After a while, each system is full of software it doesn't use. With emerge --depclean it's possible to track down these packages - they will be removed instantly. However, do this only after an correct update like emerge --pretend --verbose --newuse --deep --update world. Now you can easily run emerge --depclean and finally revdep-rebuild.

emerge --newuse --deep --update --verbose --tree --ask world
emerge --depclean
revdep-rebuild

/etc/portage/package.keywords
/etc/portage/package.unmask
/etc/portage/package.use
/var/lib/portage/world
/usr/portage/profiles/use.desc
/etc/make.conf
/etc/layman/layman.conf

Sometimes, an emerge -C package doesn't delete all installed files. Files in $CONFIG_PROTECT will not be touched. So this is the clean deinstalling mechanism:

CONFIG_PROTECT="-*" CONFIG_PROTECT_MASK="-*" emerge -Cv package 

You can check your world-file /var/lib/portage/world with the following command:

emaint --check world

If you want to test a huge software with lots of dependencies, you can't just do an emerge -aC programname because the dependencies wouldn't be uninstalled. You could use emerge depclean, but there is another, cleaner and more secure way.

Let's say you want to test kdevelop - but you do not know whether you want to use it for a long time. So create the dependencies-file:

emerge kdevelop -p | 
sed -ne '/^\[ebuild[^]]*N[^]]*\]/{s/^\[ebuild[^]]\+\] ([^[:space:]]\+\).*$/=\1/;p}' \ 
> kdevelop.deps

This ensures that no packages are listed in the deps-file which has just been updated and not newly installed. If you want to get rid of kdevelop in the future, use this:

emerge -aC $(

		More infos on forums.gentoo.org

	

First, one has to specify locales for glibc. Add the locales into the /etc/locale.gen and run locale-gen afterwards.

en_US ISO-8859-1
en_US.UTF-8 UTF-8
de_DE ISO-8859-1
de_DE@euro ISO-8859-15
de_DE.UTF-8 UTF-8

Test with locale -a whether the locales were installed correctly and to list their correct names. It should something like that:

C
POSIX
de_DE
de_DE.utf8
de_DE@euro
en_US
en_US.utf8

Next, the environment variables LANG and LC_ALL have to be set to a correct UTF-value. This can be done for example in the /etc/env.d/02locale. Test with env-update && source /etc/profile && locale whether the output of the environment variables is correct.

Open the image and go to Ebene - Farben - Kurven. Play with the curve to see the effects immediately. That saved me from hell,

Just run hdparm -Tt /dev/hda to get the current speed of your hard disk. After that you can tune it a bit: therefore, just run hdparm without any parameters on your hard disk: hdparm /dev/hda. This will print some attribute values (e.g. it will tell you whether DMA is activated or not).

multcount gives the number of sectors which should be read in with each I/O-interrupt. Possible values are normally 2, 4, 8, 16. You can get the value of your hard disk by running hdparm -i /dev/hda and reading the MaxMultSect value. But be cautious: this value is not always true, data loss could be the consequence. If you're trying it anyway, then do hdparm -mnum /dev/hda.

Another tuning parameter is the I/O-Support-Mode. This should be nowadays 3, so try -c3 and see man page for more infos.

The Using_dma flag can be enabled with -d1, -X66 tells the mode (which is here Ultra-DMA-Mode 2) and -u1 tells the controller driver to unmask other interrupts.

To keep the settings also if an IDE reset appears, add -k1 to your parameters.

If you have found your best configuration parameters for hdparm, write it to local.start so that it runs on each boot!

Just do a convert -transparent black if you want to get the black color transparent.

There is a built-in function called ignore to get rid of several lines in irssi (see /help ignore). Here some examples which should be self-explanatory:

/ignore #channel JOINS
/ignore ALL QUITS

Instead of JOINS and QUITS you can use one of the following levels:

CRAP, CTCPS, JOINS, PARTS, QUITS, KICKS, TOPICS, PUBLIC
SNOTES, NOTICES, MODES, DCC, NICKS, ACTIONS, WALLOPS, MSGS
INVITES, DCCMSGS, CLIENTNOTICE, CLIENTCRAP, CLIENTERROR

Another way to hide special lines use the script crapbuster.pl in the autorun folder and test it:

/script load autorun/crapbuster.pl
/set crapbuster_levels QUITS KICKS PARTS JOINS

If it works, add this line to the perl/core/scripts section of your irssi config:

crapbuster_levels = "QUITS KICKS PARTS JOINS"

The difference to ignore is that with crapbuster the lines arent hidden at once; instead you have to invoke /crapbuster. Have fun!

IBM Laptops have the ThinkLight technology in their notebooks: you can use that as an information for incoming messages. Load the script beep_beep.pl in the autorun folder. In the core settings:

beep_msg_level = "PUBLIC MSGS";

And in the perl/core/scripts section:

beep_cmd = "echo on > /proc/acpi/ibm/light; sleep 0.1; echo off > /proc/acpi/ibm/light";

Keep in mind that you must have the permission to set the light file under /proc!

$\mid$ will print |

There is a bug with utf8-encoded latex files - utf8.def is missing. Just invoke the following commands and the problem is gone:

cd /usr/share/texmf/tex/latex/latex-unicode/
ln -s utf8x.def utf8.def
texhash

Add a math formula in the same line:

\( MATH-FORMULA \) 

Add an indented math formula:

\[ MATH-FORMULA \]

Adding vector arrow to a variable:

\( \vec{x} = ... \)

Using lower and higher indices:

\( x = x_{1} + x_{2}^{3} \)

What's the world without infinity...

\( \infty \)

And what about some Greek letters...

\( \alpha \beta \gamma \delta \epsilon \phi \psi ... \)

Fractions can also be displayed easily:

\( one half = \frac{1}{2} \)

A sum sign (look at the \limits):

\( \sum\limits_{i=0}^{10} i \)

An integral sign (look at the \limits):

\( \int\limits_{i=0}^{10} i \)

Search your package with your favourite search engine: "ctan csquotes". Unpack it in /usr/share/texmf/tex/latex/csquotes and run texhash.

For a swap partition just invoke the following command:

swapon /dev/hda9

For a swap file invoke these commands:

dd if=/dev/zero of=/swapfile bs=1024 count=65536
mkswap /swapfile 65536
sync
swapon /swapfile

Each file which is created gets a default file permission. This can be controlled via umask. You have to give umask one argument - the opposite octal value of the mode you want. i.e. you wanna have 700 as default file permission, you have to invoke umask 077 (700 + 077 = 777 as a hint).

For your X session (/etc/X11/xorg.conf):

Section "Monitor"
	...
	Option	"DPMS"
	...
EndSection

...

Section "ServerLayout"
	Option "BlankTime" 		"2"
	Option "StandbyTime" 	"4"
	Option "SuspendTime" 	"6"
	Option "OffTime" 		"8"
EndSection

For the console (~/.bashrc):

setterm -blank 2
setterm -powersave on
setterm -powerdown 5

On my IBM T43 some hotkeys - the ones using Fn - are deactivated. To enable functions like disabling the LCD write this to /etc/conf.d/local.start. Don't forget to load the kernel module called ibm_acpi / thinkpad_acpi.

echo enable,0xffff > /proc/acpi/ibm/hotkey

Example entry in /etc/fstab for a reiserfs partition:

/dev/hdb1 /mnt/storage reiserfs users,rw,noatime,owner 0 0

The mount point should have 777.

Feeling the red vine of last night in my guts, it's better not to study... better start playing again :-D

Does your login also look so boring and standardized? Make it a bit cooler using some nice tools (checkout Gentoo Forums).

Some "background knowledge": /etc/issue is responsible for the output before the user login and /etc/motd is the message of the day displayed right after the login of a user. Here comes linux-logo into play, and of course don't forget cowsay (moooooh :-)) and fortune...

echo "games-misc/fortune-mod offensive" >> /etc/portage/package.use
emerge linux-logo cowsay fortune-mod fortune-mod-starwars
linux_logo -L 3 > /etc/issue
echo "fortune starwars | cowsay > /etc/motd" >> /etc/conf.d/local.start

May the force be with you! Good night!!

