home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Software Vault: The Gold Collection
/
Software Vault - The Gold Collection (American Databankers) (1993).ISO
/
cdr55
/
lxfaq160.zip
/
LINUXFAQ.06
< prev
Wrap
Internet Message Format
|
1993-06-09
|
62KB
From getrz!chsun!mcsun!news.funet.fi!hydra!klaava!wirzeniu Wed Jun 9 14:02:53 1993
Xref: getrz comp.os.linux:16159 news.answers:3008
Newsgroups: comp.os.linux,comp.os.linux.announce,comp.answers,news.answers
Path: getrz!chsun!mcsun!news.funet.fi!hydra!klaava!wirzeniu
From: corsini@victor.greco-prog.fr
Subject: Linux Frequently Asked Questions 5/5 [monthly posted]
Message-ID: <1993Jun8.144528.18868@klaava.Helsinki.FI>
Followup-To: poster
Summary: Linux, a small and free unix-like for 386-AT computers.
Sender: wirzeniu@klaava.Helsinki.FI (Lars Wirzenius)
Reply-To: linux@numero6.greco-prog.fr
Organization: Greco Prog. CNRS & LaBRI, Bordeaux France
Date: Tue, 8 Jun 1993 14:45:28 GMT
Approved: linux-announce@tc.cornell.edu (Lars Wirzenius)
Expires: Wed, 21 Jul 1993 13:49:58 GMT
Lines: 1661
Archive-name: linux-faq/part5
Last-Modified: 93/06/07
Version: 1.18
*********************************************************
* *
* Answers to Frequently asked questions about Linux *
* *
*********************************************************
This post contains Part 5 of the Linux FAQ (5 parts).
It must be read *after* the 4 first parts.
===================================8<====>8============================
CONTENTS (of this part)
XII. X11, THE MAXIMUM and MORE (part5)
XIII. NETWORKING and LINUX (part5)
XIV. EVERYTHING FOR PRINTING (part5)
===================================8<====>8============================
XII. X11 THE MAXIMUM and MORE
=============================
*** This section is maintained by Krishna Balasubramanian
*** <balasub@cis.ohio-state.edu>. Mail him if you have corrections,
*** additions, etc.
*** Last update: Thu, 13 May 93 00:35:45 -0400
XII.A. X386 GENERAL INFORMATION
XII.B. HARDWARE REQUIREMENTS: Supported Video cards, mice.
XII.C. LINUX DISTRIBUTION: Files required, Current version.
XII.D. LEARNING/USING X: Pointers to X documentation.
XII.E. DEBUGGING STARTUP PROBLEMS: Checklist, Screen restoration, Hanging.
XII.F. XCONFIG: Video mode settings and common errors in Xconfig.
XII.G. X-APPLICATIONS: Compiling X programs.
XII.H. ATI: SVGA server for ATI boards.
XII.I. BUGS
The X11 directories on linux systems are:
XLIB = /usr/X386/lib/X11/ (or /usr/lib/X11/)
XBIN = /usr/X386/bin/ (or /usr/bin/X11/)
XDOC = XLIB/etc/
cwxi = the comp.windows.x.i386unix newsgroup
Subscribe to this group if you are an xfree86 user. Post general
questions on xfree86 to cwxi instead of c.o.l. Very few problems
with using xfree86 are Linux specific.
XII.A. X386 GENERAL INFORMATION
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
XII.01) What is the X11 release supported by Linux?
ANSWER: It's the X11R5 (xfree86-1.2).
There are (currently separate) servers for 8514 and S3 chips.
xfree86-1.3 should be available in a few weeks. Major changes are
Improvement of video restoration, expanded support for WD chips
and the inclusion of PEX. You should get newer versions of any
applications that are older than the xfree86-1.2 release.
XII.02) What is X386/xfree86?
ANSWER: X386 is the port of the X11 server to System V/386 that was
done by Thomas Roell (roell@informatik.tu-muenchen.de). It supports a
wide variety of VGA boards. X386 1.2 is included in MIT's X11R5
distribution.
The Linux X386 port was based on the stock distribution from X11R5,
from MIT and was done by Orest Zborowski (obz@sisd.kodak.com).
It has since moved to becoming part of the standard xfree86 distribution.
See the FAQ on cwxi for more information on xfree86.
XII.03) Where can I get X386 1.2 (X11R5)?
ANSWER: The X386 1.2 and xfree86 sources are available at any site that
distributes the X11R5 source (too numerous to list here, but includes
export.lcs.mit.edu)
XII.04) Any tips on compiling X11R5?
ANSWER:
- Dont do it.
- XFree86 is distributed with a link kit so you can optionally
include what you like in the server.
- Join the xfree86 beta team (how to? see cwxi FAQ)
- Instructions are in the README file in XDOC/ and the cwxi FAQ.
XII.B HARDWARE REQUIREMENTS:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Approx: at least 4 megs of ram + swap ...slooooww. 10 Meg disk for X.
Another 6-10 meg of disk for GCC if you want to compile X11 programs.
XII.05) What VGA boards are supported?
ANSWER:
et3000, et4000, gvga, pvga1a, wd890c00, tvga8900, ati ver. 5 or 6,
8514/A. (X386mono supports generic vga's and hercules).
Diamond cards are not supported and will not be supported. If
you are the unfortunate owner of such a card, you can probably
get the server up by booting in specific modes or using dos to
set your modes before warm booting into linux or using an
external clock setting program. You will have to bear with these
irks until you can convince diamond to alter their policy.
Standard x11v1.1 or xfree86 server:
ET3000 (for ex. GENOA 5300/5400)
ET4000 (Tricom, STB PWR Graph, Sigma Legend, etc.)
GVGA (Genoa 6400)
PVGA1A (Paradise VGA Professional)
WD90C00 (Paradise VGA 1024)
supported by xfree86:
TVGA TRIDENT 8900c, 9000, support is in xfree86
ATI See the ATI section below.
Those with 8514 compatible cards may want to get
the X8514 server for speed (~2x xstones?).
MONO Any vga card should be able to use X386mono server. (At least
640x480 with 800x600 virtual). Use the vga2 section of Xconfig.
Support for hercules monochrome card (usable as 2nd display).
The following servers will usually not handle all the options supported by
xfree86. Read the documentation that comes with them carefully. They
will be merged with the xfree86 distribution in some time.
8514 ATI graphics ULTRA, ATI graphics Vantage
Should work with any VESA standard 8514/A register compatible
card? Courtesy Kevin Martin (martin@cs.unc.edu).
Scott Laird (lair@midway.uchicago.edu) writes:
I uploaded a new version of the X8514 X Server to sunsite and
tsx-11. It is in /pub/Linux/X11/X-servers/X8514/X8514scale.tar.Z
on sunsite. It's linked with version 4.2 of the jumptable libraries,
includes TCP/IP support, support for compressed bitmap fonts, Type 1
and Speedo scalable fonts. There's a README file in the same
directory that will answer more questions.
XS3 S3 chipset server (Jon Tombs jon@robots.ox.ac.uk)
Get the FAQ on ftp.robots.ox.ac.uk (pub/linux/S3 check sunsite).
Xega Generic 640x480x16 compatible server (originally for laptops).
This requires a microsoft mouse at /dev/mouse for now
and it does not use Xconfig so use environment variables
to define the font path etc. in .xinitrc:
export FONT_PATH=/usr/lib/X11/fonts/misc:/usr/lib/X11/fonts75dpi
Works better with courier fonts so add to .Xresources:
*Font: -*-courier-medium-r-*--10* ..or whatever..
A link kit is available at tsx-11 (you need gcc2.2.2).
in pub/linux/ALPHA/Xega/X386.ega.T.Z
(Obselete? Unusable? Anyone?)
WARNING: Do not try to bring up an Xserver that does not support your
hardware. There have been cases where damage has resulted from pushing
the monitor (specially fixed frequency monitors) beyond its specs.
XII.06) What Mouses are supported?
ANSWER: Serial : Logitech, Microsoft, MouseSystems .... compatibles ...
Busmouses : Logitech, microsoft, ATI_XL, PS/2 (aux).
XII.07) Does anyone have a working PS/2 mouse? Has anyone gotten the
"Mini-DIN" mouse on an HP Vectra 486/33T to work? The slight info I've
been able to find says it's PS/2 compatible. Does anyone have a
working MouseMan on a PS/2 port ?
ANSWER: (heeb@watson.ibm.com)
First you need to create an entry in /dev for it:
mknod /dev/psaux c 10 1
The other steps depend on the version of XFree: for XFree86-1.2
add the following line to your Xconfig (and you are done):
PS/2 "/dev/psaux"
This works even for 3 button mice (e.g. the MouseMan)!
There is no direct PS/2 mouse support in XFree86-1.1, so to use the
mouse with X you'll have either to upgrade or use the mconv mouse
protocol conversion utility, which can be found on nic.funet.fi, in
/pub/OS/Linux/utils/tools/mconv.c. This program converts the packets
sent by the PS/2 mouse into the corresponding ones from a Microsoft
mouse, so you can fool X telling it you have a Microsoft serial
mouse instead. Instructions for use are included in the source file.
(Johan Myreen jem@cs.hut.fi)
XII.08) I have trouble with my logitech Pilot mouse and X under
Linux, any clue ?
ANSWER: (Thomas Roell?) There are TWO lines of Logitech mice out there.
One is the programmable and uses MouseSystems protocol at startup.
X386 reprogramms them to use another protocol. If you specify
'Logitech' in the Xconfig, X386 assumes a mouse like C7 or S9 (notC7-M).
The second line is the MicroSoft compatible. Currently all newer
Logitech mice follows this practice, like the MouseMan. In that case
you have to say 'MicroSoft' or 'MouseMan'.
From: jliddle@rs6000.cmp.ilstu.edu (Jean Liddle)
I just purchased a new logitech mouseman, and yes, the new X-Windows
(XFree86 as opposed to X386) requires that you use the "microsoft" mousetype.
However, if you turn Third Button Emulation off, the middle button WILL work.
XII.C. LINUX DISTRIBUTION
~~~~~~~~~~~~~~~~~~~~~~~~~
XII.09) What is the current version?
ANSWER: The newest public Linux release is xfree86-1.2.
It was compiled with gcc-2.3 (jump tables 4.3) and is available
for example at tsx-11 in pub/linux/packages/X11/XFree86-1.2/.
XII.10) What Files do I need to download?
ANSWER: For xfree86: There is a README file in the distribution directory.
Bring this down first and read it. Then get xbin, xman and xfonts.
You need xprog and xlibman only if you plan to compile X applications.
You need xkit if you want to relink the server.
**** You need The shared images libc.so.4 (4.3 or later) and libm.so.4 (4.0)
in /lib/. All the X11 binaries use these. These are available with the GCC
distribution or any new rootdisk. To compile programs you should
get gcc-2.3 or later and the 4.3 libraries.
To get a different server, see the notes in the previous section.
X11 directories in the archive sites are:
tsx-11 : pub/linux/packages/X11, pub/linux/usr.bin.X11
sunsite : pub/Linux/X11/
XII.11) Where do the X11 files go? What are they?
ANSWER: The linux X binary distribution looks something like:
/usr/X386/
lib/
libX11.a ... (X libraries needed for compiling stuff).
X11/
config/ (template files for compiling ... xmkmf)
fonts/
misc/ , 75dpi/ ....
etc/ (documentation).
bin/ (X386, X386mono, xterm, X linked to X386).
include/X11/ (include files needed for compiling).
bitmaps/ (icons, bitmaps used by some applications).
man/
man1/, man3/, cat1/, cat3/
/lib
libX??.so.?? (Shared images needed by all X11 binaries).
Its best you do a tar -tvzf on the distribution you get or find
some means of extracting the distribution file list.
The support files include default/example Xconfig, xinitrc, twmrc
which you could copy to your HOME directory (as Xconfig, .xinitrc
and .twmrc) and edit them to define your hardware and X11 setup.
The directory XLIB/etc/ contains much documentation on how to
configure video modes.
Read the files XDOC/README and XDOC/README.Linux.
XII.D. LEARNING/USING X
~~~~~~~~~~~~~~~~~~~~~~~~
XII.12) Where can I find the basic help for learning/using X ?
ANSWER: Try man X386; man Xserver; man xterm; man twm; man xinit
on any machine with X11. The man pages on xterm and twm will reflect
behaviour on Linux systems quite accurately. The others will give you
some bearing though not every option will work. These man pages are
part of the xfree86 distribution.
XII.13) What docs are available besides man pages?
ANSWER: The FAQ in comp.windows.x is extensive and has a bibliography.
There is also the Xt-FAQ. Both are available from export.lcs.mit.edu in
pub/contrib/. There is a collection of (mit) X11 documents on tsx-11
(pub/linux/doc/xdoc.tar.Z ?) or export (pub/R5-Untarred) if you are
interested in internals.
Also of interest if you use motif, open-look/xview or interviews:
comp.windows.x.motif, comp.windows.open-look, comp.windows.interviews.
XII.14) How do I start up X?
ANSWER: Type 'startx' at the shell prompt. Before doing this you should:
0) Untar the xfree86 distribution from / (read XDOC/README, XDOC/README.Linux).
1) copy the file Xconfig.sample from the XLIB/ directory to your
HOME directory (/usr/root/Xconfig?) or to XLIB/Xconfig.
2) Edit Xconfig and set the video data for your card and monitor.
Use the clocks appropriate for your card and the modes appropriate
for your monitor for those clock values. To determine what values are
appropriate the Xconfig section below and the files in XDOC/.
3) Set the mouse device in Xconfig correctly....Use the busmouse entry
for Logitech only. For other busmouses use microsoft (try all :-).
4) Ensure that there is a termcap entry for xterm in /etc/termcap
(One is available in the subdir /doc/).
5) Ensure that the X11 binaries are in the path example:
add the line PATH=${PATH}:/usr/X386/bin to ~/.profile or
look at bash.ad in /doc/.
WARNING: NONE of the modes (in the Modes line in Xconfig) should use a
clock your monitor cannot handle.
Now type startx.
XII.15) How do I configure X .. colors, menu, keyboard?
ANSWER: makes me ... dizzy ...
server : video, keyboard
Xconfig : man X386, man Xserver, XDOC/VideoModes.doc
man X386keybd, man xset, man xmodmap.
resources: global and application preferences.
XLIB/xinit/.Xresources or ~/.Xresources : man xrdb?
twm : menus, title bars, colors .. look and feel.
~/.twmrc or XLIB/twm/system.twmrc : man twm
xinit : startup.
~/.xinitrc or XLIB/xinit/xinitrc -- man xinit
applications : XLIB/app-defaults/Xxx (also ~/.Xresources)
man xxx
You can start with files from any X11 setup. Backup the files
you change. See the FAQ on comp.windows.x.
XII.16) [suggested by Thomas Koenig]
How do I support national keyboards in X11 ?
