SEO

Hi all,
I have the following problem:
When running a X -query session from a linux box to a HPUX 11 machine,
the keyboard behaves strangely, in that the numerical pad doesn't work
at all (as it does, for example, when running the same session toward a
Sun Solaris or another Linux): does anyone have the same problem and a
solution? on the Linux box, my kbd is set to pc/it, and as I said, works
correctly on other OS (except for a remapping of @ char).
Another problem I'm facing is that of a wrong colormap between the
systems (e.g, red becomes purple, green becomes cyan and so on..). I
noticed that the HPUX rgb.txt file is richer than linux's one, is it a
problem of missing mappings? Can I solve this problem replacing linux
file with HPUX's one?
Thanx for your help,
Stefano


--
Posted via Mailgate.ORG Server - http://www.Mailgate.ORG

On Wed, 14 Jul 2004 12:35:52 +0000, Stefano Cailotto wrote:

> Hi all,
> I have the following problem:
> When running a X -query session from a linux box to a HPUX 11 machine,
> the keyboard behaves strangely, in that the numerical pad doesn't work
> at all (as it does, for example, when running the same session toward a
> Sun Solaris or another Linux): does anyone have the same problem and a
> solution? on the Linux box, my kbd is set to pc/it, and as I said, works
> correctly on other OS (except for a remapping of @ char).

How exactly do you want the keypad to behavior? Linux provides the
XKEYBOARD extension with hooks in Xlib/Xext, but HP-UX does not implement
these. This can cause the keypad to behave oddly with certain clients,
especially "hpterm". Can you explain exactly what client you are talking
about and what behavior you get that is unexpected?

> Another problem I'm facing is that of a wrong colormap between the
> systems (e.g, red becomes purple, green becomes cyan and so on..). I
> noticed that the HPUX rgb.txt file is richer than linux's one, is it a
> problem of missing mappings? Can I solve this problem replacing linux
> file with HPUX's one?

Green <-> Cyan? Red <-> Purple? I don't know about these, but I do
know that there is a difference in the CDE "welcome" screen. This uses a
background of "CornFlowerBlue". XFree86 lists this as the RGB triple:

100 149 237 CornflowerBlue

versus the HP-UX triple of:

34 34 152 CornflowerBlue

This is a pretty big difference but I've never seen anything as drastic as
the color changes that you mentioned.

"Chuck Slivkoff" <slivkoff-AT-hp-DOT-com@localhost.invalid> wrote in
message newsan.2004.07.27.00.10.35.32071@localhost.inval id

> On Wed, 14 Jul 2004 12:35:52 +0000, Stefano Cailotto wrote:
>
> > Hi all,
> > I have the following problem:
> > When running a X -query session from a linux box to a HPUX 11 machine,
> > the keyboard behaves strangely, in that the numerical pad doesn't work
> > at all (as it does, for example, when running the same session toward a
> > Sun Solaris or another Linux): does anyone have the same problem and a
> > solution? on the Linux box, my kbd is set to pc/it, and as I said, works
> > correctly on other OS (except for a remapping of @ char).
>
> How exactly do you want the keypad to behavior? Linux provides the
> XKEYBOARD extension with hooks in Xlib/Xext, but HP-UX does not implement
> these. This can cause the keypad to behave oddly with certain clients,
> especially "hpterm". Can you explain exactly what client you are talking
> about and what behavior you get that is unexpected?
>
> > Another problem I'm facing is that of a wrong colormap between the
> > systems (e.g, red becomes purple, green becomes cyan and so on..). I
> > noticed that the HPUX rgb.txt file is richer than linux's one, is it a
> > problem of missing mappings? Can I solve this problem replacing linux
> > file with HPUX's one?
>
> Green <-> Cyan? Red <-> Purple? I don't know about these, but I do
> know that there is a difference in the CDE "welcome" screen. This uses a
> background of "CornFlowerBlue". XFree86 lists this as the RGB triple:
>
> 100 149 237 CornflowerBlue
>
> versus the HP-UX triple of:
>
> 34 34 152 CornflowerBlue
>
> This is a pretty big difference but I've never seen anything as drastic as
> the color changes that you mentioned.

Thanks for your answer, Chuck!
The "client" I'm talking about is a Linux box, running debian (or
redhat, no matter)
and I start the session with X -query <hpserver>. I saw many messages
regarding
kbd problems, so I don't know if it may be solvable...I discovered that
keeping the Shift key pressed makes the numpad working always correctly
(funny, no?!),
otherwise there's a totally unpredictable behavior, as the key works
only sometimes, w or w/o numlock: for example, if you activate numlock
most of times numpad works, if you deactivate numlock only the last
pressed
number works, if you reactivate NumLock it does/does not work (that's
why
I said unpredictable!)

For what is concerned colors, the problems arise especially with one
application (Marconi ServiceOn), but only from certain linux machines:
for example, from a Macintosh running Debian, everything works
correctly,
while from other two boxes (with Debian and LTSP Linux) I get the color
problem. I tried also changing
X server version, but without success. It might be a problem of the
application
itself, but it should happen for every machine!

One question: How should I do to recompile rgb colormap under HPUX (ie
to
get binary rgb from rgb.txt)?

Thanks a lot,
stefano


--
Posted via Mailgate.ORG Server - http://www.Mailgate.ORG

On Thu, 29 Jul 2004 09:08:33 +0000, Stefano Cailotto wrote:

> I saw many messages regarding kbd problems, so I don't know if it may be
> solvable...I discovered that keeping the Shift key pressed makes the
> numpad working always correctly (funny, no?!), otherwise there's a
> totally unpredictable behavior, as the key works only sometimes, w or
> w/o numlock: for example, if you activate numlock most of times numpad
> works, if you deactivate numlock only the last pressed number works, if
> you reactivate NumLock it does/does not work (that's why I said
> unpredictable!)

The "numlock" functionally is not well-defined under HP-UX. You should not
expect it to work as you would on a PC. You might, though try adding to
~/.dtprofile:

export XKB_DISABLE=1

This should let X11R6 clients from HP-UX behave a bit better when the X
server XKEYBOARD extension is detected. (See "xdpyinfo" output.) That is,
the numeric keypad will likely function either as numbers or as arrows,
but not both.

> For what is concerned colors, the problems arise especially with one
> application (Marconi ServiceOn), but only from certain linux machines:
> for example, from a Macintosh running Debian, everything works
> correctly, while from other two boxes (with Debian and LTSP Linux) I get
> the color problem. I tried also changing X server version, but without
> success. It might be a problem of the application itself, but it should
> happen for every machine!

In all cases, the X11 client is running on HP-UX, but displaying to
various Linux displays, correct? You might want to compare the outputs
from xdpyinfo for each display. This might be a case where you may have
the visual set to a depth 15 or 16, when it expects 8 or 24.

> One question: How should I do to recompile rgb colormap under HPUX (ie
> to get binary rgb from rgb.txt)?

This will only have effect for an X server running under HP-UX. Color
names requested by the client will be resolved by the display server
(Linux), not client.

rgb rgb < rgb.txt

Hope this helps,

-chuck

On Mon, 02 Aug 2004 20:00:30 +0000, Chuck Slivkoff wrote:

> The "numlock" functionally is not well-defined under HP-UX. You should not
> expect it to work as you would on a PC. You might, though try adding to
> ~/.dtprofile:
>
> export XKB_DISABLE=1
>
> This should let X11R6 clients from HP-UX behave a bit better when the X
> server XKEYBOARD extension is detected. (See "xdpyinfo" output.) That is,
> the numeric keypad will likely function either as numbers or as arrows,
> but not both.

You might also try disabling the extension on the Linux side using "-kb".