Normally, you should install Windows first and then Linux because if you do it the other way, the Windows installation will delete your MBR (master boot record) installed by grub.

Of course there is a possibility to achieve the other way: After having installed Windows boot with a live cd and run the following commands (taken from Gentoo Weekly Newsletter):

mount /dev/rootpart /mnt/gentoo
mount /dev/bootpart /mnt/gentoo/boot
mount -t proc none /mnt/gentoo/proc
mount -o bind /dev /mnt/gentoo/dev
chroot /mnt/gentoo /bin/bash
env-update && source /etc/profile
grep -v rootfs /proc/mounts > /etc/mtab
(Assuming your hard drive is /dev/hda)
grub-install /dev/hda

I've got a new mp3 player, the Sansa SanDisk e250 with 2GB hard disk. Before you plugin the sansa, change the usb setting to MSC. After that add a udev rule as described below:

nano /etc/udev/rules.d/10-local.rules
KERNEL=="sd?1", BUS=="usb", SYSFS{product}=="Sansa e250", NAME="%k", SYMLINK+="sansae250"

mount the device as a normal storage device: mount /dev/sda1 /mnt/sansa

First of all some nice shortcuts (Emacs mode):

TAB auto-complete
Ctrl-r search for commandos
Ctrl-l clean screen
Ctrl-e go to end of line
Ctrl-a go to beginning of line
Ctrl-u delete whole line
Ctrl-k
Ctrl-d delete character under
Alt-b go one word backward
Alt-f go one word forward

If you prefer vi-Mode, then type set -o vi

You can see all bindings using bind -p; bind -l will list all available functions! Use bind -u function and bind -q function to set new shortcuts.

dd stands for "convert and copy" - weird? No, because cc is the C compiler :-D

So for what is dd good for? Perhaps for saving your MBR (master boot record):

dd if=/dev/hda of=bootsector bs=512 count=1

Or raw copy a whole cd (not mounted):

dd if=/dev/cdrom of=foobar.iso

This is fine for raw data - but CDs (compact discs) need a iso files: mkisofs is your friend

mkisofs [options] -o outputfile.iso /tmp/data

options can be -r (Rockridge), -j (Joliet), -R (Rockridge + Owner + Permissions), and many more, see the man page!
Here some more: -V "Volume-ID" sets a volume ID for the CD

Before you actually write your data onto the CD, you can test it using the loop device:

mount -o loop foobar.iso /mnt/tmp

After testing your data unmount /mnt/tmp and finally burn it on CD:

cdrecord --dev=ATAPI:0,0,0 driveropts=burnfree foobar.iso

Home directories and other data could be burned as follows:

mkisofs -o /tmp/backup.iso /home/user
cdrecord -v --dev=ATAPI:0,0,0 driveropts=burnfree /tmp/backup.iso

But what to do when the file is too big for one compact disc? Use split!

split -b 695m backup.iso backup.iso.split.

This command will create files which have exactly the size of 695 MB (except the last one). Restoring the original file is as simple:

cat backup.iso.split.* > backup.iso

The only problem with this scheme is that files are twice on the machine! If you have a small disk with vast space, use the pipe:

tar cz /home/user | split -b 695m - archiv.tar.gz.split.
cat archiv.tar.gz.split.* | tar xz

diff file1 file2 assumes that file2 is newer than file1. Thus, the output has to be interpreted as which changes have to be made to file1 to get the contents of file2!

13c13 

means: line 13 in file1 has changed to line 13 in file2

81a82,84 

means: lines 82 through 84 of file2 follow line 81 of file1

with large outputs use these ones:

diff file1 file2 | less 
diff file1 file2 > changes.diff

If you're just interested if two files are identical than invoke:

diff --brief file1 file2 (or just -q)

Another perspective can be shown:

diff --context file1 file2 (or just -c)

! stands for changed lines, + for added lines and - for removed lines.
Patch files are changes in the Unified-Diff-Format. You can get them using the --unified (-u) parameter.

If you don't have the smallest idea about unicode / UTF-8, read this article.

Be sure that your locales are configured correctly:

locale
locale -a
locale -m

Since I have a keyboard with US layout (which really rocks at programming) I miss my German Umlaute ae, oe, ue, ss and the Euro sign. Add this to your .xinitrc:

 
xmodmap -e "keycode 113 = Mode_switch Mode_switch"
xmodmap -e "keysym a = a NoSymbol Adiaeresis NoSymbol Adiaeresis"
xmodmap -e "keysym o = o NoSymbol Odiaeresis NoSymbol Odiaeresis"
xmodmap -e "keysym u = u NoSymbol Udiaeresis NoSymbol Udiaeresis"
xmodmap -e "keysym s = s NoSymbol ssharp NoSymbol"
xmodmap -e "keysym e = e NoSymbol EuroSign NoSymbol"

AltRight has keycode 113 and is therefore used as a mode_switch for my Umlaute and special characters. Look at /usr/include/X11/keysymdef.h to find the correct names for the xmodmap commands.

If you want to test your console client, load a unicode test file and cat it. For comparison, watch it in firefox (which has decent UTF-8 support) or run the cat command within a xterm invoked by LC_CTYPE=en_US.UTF-8 xterm -fn '-Misc-Fixed-Medium-R-SemiCondensed--13-120-75-75-C-60-ISO10646-1'.

Another way to test your console client is to invoke some perl one-liners (should print the Euro sign):

 
perl -C -e 'print pack("U",0x20ac)."\n"'
perl -C -e 'print "\x{20ac}\n"'

It's a combination of client-server-architecture with IMAP / POP support. The main idea is: keep the mails within the imap folders on the server and have several clients syncing with this server. The problem was that there are a lot of different mail addresses with whom I had to struggle with.

Here's what's happening, in short: I'm outside with my laptop and want to receive mail. First, I check whether there are differences on my IMAP server using offlineimap. After that, I use a fetchmail and procmail to receive mail via POP3 from my different mail addresses - I use several filters in procmail and mutt to keep the messages in the correct (local) IMAP folder. Before quitting, changes to these folders are synchronized to the IMAP server so that access in the future from other PCs for example results in a consistent and up-to-date mail directory!

The quick and dirty way: copy the mutt, procmail, fetchmail, offlineimap and msmtp configuration into the home directory, see the configuration files on the other page on this website. Adjust them for your own needs!

So what to do when you want to add a mailing list for example. First, create the mailboxes in your .maildir/ folder and on the remote server where the IMAP server is. Of course the names must be identical. After that you can move the mails from INBOX.FOO to INBOX.FOO.2007-10 using mutt: tag the messages you want to move with t, and save them using ;s. Mutt has to know about mailboxes etc. and because of that you have to configure it in the .muttrc. Look at your mbox-hooks, lists and mailboxes command. Also check your procmail configuration so that the mails are stored in the correct mailbox.

Starting mail without any parameter, the INBOX is shown in an interactive window. The handling is a bit difficult and circumstantial, but if no other MUA is available, this is just fine. Here some shortcut infos (similar to mutt):

? help screen
+ go to next mail
- go to previous mail
NUM read message number NUM
d delete (also d 1 2 or d 1,5 or d *)
u undelete
q quit
x exit
r reply
m mail

You write mails until a line appears which consists only of ".". If you wanna switch to your favorite editor type ~v after you've typed in your subject line (the $VISUAL variable is read out).

The real strength of mail comes out when it is used automatically (e.g. in scripts):

echo "foobar bla bla" | mail -s "Subject" -b "bcc1@domain.de" -c "cc1@d.de,cc2@d.de" foo@bar.de
mail -s "Subject" foo@bar.de < file

Use -v if you wanna watch the SMTP connection and get more infos about the mail delivery!

C-x j = show background jobs
C-s = search within directory
insert = select file / directory

mplayer -ao pcm movie-file

Note: this would have gone faster if I had used -vc null -vo null -ao pcm:fast as arguments!

for i in *.wav; do lame -h $i `basename $i wav`mp3; done;

Add this to your existing mutt-configuration:

set print_command="muttprint"
set print="yes" 
set print_split

macro index p "display-toggle-weed> print-message> display-toggle-weed> exit>"
macro pager p "display-toggle-weed> print-message> display-toggle-weed>"

I didn't have luck when I tried it the first time. muttprint ended with "error in charset conversion". After looking in the log-file (debug set to "1" in ~/.muttprintrc) I found out that latex was missing a utf8.def file. To solve this problem you must install latex-unicode and change utf8 into utf8x in commondefs.tex.

Linux User article about muttptrint

If you dont know what you should write into the /etc/resolv.conf :D

search freenet.de
nameserver 62.104.196.134 #DNS Freenet
nameserver 192.76.144.66 #DNS MSV

search t-online.de
nameserver 212.185.252.73 #DNS tonline
nameserver 62.225.252.16 #DNS tonline
nameserver 194.25.2.129 #DNS tonline

Kernel configuration:

Filesystems
	Network Filesystems
    	M NFS Filesystem Support 
    	* provide NFSv3 client support 
    	M NFS Server support 
	    * provide NFSv3 server support

Steps the client / server has to do:

both: emerge nfs-utils
both: vim /etc/hosts
NFS-Server: echo "/mnt/media thinktuxwireless(async,no_subtree_check,rw)" >> /etc/exports
NFS-Client: mount -t nfs gateway:/mnt/media /mnt/nfs/

nmap is a scanning-tool for hackers and admins - it detect ports / services on hosts.

possible responses:

open - connection was accepted by one service
filtered - a firewall filters this port
unfiltered - there is no firewall

Scan methods:
-sS = tcp syn scan: "half-open" scanning; only the syn paket is sent; if syn/ack returns, this port is listened and finally RST is sent; if port returns with RST, there is no service!
-sT = tcp: "full" scans, full connection establishment!
-sF = FIN scan: use FIN package to test if port is open, because many firewalls do not accept SYN packages
-sX = Xmas Tree... like -sF, but FIN, URG and PUSH are used
-sN = Null scan, turns off all flags
-sP = ping scans; either with ICMP echo requests, ACK requests or normal SYN packages
-sU = UDP scans: one 0 byte UDP package is sent to the host; if ICMP message "port unreachable" returns, then port is closed; otherwise it is assumed it's open;
-sO = ip protocol scans: scans protocols; if "protocol unreachable", protokoll doesn't exist on host
-sA = ACK scans: test, whether a stateful firewall is used; if RST, then port is unfiltered, otherwise it's filtered;
-sR = rpc scan
-sL = list scan: list all hosts without pinging or scanning ports

Options:
-P0 = do not ping before port scan
-O = fingerprinting
-oN file = write results to file
-p 20-30,139,6000- = list of ports to be scanned
-n = do not do any reverse dns
-v = verbose mode

Convert the doc file into an OpenOffice document (odt) and unzip it. That's it!

Check version of CGI.pm modul: perl -MCGI -e 'print $CGI::VERSION'

/etc/apache2/mod_evasive20.conf contains the configuration of the evasive-module for apache which prevents DOS-attacks.

DOSHashTableSize 3097
DOSPageCount 10
DOSSiteCount 100
DOSPageInterval 1
DOSSiteInterval 1
DOSBlockingPeriod 5
DOSEmailNotify email@addy.de
DOSLogDir "/tmp/evasive"

This has to be interpreted as follows:
It will log an attack to /tmp/evasive if the threshold for the same page / URI (PageCount) is exceeded per PageInterval or if the same client exceeds the SiteCount threshold per SiteInterval. The IP will be blocked for BlockingPeriod seconds and a notification Email is sent to EmailNotify.

To receive this Email, you have to check whether the system variable $MAILER is set. You can check this running echo $MAILER on console. If it's not set like on my system, edit ~/.bashrc and type in it:

export $MAILER="/bin/mail -t %s"

Use the incoming log mails to tune your configuration! More information under [1].

[1] README

If you want to convey mails with special characters using utf you have to transmit several headers and you must encode your strings using utf8_encode().

The headers are:
Content-type: text/plain; charset=utf-8 Content-encoding: utf-8 Content-transfer-encoding: utf-8

If you do not want to use the whole ps file with all its pages, you can use psselect to select only the even or odd pages for example.

psselect -e infile outfile will produce outfile with the even pages, -o with the odd pages and with -ppages you can give the number of pages you want as a comma separated list or with the format start-end.