ANSWER: Put a .Xmodmap file into the user's home directory specifying
which keys to use for what; or else replace the
/usr/lib/X11/xinit/.Xmodmap with the national keyboard keys
customized. Do man xmodmap for more details.
enclosed an excerpt of mine (french keyboard):
keycode 8 =
keycode 9 = Escape
keycode 10 = ampersand 1
keycode 11 = eacute 2 asciitilde
keycode 12 = quotedbl 3 numbersign
keycode 13 = apostrophe 4 braceleft
keycode 14 = parenleft 5 bracketleft
keycode 15 = minus 6 bar
keycode 16 = egrave 7 grave
keycode 17 = underscore 8 backslash
keycode 18 = ccedilla 9 asciicircum
keycode 19 = agrave 0 at
keycode 20 = parenright degree bracketright
keycode 21 = equal plus braceright
keycode 22 = BackSpace
keycode 23 = Tab
keycode 24 = A
........
XII.E. DEBUGGING STARTUP PROBLEMS:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
XII.17) I cannot type in my xterm
ANSWER: /dev/console must have major=4, minor=0
rm -f /dev/console; ln /dev/console /dev/tty0
XII.18) What are some of the common omissions and errors?
ANSWER: (Peter)
1) Add /usr/X386/bin/ (or /usr/bin/X11) to your path.
2) Add the xterm termcap entry to /etc/termcap.
3) ln -s X386 X in the XBIN directory.
ln -s X386mono X (if you want to use the mono server).
4) Run X as root first. This avoids some trivial problems.
5) If you change the font path entry .... X386 is finicky about
the syntax.
6) Make sure your mouse entry in Xconfig points to a legitimate serial
device (or busmouse)... usually something like: microsoft "/dev/ttyS1"
serial devices have major # 4 minor 64,65 ... busmouses are major 10.
ls -l in /dev will show you the major and minor entries.
7) Need read/write/exec access to /tmp directory
8) Use startx 2> x.err to log your error messages. If the server sticks
and you have to reboot you'll probably loose this stuff... see notes
on `hanging' below.
9) You must have a free vt (eg. /dev/tty5 with no getty (see /etc/rc)).
10) Delete /tmp/.Xn-lock if recovering from a crash.
11) Note the new probeonly option for the server (man X386):
startx -- -pn -probeonly
12) See Xconfig problems and the device list below.
XII.19) Why is the server unable to find some of the fonts?
ANSWER: First check that the directories listed in the font path exist
and have font files in them. Some of the servers are not set up to use
compressed fonts. In such cases you have to uncompress the fonts in the
directory and run `mkfontdir .`. Read the man page on mkfontdir.
XII.20) My server "hangs" Why?
ANSWER: (Peter) Quite often it isn't a hang. If the server grabs the
screen(and enters graphics mode) and then dies, it may return you to
bash *without* restoring the screen. In other words, just because you
see no output, don't assume it's running/hanging. The way to test
this is to see if you can use a bash command to spin your disk. eg.:
sync; ls -l /bin.
Also - I *suspect* that (at least for me), you can kill the
server if you include the servernum option in the Xconfig (Obselete?)
You can kill the server when you want to by typing ctrl-alt-backspace.
Then a couple of ctrl-C 's in case your stuck in xinit.
The two line message
X386 version .....
(protocol Version ....
is returned from the X386 programme itself, so if you get
this (or if the screen blanks) X must be starting.
XII.21) When I started X11 I got "Cannot connect to
server" or "process does not exist", any clue ?
ANSWER: The cause might be a wrong fontpath variable, mouse device
or video settings in your Xconfig.
XII.22) Why cant I run more than 4 xterms?
How can I have more than 4 pty's ?
ANSWER: set the number in the header include/linux/tty.h and
recompile the kernel. Also make nodes :
mknod /dev/ptypxx 4 minor (where minor = 128 + xx)
mknod /dev/ttypxx 4 minor (minor = 192 + xx)
This is for before linux-0.97pl6.... for later versions
you only need to make the nodes.
XII.23) How does X11 start up?
ANSWER: The startup involves the server (XBIN/X), and some
programs like startx and xinit.
Typing startx runs the script XBIN/startx. Look through startx.
For explanation of the server arguments try man Xserver.
startx does little other than gather arguments and then call xinit.
xinit runs the server X (X is linked to X386) and the client programs
specified in ~/.xinitrc. xinit continues to run, and is often the
source of the error messages you see when you exit. If you get the
message "giving up", it means X has died (quite probably it died
immediately) and xinit has been unable to start clients from the
xinitrc script. (The scripts must be executable).
XII.24) What devices does X depend on:
ANSWER: check your device numbers with ls -l /dev. You should have:
5 0 tty
4 0 tty0
4 0 console
4 1 tty1 (etc) virtual consoles (vt's) : eg. startx -- vt8
1 1 mem
4 128 ptyp0 pseudo ttys used by xterm/emacs etc
4 129 ptyp1 (etc) to talk to unix programs.
4 192 ttyp0
4 193 ttyp1 (etc)
4 64 ttyS1 one of these is the mouse
4 65 ttyS2 or you have a busmouse.
10 x mouse busmouse x = 0 => logitech 1 => ps/2 (aux)
2 => microsoft 3 => ATI_XL.
/tmp/.X11-unix/Xn n = display number. Socket used by X11 programs to
talk to the server. This is created by X386.
/tmp/.Xn-lock n = display number. Lock file.
Note that unix programs dont expect standard names for serial devices
and mice. So you can give these devices any name as long as you tell
the server what its called (in Xconfig).
XII.F. Xconfig: Xconfig and Video mode settings.
~~~~~~~~~~~~~~~
XII.25) What do the mode names in Xconfig mean?
ANSWER: Mode names like "640x480" are used to match entries in
the modes line with the modeDB entry which actually defines the mode.
You can toggle between modes in the modes line with ctrl-alt-numericplus.
If you're experimenting with 640x480 you can define
modes "640x480a" "641x480b" "642x480c"
with different entries for each under ModeDB so you can toggle between
these three modes.
XII.26) How do I compute the numbers for the video modes?
ANSWER: See if there is an entry that matches your chip/monitor in
XDOC/modeDB.txt. Use the clock entries for your chip
and the mode entry for that clock appropriate for your monitor.
If you cant find a mode entry for a particular clock 'needed_clock'
but can find a mode entry (for your monitor) for a nearby clock value
'found_clock', scale all numbers in the modeDB entry by
(needed_clock/found_clock). Read XDOC/VideoModes.doc
XII.27) How can I determine the clocks on the video chip?
Where can I find clock.exe or clock.pas ?
ANSWER: The XFREE Team discourages the use of clocks.exe and wants it
to be removed from all docs, since the new Server is better at
guessing the clocks than clocks.exe.
You can get X to tell you what it finds by commenting the clocks entry
in Xconfig. Use 'startx -- -probeonly 2> x.err' to log the output.
If X starts up this file gets the clocks only after you exit.
I strongly advise using the 'modegen' spreadsheets to generate the
appropriate settings. If I remember correctly you can find them in
'pub/linux/packages/SLS/x4' called modegen.taz or something equally
as obvious ... Good luck. (Andy asb@cs.nott.ac.uk)
The clocks entry in Xconfig serves as an index. example:
clocks 1 2
ModeDB
1 640x480 .....
2 800x600 .....
means the first clock on the chip will be used for the 640x480 mode
and so on. Usually one uses the MHz value for the clock for convenience
instead of 1 2 you probably have 25 28 ... etc.
XII.28) What are the settings used with a trident-8900C?
ANSWER: To get X/SLS/Linux working with a trident-8900C :
1) From Frank Houston fh8n@uvacs.cs.Virginia.EDU
vga256
Virtual 1024 768
ViewPort 0 0
Modes "1024x768" "800x600" "640x480"
Use the VESA monitor timings from the modeDB in the example Xconfig.
You do not need to specify the chipset or clocks.
