete to the left.
I don't have access to a Debian system at the moment, so I can't
test this.  I wonder if this behaviour is standardized by X? If so,
it would mean the server is non-compliant.

I suspect your fix will break the behaviour of "Delete" under the
Debian server :-(

Solutions? Perhaps you can edit your .xsession to change your xresources
entries using xrdb depending on the value of "$DISPLAY"? Not
a very good solution though.

If anyone has any better ideas...


Reply to: