X11R4 problems
Darryl Okahata
darrylo at hpnmdla.HP.COM
Tue Feb 26 05:57:51 AEST 1991
Here is a document that describes how to get Herr Roell's X11R4
package running with Esix. I originally sent it out to the Esix-user's
mailing list, but I never posted it here.
-- Darryl Okahata
UUCP: {hplabs!, hpcea!, hpfcla!} hpnmd!darrylo
Internet: darrylo%hpnmd at relay.hp.com
DISCLAIMER: this message is the author's personal opinion and does not
constitute the support, opinion or policy of Hewlett-Packard or of the
little green men that have been following him all day.
===============================================================================
Here is a summary of my experiences with running Thomas Roell's
X11R4 version 1.1 with Esix. Some of this is documented elsewhere, or
has already been said, but it's nice to have a single document
describing everything.
First of all, the /usr/lib/X11/X386/install.sh script will not work
"as is" (it appears to be ISC-specific). You have to comment out the
tests for the "sp" (stream pipes) driver. *HOWEVER*, this does not mean
that you do not need the "sp" driver -- you do need it. With Esix, the
stream pipes driver is part of the "Network Servers" disk, and so you
must have the contents of this disk installed on your system before
X11R4 will work. Also, the install.sh script will not work properly
unless you have /usr/bin/X11 in your $PATH.
The Microsoft mouse driver in this X11R4 release seems to be
broken. As a workaround, if you have a Microsoft mouse (I do), use the
following mouse configuration line in /usr/lib/X11/X386/Xconfig:
MouseSystems 2 MS "/dev/tty00"
(The Mouse System's mouse apparently uses the same protocol as the
Microsoft Mouse.) Don't forget to edit the above line to reflect the
serial port of the mouse. Thanks to Chuck Murcko
(cmurcko at topsail.Topsail.ORG) for this fix. Microsoft mice also have a
problem where button presses are not recognized by the server until the
button is released or the mouse is moved. I don't have a fix for this.
Logitech mice (which most people seem to have) work fine.
Don't run X11R4 from the "console" console; use one of the other
nine virtual consoles. Running xinit from the console requires that a
virtual console be free (no getty or user session attached to that
console). If you, like me, have gettys running on all nine virtual
consoles, this means that you have to edit /etc/inittab (and the other
associated file below /etc/conf) to disable one of the virtual consoles.
The easiest way to avoid having to do all this work is to run xinit from
a virtual console.
Make sure that the X server /usr/bin/X11/X386 is linked to
/usr/bin/X11/X, and that it has setuid root permissions. Initially (the
first time you ever run it), the server needs to create a /dev/X
subdirectory with world-writable/readable permissions.
You'll get an error, "Creating TCP socket: Invalid argument". You
can ignore this.
Mike Knister (mknister at eecs.umich.edu) says that the cursor keys,
etc. do not work well, and that the solution is the following (I haven't
had the time to test it):
------------------------------------------------------------------------------
The following file helps with the cursor and delete keys. Put it into
a file like .xmodmap, and load (after xinit) it using:
xmodmap .xmodmap
! ---------xmodmap file to map life bearable under Esix:----------
! (for X386 1.1)
keycode 90 = Delete period
keycode 89 = Insert KP_0
keycode 86 = End KP_1
keycode 87 = Down KP_2
keycode 88 = Next KP_3
keycode 82 = Left KP_4
keycode 84 = Right KP_6
keycode 78 = Begin KP_7
keycode 79 = Up KP_8
keycode 80 = Prior KP_9
! Backspace key: I personally like it to be delete
keycode 21 = Delete
!-------------end of xmodmap file -----------------------
-------------------------------------------------------------------------------
Here are some useful screen sizes for people with the NEC Multisync
II (these were taken from Chin Fang's (fangchin at portia.Stanford.EDU)
postings in comp.unix.sysv386). I have a Multisync II and have found
that the Xconfig timings for the 800x600 resolution cause a fair amount
of flicker, not to mention poor resolution due to the large color mask
size. While not all people can use the following, due to differences in
clock frequencies, they are much nicer than using a 640X480 resolution.
-------------------------------------------------------------------------------
name clock horizontal timing vertical timing flag
# 704x528 -> ~66Hz vert freq
"704x528" 32 704 736 856 888 528 528 533 545
# 752x564 Multisync II - only -> ~64 Hz
"752x564" 36 752 788 916 952 564 564 569 594
# "No-flicker" VGA-resolution -> ~68 Hz
"640x480" 28 640 676 776 812 480 480 485 505
-------------------------------------------------------------------------------
I've only tested the "704x528" resolution.
-- Darryl Okahata
UUCP: {hplabs!, hpcea!, hpfcla!} hpnmd!darrylo
Internet: darrylo%hpnmd at relay.hp.com
DISCLAIMER: this message is the author's personal opinion and does not
constitute the support, opinion or policy of Hewlett-Packard or of the
little green men that have been following him all day.
More information about the Comp.unix.sysv386
mailing list