2) From: Henk Vandecasteele henkv@cs.kuleuven.ac.be
TRIDENT TVGA 8900C card with BIOS C2.11, A cheap color monitor AOC (CM-335)
with a maximum horizontal frequence of 39.5
vga256
Virtual 1024 1024
ViewPort 0 0
Modes "912x684" "800x600" "640x480"
# This mode drives my monitor to the limit (VESA timings for the rest).
"912x684" 45 912 944 1104 1136 684 684 690 712
XII.29) Oh what, Oh what, Oh what can I do with a trident 8900b?
ANSWER: (drew@cs.colorado.edu)
1. While the 8900b works with the Xfree 8900c driver, you must
specify a chipset line for the 8900c in your Xconfig.
2. The tridents are miserably slow in a Color mode. If you don't
mind black and white, you should use X386mono. Not only is it
faster, but it requires less video memory, and you'll be able
to run a 1k X 1k virtual desktop. Again, you must specify the
Chipset line to use the banked mono server for virtual
resolutions > ~800x600.
XII.30) What are common problems with Xconfig?
ANSWER: If X isn't firing at all, or even if it clears the screen and
crashes (see below) you almost certainly have a fault in Xconfig.
1. Note that the server X interprets Xconfig. (not xinit)
2. test with only one graphics mode included - one you are most sure of.
3. If you make a mistake in the mouse section, you will (I understand)
get an error message, so if you don't get some mouse error, try
elsewhere first. (eg "No such file or directory" indicates you have a
fault in the "/dev/ttyxx" line, or /dev/ttyxx doesn't exist.
4. *NB* X looks in your home directory first to find an Xconfig. make
certain it's using the one you've been working on!
5. you need double quotes (") around your chipset and modes: eg
chipset "et4000"
Modes "640x480" "640x480a" "800x600"
and around the /dev/ttyxx eg.:
microsoft "/dev/ttys1"
6. Check your VGA section and modeDB first.
9. ensure ModeDB clock speeds match the VGA section. Comment out any
unused lines in MODEDB.
eg clocks 25 36 00 00 00 00 00 00
then you *must* have (and only have) lines under modeDB corresponding
to speeds 25 and 36. If you have one with (say) 62 uncommented out in
the modeDB section in place of the 36 line, you will cause the server to
die immediately with no error message.
10. videoram must equal the amount of display ram on your chip.
11. The virtual resolution screen must fit in the videoram.
eg: 512kB => 800x600 (for the 1 byte per pixel color server)
XII.31) What do I use in my Xconfig file to use the bus mouse?
Logitech doesn't work with my Logitech busmouse.
ANSWER: (Nathan I. Laredo)
Here are the mouse lines from my Xconfig:
BusMouse "/dev/mouse"
You may add a sample rate if you want, but chances are, unless the model
number on your mouse matches mine: PC-93-9MD it probably will cause your
machine to lock up, as I've gotten several reports saying that.
According to the X386 documentation, non-logitech mice do not support
a sample rate.
XII.32) Where can I get Xconfig-files for various hardware?
ANSWER: If you have access to e-mail, e-mail bcr@physics.purdue.edu
a message with the subject: help
If you do create a new Xconfig file or improve on an old one please mail
it to bcr@physics.purdue.edu to save others from having to duplicate
your work.
XII.G X-APPLICATIONS: Compiling X Programs.
~~~~~~~~~~~~~~~~~~~~~
XII.33) Does anyone know where I can find the libobz.a library ?
ANSWER: This library contained the socket stuff, when X386 V0.1 was
released. Meanwhile it has found it's way into the kernel, so
remove -lobz from the link step in the Makefile. Any program
asking for libobz needs to be recompiled anyway.
XII.34) I get _setjmp undefined when linking xv from .a files.
ANSWER: You are trying to link a program compiled with gcc-2.2.2
and setjmp has changed sice then. Recompile xv from the source
using the latest gcc.
XII.35) How do I compile an X application that has an Imakefile?
ANSWER:
Type xmkmf to generate the Makefile.
Type make.
XII.36) How do I compile a program that has no Imakefile?
ANSWER: Use compiler options generated with programs that do have
Imakefiles. Look in XLIB/config/ for the linux definitions.
In particular you need -fwritable-strings.
XII.37) gcc complains the X libraries are not found or links static.
ANSWER: Check out the -nojump -static options of GCC ... read the GCC FAQ.
You can tell gcc what directories to search with -L.
XII.38) When using Makefiles generated by xmkmf why do I get many error
messages?
ANSWER: The old xmkmf under Linux doesn't like # comments in Imakefiles.
Therefore you must change the #'s to XCOMM. But be careful. Don't change
any pre-compiler directives. (mal11@po.CWRU.Edu Matthew A. Lewis)
XII.39) (Ton van Rosmalen wrote):
I recently compiled XFree86 1.2A for Linux and it compiled fine with
gcc-2.3.3 and libc-4.3.3 iff I skipped the make depend.
ANSWER: (Rik Faith)
The source for ed.c is available from (Linux binaries are in utilb13.taz)
ftp.cs.unc.edu:/pub/faith/linux/utils/utilb-1.3.src.tar.Z
This version of ed.c has a 4096 character line limit and will work
correctly when building XFree86 under Linux.
XII.40) What are the development tools available for X under Linux?
ANSWER: Some of the X development packages available on Linux:
(XS = sunsite.unc.edu pub/Linux/X11)
(Contact the people who distribute them in case of problems.)
Libraries:
Xpm, tiff .. are in XS/libs/. If you are compiling a package
that uses these libraries it is best to pick up the standard DLLs.
hooft@chem.ruu.nl, mitchum.dsouza@mrc-apu.cam.ac.uk
(Some older libraries are in XS/libs/oldlibs).
Interviews: SLS package
Shared libraries are not yet available so doc ~1M, idraw ~1.5M, ibuild?
KHOROS:
This is the second release of my patches (~15K XS/xapps) to the Khoros
system for image processing. If you want to compile it, you
need Khoros 1.0, patchlevel 5 (source distribution) and up to 100MB
free disk space (including package 0.).
Wolfram Gloger wmglo@dent.med.uni-muenchen.de Apr 04 1993.
MOTIF XS/$$$:-)
Metro Link, Inc. is pleased to announce the availability of OSF/Motif 1.2.2
for Linux at the introductory price of $199 for a complete runtime and
development system (shared libs, UIL, source for some demos + 1 Orielly book).
You can order OSF/Motif 1.2.2 for Linux by calling Metro Link,Inc. at
(305) 970-7353 (voice) or (305) 970-7351 (fax) or by mailing us at
sales@metrolink.com.
SUIT: XS/devel
Simple User Interface Toolkit.
Included are shared libs and static libs. Compiled with the SLS kit.
To use these shared libs you need libX11 3.0 and libc 4.3.
Rob Robert-Jan Kooper rob@is.twi.tudelft.nl kooper@dutiws.twi.tudelft.nl
tk+tcl: XS/devel
Here is tk3.0 and tcl6.5 sources and libraries. Compiled this with gcc2.3.3
I have also included the wish (X11) binary.
Mitch mitchum.dsouza@mrc-apu.cam.ac.uk mitch%markab@mrc-apu.cam.ac.uk
xview: XS/devel
New libxview.a and libolgx.a that correct some problems with the
ones found in SLS (and maybe in the other xview releases, I don't know).
Kenneth Osterberg ??
XII.H ATI: SVGA server for ATI boards.
~~~~~~~~~
Last Revised: Mon May 10 15:17:57 1993 by faith@cs.unc.edu
XII.41) What's new?
ANSWER:
For XFree86 version 1.2 (and later versions), there are two changes to the
extregPlusXLAndOrMasks table:
1) ER_B0 was changed from 0x28 to 0x31. (In the Linux world, this was
distributed as a binary patch long ago.)
2) ER_BE was changed from 0x08, but only for cards EARLIER than the
28800-5.
Both of these changes allow more 28800-4 cards to work, although either one
may be sufficient for any one card. You can find these changes in the
driver by searching for the Feb 3 timestamps.
XII.42) What is the ATI SVGA driver?
ANSWER:
The ATI SVGA driver is a 256 color driver for the XFree86 server. The
driver was written for the ATI WONDER series of graphics adapters. In
general, the following modes are supported for cards with 512kb of memory:
640x480x256
600x800x256
The following modes are supported for cards with 1024kb of memory:
1024x768x256
XII.43) What is the ATI SVGA driver *NOT*?
ANSWER:
The ATI SVGA driver is *NOT* a black and white (2 bit) driver. Therefore,
it will *NOT* work with the XFree86 monoserver.
The ATI SVGA driver is *NOT* a 16 color driver. Even though your manual
says that your graphics adapter has a 1024x768x16 mode, the ATI SVGA driver
will *NOT* use this mode. If you want to use 1024x768, you *MUST* have 1MB
of memory on your board and use the 1024x768x256 mode.
The ATI SVGA driver will *NOT* support more than 8 bits of pixel depth.
Therefore, even though your manual says that your graphics adapter supports
modes using more than 256 colors, the ATI SVGA driver will *NOT* use these
modes.
The ATI SVGA driver is *NOT* an accelerated driver. If you have an ATI
GRAPHICS ULTRA series card, the ATI SVGA driver will *NOT* use the
accelerated hardware. It will only use the SVGA hardware. This will
probably not be any faster than a VGA WONDER card, unless the card is on a
local bus (and then, it won't be faster than a VGA WONDER card on a local
bus, if a card of this type existed).
XII.44) What cards will the driver work with?
ANSWER:
This is a difficult question, because there are many different ATI video
cards, and it is sometimes difficult to tell them apart. The best way is
to look on the card itself for the ATI chip numbers. Here is a partial
list of ATI VGA WONDER cards. The MOST IMPORTANT number is the CHIP
VERSION.
Adapter version Dot clock Chip version
VGA WONDER V3 ATI18800 NOT SUPPORTED
VGA WONDER V4 ATI18800-1 NOT SUPPORTED
VGA WONDER V5 ATI18810 ATI18800-1 NOT SUPPORTED
VGA WONDER + (V6) ATI18810 ATI28800-2 ? MAY WORK ?
VGA WONDER (1 MB) ATI18810 ATI28800-4 SUPPORTED
VGA WONDER (1 MB) ATI18810 ATI28800-5 SUPPORTED
VGA WONDER XL ATI18810 ATI28800-5 SUPPORTED
Note that the ATI SVGA driver code has hooks and some support for all of
the cards listed. However, with the earlier cards, there are serious
differences in the programmer's interface. The only people who have ever
worked on the code have had 28800-4 and 28800-5 chips, so these chips are
the only ones that are actually known to work.
The ATI GRAPHICS ULTRA chips seem to have SVGA support which is equivalent
to the 28800-4, 28800-5, and 28800-a, although I can find no documentation
on this in the technical manuals provided by ATI. The ULTRA PLUS and ULTRA
PRO boards have been tested to some extent, and seem to work fine, but the
original ULTRA may have some problems at greater than 640x480 resolution.
XII.45) What should I put in my Xconfig file?
ANSWER:
THE MOST IMPORTANT LINE IS THE CLOCKS LINE:
Clocks 18 22 25 28 36 44 50 56
30 32 37 39 40 0 75 65
The first row of clocks is usable on all VGA Wonder cards. The second row
is usable only on V5, PLUS, and XL cards (28800-2, 28800-4, and 28800-5).
However, you should include *BOTH* rows for *ALL* cards.
The specification is different from the specification used with the ET-4000
cards. Programs with claim to detect the clocks usually work only with
ET-4000 based cards. No such program exists for the ATI cards, since it
isn't needed. THE CLOCKS LINE SHOULD NEVER BE CHANGED. The order of
numbers in the clocks line *is* significant.
The actual frequencies, which can be used to calculate the ModeDB lines,
are as follows:
18 = 18.000
22 = 22.450
25 = 25.175
28 = 28.320
36 = 36.000
44 = 44.900
50 = 50.350
56 = 56.640
30 = 30.240
32 = 32.000
37 = 37.500
39 = 39.000
40 = 40.000
0 = 56.640
75 = 75.000
65 = 65.000
The Chipset will be automatically detected. The chipset name for this
driver is "ati".
Here is a sample Xconfig (which I use):
RGBPath "/usr/lib/X11/rgb"
FontPath "/usr/lib/X11/fonts/misc/,/usr/lib/X11/fonts/75dpi/"
Keyboard
AutoRepeat 500 5
ServerNumLock
#Microsoft "/dev/mouse"
#MouseSystems "/dev/mouse"
#MMSeries "/dev/mouse"
#Logitech "/dev/mouse"
#MouseMan "/dev/mouse"
Busmouse "/dev/mouse"
# BaudRate 1200
SampleRate 150
# Emulate3Buttons
vga256
ViewPort 0 0
#Modes "1024x768v" "800x600v"
Modes "800x600v"
Clocks 18 22 25 28 36 44 50 56
30 32 37 39 40 0 75 65
Displaysize 260 184
ModeDB
# From NEC MultiSync 3FGx manual (copied by faith@cs.unc.edu)
# active_area front_porch pulse_width back_porch for horiz and vert
# VESA 640x480 at 72Hz (37.860kHz)
# 20.317us 0.762us 1.270us 4.603us; 12.678ms 0.238ms 0.079ms 0.740ms
# (I couldn't find a good clock rate for this one)
# VESA 800x600 at 72Hz (48.077kHz)
# 16.000us 1.120us 2.400us 1.280us; 12.480ms 0.770ms 0.125ms 0.478ms
"800x600v" 50 800 856 976 1040 600 637 643 693
# VESA 1024x768 at 60Hz (48.363kHz)
# 15.754us 0.369us 2.092us 2.462us; 15.880ms 0.062ms 0.124ms 0.600ms
"1024x768v" 65 1024 1048 1184 1344 768 771 777 806
XII.46) What is the history of the driver?
ANSWER:
The complete history of the driver is cloudy. I have tried to reconstruct
the history from several email messages from a few of the people involved.
The history is probably incomplete and inaccurate.
Apparently, Per Lindqvist (pgd@compuram.bbt.se) first got an ATI driver
working with an early ATI card under X386 1.1a. (This original driver may
have actually been based on an non-functional ATI driver written by Roell.)
Then Doug Evans (dje@cygnus.com) ported the driver to the ATI XL, trying in
the process to make the driver work with all other ATI cards.
I (Rik Faith) obtained the X11R4 driver from Doug Evans in the summer of
1992, and ported the code to the X386 part of X11R5. This subsequently
became part of XFree86. Per and Doug did the majority of work, making the
port somewhat trivial. I am grateful for their contribution.
The port to X11R5 required writing the ATIProbe and ATIEnterLeave routines,
and re-organizing the ATISave and ATIRestore routines. Someone else added
the operating system specific code, but I don't know who.
The major reference that I used was manual from ATI:
"VGA WONDER Programmer's Reference"
ATI Technologies, 1991.
Release 1.2 -- Reference #PRG28800
(Part No. 10709B0412)
However, Chapter 11 (ATI 18800 ATI VGAWONDER) of George Sutty and Steve
Blair's "Advanced Programmer's Guide to SuperVGAs" (Brady/Simon & Schuster,
1990) was also useful. Further, someone e-mailed be a random document from
the ATI BBS which was dated 3Jul91 and which may have been named
PROGINFO.DOC.
XII.47) What is the future of the driver?
ANSWER:
I no longer use an ATI Wonder card and have no interest in this driver.
Since I have received no bug reports or patches for the driver during the
past 3 months, I think it is about as stable as it is going to get. (I
suspect that support for the 28800-2, 28800-4, and 28800-5 chips can be
made more robust, but I doubt that the other chips will ever be fully
supported.)
I have absolutely *NO* intention of porting the code to the mono or 16
color servers. If *you* want ATI support in the mono or 16 color servers,
they *you* should do the port. (Since I no longer use my ATI Wonder card,
please feel free to take over the current driver as well.)
XII.I BUGS:
~~~~~~~~~~~
Restoration of the text screen fails on some hardware. You can
get the program runx from the vgalib distribution (tsx-11
pub/linux/sources/libs/). Report the problem to the xfree86 team.
Cant compile Xaw programs without -static in versions x11v1.1 or older.
Fix for missing numlock control in xfree86-1.1.
use xmodmap to change the mapping:
clear mod1
add mod1 = Alt_L
keysym Alt_R = Mode_switch
add mod5 = Mode_switch
David (Dawes)
Sources for X11_FAQ:
Steve Kotsopoulos, Peter Hawkins, John Morris, MM. Corsini,
K. Balasubramanian.
Direct comments, questions, complaints to krishna at:
balasub@cis.ohio-state.edu
XIII. NETWORKING and LINUX
============================
*** This section should be maintained by Philip Copeland
*** (p_copela@csd.uwe.ac.uk).
XIII.01) Where can I find useful information about networking for
Linux ?
ANSWER: Join the NET channel of the mailing list. Also, you need to
read the NET-FAQ by Phil Copeland, it's on sunsite and tsx-11 and
posted to c.o.l and the NET channel every couple of weeks.
Basically, all of the information on networking and TCP/IP for Linux
is in the NET-FAQ, because it's maintained seperately than this
document (and we don't want to waste the space reprinting it all
here--- it's big!).
The NET-FAQ has been completely rewritten as of Feb 1993. Get the new
version.
XIV. EVERYTHING FOR PRINTING
=============================
**** This is the lpd FAQ version 1.2 (last update 1st june), this
**** section is maintained by Brian McCauley. Please send any
**** suggestion to: B.A.McCauley@bham.ac.uk
XIV.01) What _is_ the lpr/lpd package?
ANSWER: To quote the net-FAQ:
Lpr is a utility that is used to print files out to local and remote
printers. As with all networking, there are configuration files which
have to be tailored to your needs. These files are concerned with
access control, accounting and page setup of your printer(s). Now all
control of the printer is performed by a process called lpd
Lpd is a laser printer daemon (yes I know an epson isn't a laser
printer but thats besides the point). The program lpd is normally
started in /etc/rc.local in either the forground or the background.
What happens is that lpd will sink back as a daemon, only ever being
invoked when a printer request is generated. lpd itself has some
options which are of possible intrest to some people chiefly that of
the '-l' flag which causes lpd to log requests recieved from the
network (useful for debugging). The other option is for what port lpd
should listen for data on and is generally not used.
[ Note: lp really stands for line printer but as technology has
advanced laser printer makes more sense. ]
XIV.02) What is _the_ Linux lpr/lpd ?
ANSWER: Up until 28th May 1993, as far as there is a definitive Linux
version of anything _the_ Linux lpr/lpd is <bir7@leland.stanford.edu>
Ross Biro's port of the BSD sources dated 20th November 1992 (so if
you're used to AT&T's you'll find some differences). Ross's binaries
and the diffs are found in:
tsx-11.mit.edu:pub/linux/binaries/usr.bin/lpr.tar.Z
sunsite.unc.edu:pub/Linux/system/Daemons/lpr.tar.Z
The SLS file lpr.tgz on disk b2 contains the same stuff.
However this is all about to change so from now on this version will
be called the "old" binaries. As of 28th May 1993 there appears be two
different new releases of lpd expected RSN.
In addition to the standard BSD lpr capabilities Ross Biro's port
should have had better support for serial printers (but it didn't work
in the old binaries).
Ross's port was not based on the latest BSD sources (by about 3 years,
'88 vs '91). His new release will also be based on these sources which
can be found are at:
gatekeeper.dec.com and /afs/ir.stanford.edu/src/...
Recent BSD sources are all over the place.
ftp.uu.net:systems/unix/bsd-sources/usr.sbin/lpr
src.doc.ic.ac.uk:unix/bsd-sources/usr.sbin/lpr
to name but two. The BSD sources have unGNU things in their Makefiles
so they don't compile ``out of the box'' but have now been ported by
<waltje@uWalt.NL.Mugnet.ORG> Fred van Kempen and will be available RSN
(if not now) as part of the complete overhall of the Linux Net stuff.
Ross agrees that Fred's version should be thought of as _the_ Linux
lpd package. I <B.A.McCauley@bham.ac.uk> have not had a chance to look
at Fred's port yet (I'm waiting for the full release of 0.99pl10
before I update anything) but most of this FAQ refers to the BSD lpd
package or Linux in general so this is probably not too important.
There are some other versions of lpd stuff about but I have to draw
the line somewhere. If anyone whants to maintain FAQ sections on these
they are welcome to do so. The other versions include plp and the ka9q lp
facility. There are also romours of SysV spoolers being ported.
XIV.03) How do I get lpr to work properly other than for root? Where
do the files go and what should their permissions be?
ANSWER: Put /etc/lpd& in your /etc/rc or /etc/rc.local (usually in
/etc/rc.local after you start syslogd (if you use syslogd)).
Set the group fields of the file permissons/ownership as follows:
-rwxr-s--- 1 root daemon 37892 Nov 19 23:32 /etc/lpd
-rwx--s--x 1 root daemon 21508 Nov 19 23:32 /usr/bin/lpc
-rws--s--x 1 root daemon 17412 Nov 19 23:32 /usr/bin/lpq
-rwx--s--x 1 root daemon 17412 Nov 19 23:32 /usr/bin/lpr
-rwx--s--x 1 root daemon 17412 Nov 19 23:32 /usr/bin/lprm
...and for each of the spool directories listed in the sd fields of
/etc/printcap...
/usr/spool/lp1:
total 5
drwxrwxr-x 2 root daemon 1024 May 18 23:00 .
drwxr-xr-x 11 root root 1024 Feb 19 20:56 ..
-rw-rw-r-- 1 root daemon 4 May 18 23:00 .seq
-rw-rw-r-- 1 root daemon 18 May 18 23:00 lock
-rw-rw-r-- 1 root daemon 25 May 18 23:00 status
Note these 3 files are created by lpr and lpd so if you've never
run these they could be missing so touch them into being.
Some of these permissions may be overkill so don't be surprised if
your system works with different permissions. The fact that lpq needs
to be suid root is inelegant and may change in the future.
Alternative approach: just make lpc, lpr and lprm setuid(root) then
you can forget the file permissions on the spool queues!
You're free to choose different directories for the executables on
your system (notably lpc is often in /etc even though it has commands
that are useful to non-root). The master lpd lock file which is always
in /usr/spool/ so you must have one of those but you needn't
necessarily keep your spool queues there.
The main configuration file is /etc/printcap although the old binaries
look for it in /usr/etc/lpd. The README says hosts.equiv and
hosts.lpd files are looked for in /etc/inet but again the more
accepted convention is to locate them in /etc.
For an simple life:
mv -i /usr/etc/inet/* /etc/inet/* /usr/etc/* /etc/
rmdir /usr/etc/inet /etc/inet /usr/etc
ln -s ../etc /usr/etc
ls -s . /etc/inet
[But make sure the files in /usr/etc and /etc/inet are not already
symlinks into /etc or you'll be in deep water.]
Eventually you'll be able to delete these symlinks once you've got rid
off all the old stuff.
XIV.04) Why do I sometimes see /usr/etc or /etc/inet or /usr/etc/inet?
ANSWER: For various reasons including the difficulties of fsck'ing /
some people like to put all things that aren't needed during bootstrap
in /usr so that / is as small as possible. Also some people like all
their internet stuff in /etc/inet. Most Linux gurus advise keeping it
all in /etc.
If you really want this quirky layout:
cd /etc
ls -s ../usr/etc/* .
ls -s inet/* .
[Make sure the files in /usr/etc and /etc/inet are not already
symlinks into /etc.]
XIV.05) Where do I get a printcap for a xxxxx?
ANSWER: This question is essentially meaningless - see next question.
XIV.06) What is the content of /etc/printcap?
ANSWER: Given the similarity in appearence and name between
/etc/termcap and /etc/printcap one could be forgiven for assuming that
they contain analogous infomation. This is not the case. Whereas
termcap contains informations about terminal *types* - (mostly escape
seqences) printcap contains information about *specific* printers
(like the directory that holds the spool queue and the device name of
the printer). The information about a printer model's escape sequences
and so on are held in the various filters which are _programs_ called
by lpd to drive the printer. /etc/printcap simply gives the locations
of these filters. For details RTFM(printcap). [Alternatively the
net-FAQ has a summary of some of the more important fileds.]. One last
point you should always specify ``suppress header'' ``:sh:'' unless
you have _text_ printer and want banners. (See the banners question
for further expanation).
XIV.07) What is the format of a /etc/printcap?
ANSWER: Ideally RTFM(termcap) (yes, I said *termcap*) but since most
people don't have TFM(termcap) here are the essentials.
Lines starting # are comments (as you might have guessed).
For each printer usable from the lpr command on your system there is
one logical line in the file. For the sake of readability each logical
line may be spread over several physical lines by making the last
character on all but the last physical line a backslash.
Each logical line has the following format:
name1|name2|name3:string_capability=string:\
:numeric_capability#number:boolean_capability:
The leading spaces and colon on the second line are for readability
only.
A printer can have as many names as you like but conventionally the
final name is used as a longhand description of the printer. (Still
people are free to say `lpr -P "Fred's grotty teletype"' if that's
the description you've given.)
The list of capabilities can be as long as needed and the order is not
significant. Each ``capability'' is denoted by a two character code.
(The name ``capability'' comes form the file format's termcap heritage
- parameter or attribute would be a more sensible terms.) [Note from
Ross Biro: capabilities with 3 character names don't work properly
which is why the serial port stuff in the old binaries failed.]
Capabilities having string value and have a = delimiter between the
capability name and the value while those having a numeric value use a
# (actually they can use either a # or an =). Boolean ``capablilties''
are true if they appear in the list and false if they do not.
Special characters in a string value can be expressed using `\'-escape
sequences as in C; in addition, `\E' stands for ESC. `^' is also a
kind of escape character; `^' followed by CHAR stands for the
control-equivalent of CHAR. Thus, `^a' stands for the character
control-a, just like `\001'. `:', `\' and `^' themselves can be
represented as `\:', `\\' and `\^'.
Example:
lp|bam|BAM's Epson FX-80:lp=/dev/lp1:sd=/usr/spool/lp1:sh:mx#0:\
:df=/usr/local/lib/magic-filter/lp.df:\
:if=/usr/local/lib/magic-filter/lp.if:
Name is lp (this is the printer that lpr uses by default). It's also
known as bam or "BAM's Epson FX-80".
The printer is on /dev/lp1 (aka AT-bus LPT1:). I don't want a burst
page. I don't want a file length limit. Files queued by `lpr -d' are
passed through /usr/local/lib/magic-filter/lp.df and those queued by
`lpr' through /usr/local/lib/magic-filter/lp.lf.
See also the next question.
XIV.08) My /etc/printcap looks identical to someone else's but it
doesn't work - why?
ANSWER: See if lpc stat reports a printer called ` :'. The last
character on a continued line must be a \. If there are whitespace
characters after the \ then it doesn't register the next line as a
continuation.
XIV.09) What's the minimum /etc/printcap?
ANSWER: This is a silly question but it _is_ frequently asked. The
answer is ``lp:sh'' (that's 6 bytes including the linefeed character
on the end). To use this /etc/printcap you must make /dev/lp a symlink
to your printer and create your spool queue directory as
/usr/spool/lpd. (You might think that if you wanted banner pages you
could loose the ``:sh'' but the termcap syntax requires at least one
charateristic per entry).
XIV.10) How do I prevent the `staircase effect'?
ANSWER: Unix terminates each line of a file with a linefeed but not a
carriage return so taken literally a Unix text file printed on an
ASCII device will start each line below the end of the previous line.
Some printers can be set to treat "linefeed" as "carriage return,
linefeed", others can't. If yours can then do simply do that. If the
printer cannot be fixed create a shell script filter that reads:
#!/bin/sh
if [ $1 = -c ]; then
cat
else
sed -e s/$/^M/
fi
echo -ne \\f
where ^M is a carriage return character not a ^ followed by a M. To
type ^M in emacs use the sequence C-q C-m and in vi use C-v C-m.
Conventionally this script is called /usr/lib/lpf. The test of $1
allows the isertion of carriage returns to be switched off by the `-l'
switch on lpr.
Alternatively your printer may have an escape sequence that will set
the way it handles linefeed characters. A simple filter that uses an
`echo -ne' command to send this sequence may be appropriate.
#!/bin/sh
# Filter for HP printers to treat LF as CRLF
echo -ne \\033\&k2G
cat
echo -ne \\f
[ Ross Biro's new release will include a filter called lpof that does
this. ]
XIV.11) How do I get my printer to go back to the default font after
each printout?
ANSWER: Define the `tr' ``capability'' in /etc/printcap to be your
printer's font reset command. For details of the format of this string
see the question on the format of printcap. This may not work if a
printout crashes in the middle of an escape sequence - putting a lot
of ^@ on the front may help but this probably won't be enough it you
were printing raster graphics when the filter died.
XIV.12) How do I prevent a formfeed at the end of every printout?
ANSWER: If you don't have an `if' specified in /etc/printcap then lpd
will automatically put a formfeed at the end of each file. If you're
using a filter then it's up to the filter to decide if it wants to put
a formfeed. To disable formfeed completely if you don't have an `if'
put :ff=: in your /etc/printcap. But please note this suppresses the
formfeed that would usually be printed if a filter dies. If you want
formfeeds after text printouts but not on printouts printed with `lpr
-l' then create the following `if' filter:
#!/bin/sh
cat
if [ "$1" != -c ]; then
echo -ne \\f
fi
If you want a formfeed after `lpr -l' to be optional you can misuse the
`-i' switch to suppress the formfeed with the following trick (after
all `lpr -i -l' would usually not be implemented).
#!/bin/sh
cat
# use lpr -i -l to print raw without trailing formfeed
if [ "$1" != -c -o "$4" = -i0 ]; then
echo -ne \\f
fi
XIV.13) How do I get burst/banner pages?
ANSWER: For a simple text printer (in particular not postscript) and a
simple text banner simply take :sh: out of the printcap record. If you
want to prevent the banner comming out in whatever font was last used
on the printer then define the `tr' ``capability'' to be your
printer's font reset command.
If you want a fancy cusomised banner (or have a postscript printer)
leave :sh: in the printcap and make each of your filters print the
banner. All the information to put on the banner is included in the
filter's positional parameters. RTFM(printcap) for details. [ If
you're using <B.A.McCauley@bham.ac.uk>'s magic-filter package then
call the code to print the banners from the config script. ]
XIV.14) How do I print text on a ps printer?
ANSWER: You need a filter based on a program called enscript. I don't
know where you can get enscript for Linux.
XIV.15) Why do files with a lot of graphics get truncated?
ANSWER: Usually because you've got a limit set on the maximum size
file that can sit in the spool queue. Put `mx#0' in your printcap.
XIV.16) Why doesn't `lpr -i' work?
ANSWER: To get lpr -i to work you need a filter istalled as `if' that
implements it. The -i switch is simply passed on by lpd to the filter.
There is such a filter called /usr/lib/lpf that can be opdtained from
bsd source archives in the directory bsd-source/lpr/filters.
XIV.17) Why doesn't `lpr -p' work?
ANSWER: Because its broken in the '88 sources from which the old binaries were
compiled. lpd always thinks that the printer is 0 characters wide
regardless of what printcap says.
XIV.18) Why does lprm say `too many users'?
ANSWER: This was a bug in the C library in the scandir() function which failed
when an unsorted directory list was requested. This has been fixed now
so just get a new libc.so. Alternatively there is a lprm with a
workround (by toy@soho.crd.ge.com) in:
sunsite.unc.edu:/pub/Linux/system/Daemons/lprm-fix.tar.z
XIV.19) Why does lpc complain that it hasn't stopped a daemon?
ANSWER: There's one lpd process that runs all the time and it spawns
off children to handle each printer as needed. These children leave
lock files lying about so lpc thinks that they are still running and
tries to kill them. This is probably a bug but it's not serious.
XIV.20) How do I print over a Network?
ANSWER: To print on the printer "foo" connected to the machine
"bar.baz.net" from the machine "mine.baz.net" you put an entry like
this in your /etc/printcap (on mine.baz.net):
foo:lp=:rm=bar.baz.net:rp=foo:sd=/usr/lpd/spool/foo:
and, of course, create the spool directory /usr/lpd/spool/foo.
There's no point specitying filters and the like in
mine.baz.net:/etc/printcap as it's the one in
bar.baz.net:/etc/printcap that will get used.
On the machine bar.baz.foo, you need to put "mine.baz.net" on a line
by itself in either /etc/hosts.equiv or /etc/hosts.lpd; note that
putting it in /etc/hosts.equiv will allow for unauthenticated logins as
well as printing. /etc/hosts.lpd is printing only.
[ Editors note: I realise there are more questions about net printing
but I don't know them. If you find any let me know (with answers if
possible). See also the net FAQ ]
XIV.21) Which /dev/lp* is my printer?
ANSWER: On an XT bus system LPT1: becomes /dev/lp0 (major=6, minor=0),
on an AT LPT1: becomes /dev/lp1 (major=6, minor=1). To be more
precise:
Name Major Minor I/O address
lp0 6 0 0x3bc
lp1 6 1 0x378
lp2 6 2 0x278
XIV.22) When the kernel boots it says `using polling driver,' is
there an interrupt driven driver?
ANSWER: Yes the latest kernels support parrallel port IRQs but to
enable them you must use `lpcntl'. The interrupt driver uses less
processor time and gives a faster throughput but sometimes one may
choose sacrifice this to free up an IRQ line. If your printer is
/dev/lp1 using IRQ 7 put:
/etc/lpcntl /dev/lp1 7
in your /etc/rc.
If you don't have lpcntl get it from:
tsx-11.mit.edu:/pub/Linux/BETA/lp/lpirq.4.tar.Z
XIV.23) What's the difference between the polling driver and an the
interrupt driven one?
ANSWER: A polling sits in a loop repeatedly reading the port hardware
to see if it is ready for another character and then gives it one. An
interrupt driver goes to sleep and is woken by the port hardware
(using a hardware interrupt) when it wants another character. Polling
drivers are easier to write but in multitasking environments they are
a bad thing.)
XIV.24) How do I write filters?
ANSWER: In normal Unix terminology, filters are just programs (so they
must have execute permission) that read a stream from their standard
input and write to their standard output.
lpd filters are filters in the sense that thay read from their STDIN
and write to their STDOUT but are not necessarily true filters in that
they may assume that their standard input is a file and perform
lseek() operations on it. I'm not sure that such filters are strictly
conforming but they seem to work OK.
All lpd filters must conform to a particular command line syntax (or
more often simply ignore command line parameters). For details of the
command line parameters RTFM(printcap).
It's easier to debug filters if you test them in an immediate shell
before you install them. (If your filter make use of its command line
arguments you'll have to specify them too).
my-new-filter <file >/dev/lp1
If you want to write a shell script filter it must have a #!/bin/sh
header. Here for example is my Epson FX-80 dvi filter:
#!/bin/sh
/usr/TeX/bin/dvips -f | \
/usr/bin/gs -q -sDEVICE=eps9high -r120x216 -dNOPAUSE -sOutputFile=- -
XIV.25) I've written a filter in shell script which works OK when I
use it from the shell prompt but fails when I tell lpd to use it?
ANSWER: You may need a #!/bin/sh header. You may also need to set PATH
within the script since the daemon's PATH may not have everything you
need.
XIV.26) When should I define an `of' filter?
ANSWER: Never. (Well strictly speaking there are circumstances but
you're unlikey to meet them until you're so familar with lpd that you
won't need this FAQ.)
XIV.27) Where do I get filters for given printers?
ANSWER: From BSD source archives (I think). [Editors note: I need a
better answer here ].
If you already have a program to print say DVI on your printer then
making it into a filter is usually a matter of writting trivial shell
script - see this FAQ. If the program you are using insists on reading
a names file as input see the next question. Text mode filters are
trivial too (see this FAQ) unless you want lpr to have a choice of
fonts in which case they are slightly harder than trivial. You will
probably want to insert and `echo -ne' command at the beginning and
end of your filter to set up the font etc to your liking.
XIV.28) How do I get a program that won't read from STDIN to act as
a filter?
ANSWER: For example dvilj2p insists on a named file as its input (and
what's more expects one with a .dvi suffix). To make a dvi filter for
a laserjet you'll have to capture the STDIN into a temporary file.
#!/bin/sh
cat >/tmp/$$.dvi
dvilj2p /tmp/$$
rm /tmp/$$.dvi
The problem with this kludge is that is may have a very high disk
overhead. (In the long term dvilj2p should be fixed).
XIV.29) The set of filters supported by lpd seem strange - why
aren't there filters for Postscript, GIF, TIFF and so on?
ANSWER: History. You can, in fact, use any of the filters (with the
possible exception of the default filter) for any reason. If you're
never going to use Benson Varian raster files you could use the -v
switch for GIF files. If you are on a network remember that the filter
setups go on the print server so you will need to consider other
people's needs too. To avoid running out of possible types use magic
filters.
XIV.30) What are magic filters?
ANSWER: Magic filters deduce their input files' types from `magic
numbers' (distictive byte patterns at particular offsets). Magic
filters are usually perl scripts, shell scripts or C programs that
simply identify the file type then call the appropriate non-magic
filter. Blatent plug :-) I <B.A.McCauley@bham.ac.uk> have a generic
magic filter bash script that selects the right filter to use based on
the output of the `file' command. With my magic ``dvi'' filter (and
existing separate dvi and ps filters) I can do things like:
lpr -d file1.dvi file2.div.Z file3.ps file4.ps.z
This is not yet on any anon-ftp sites so e-mail me for a copy. Once
it's stable and I've had a bit of (possitive) feedback I'll upload it.
Magic filters should never specified as `of' as the output filter only
gets called once if a number of files are printed without a gap.
Magic filters as `if' can also be a problem as they may prevent you,
say, _listing_ a PostScript or nroff file.
===================8<==========>8================
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# LaBRI | #
# 351 cours de la Liberation | e-mail: corsini@geocub.greco-prog.fr #
# 33405 Talence Cedex | e-mail: corsini@labri.u-bordeaux.fr #
# | #
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
--
There will be a sig when our local net is reliable.
For now, I would rather stay anonymous.