#!/bin/bash
deskid=`xwininfo -int -name 'Desktop' | grep 'Desktop' | awk -F' ' '{ print $4 }';`
root-tail -g 900x450+500+50 -font 8x13 -id ${deskid} -f \
/var/log/messages,white

For this, you do NOT have to start samba for this.

mount -t smbfs -o username="Administrator",password="xxx" //WIN-IP/C$ /mnt/samba/

Same in fstab:

//WIN-IP/C$            /mnt/samba    smbfs           username=Administrator,password=WIN-PASS        0 0

Be sure that both machines are in the same workgroup, then create this /etc/samba/smb.conf:

[global]
workgroup=LAN
browseable = yes
server string=%h (Samba %v)
dns proxy=no
name resolve order=lmhosts wins bcast host
log file=/var/log/samba/log.%m
max log size=1000
syslog=0
panic action=/usr/share/samba/panic-action %d
security=SHARE
encrypt passwords=true
passdb backend=tdbsam guest
obey pam restrictions=yes
#invalid users=root
map to guest=Bad User
passwd program=/usr/bin/passwd %u
passwd chat=*Enter\snew\sUNIX\spassword:* %n\n *Retype\snew\sUNIX\spassword:* %n\n .
client use spnego=no
load printers=no
printing=cups
printcap name=cups
socket options=TCP_NODELAY
display charset=iso8859-1
unix charset=iso8859-1
preserve case=yes
case sensitive=no
short preserve case=yes
os level=20
;   preexec = /bin/mount /cdrom
;   postexec = /bin/umount /cdrom

[public]
path = /home/smashor
public = yes
writable = yes
create mask = 0000
browseable = yes

Win98 computers (and maybe others too) only respond to smbclient when you take the netbios name. Instead of smbclient -L 192.168.0.99 or smbclient -L rentner (name of /etc/hosts) you have to use smbclient -L BLABLA. You get this name with nmblookup -A 192.168.0.99.

For the Fx keys where x is one of 0-9 use this:

bindkey -k kx command

If you have more Fx keys use this:

bindkey -k Fx command (which means 1x)

So to bind F1 to "go to window 1" and to bind F12 to "go to next window" use these lines:

bindkey -k k1 select 1
bindkey -k F2 next

S.M.A.R.T. stands for Self Monitoring Analysis and Reporting Technology - it gives information about the state of the hard disk (e.g. warmth, age, number of errors). To receive the SMART attributes one will need smartmontools (smartctl to get the information, smartd is a background daemon).

First, you have to activate SMART for your disk:

smartctl -s on /dev/hda

Then you can get an overview page:

smartctl -A /dev/hda

Since there is no standard, the developers provide information for hard disks. To see, if your hard disk is in the database, invoke:

smartctl -P show /dev/hda

The column named "UPDATED" indicates how often the attributes are updated. Always means data is updated all the time, while Offline means that the data is updated every hours e.g. You can update at once the "offline attributes":

smartctl -t offline /dev/hda

The column named "TYPE" can have two values: "Pre-Fail" or "Old_age". Old_age attributes are interesting, but they do not indicate a damage. Thus the Pre-Fail attributes exist!

The higher the absolute value of "VALUE", the more sane is your hard disk. "THRESHOLD" gives a lower bound which should not be undergone.

You should regularly look at the "Pre-Fail" values!

You can run some tests:

smartctl -t short /dev/hda
smartctl -t long /dev/hda
smartctl -l selftest /dev/hda

Some manufacturer information can be received with this one:

smartctl -i /dev/hda

By the way, the temperature of current hard disks is about 45 C, the value should not be regularly over this value!

For S-ATA disks, you have to add -d ata at each smartctl command.

You need smartctl -o on /dev/hda to update the attribute values each 4 hours.

To receive mails on faults and errors, write -m email@address.de in your /etc/conf.d/smartd and restart smartd (don't forget to add rc-update add smartd boot)

SPIM - the MIPS simulator - does not work in Gentoo out of the box (or Linux in general, don't know) if you install the package spim. After xspim the window appears, but the subwindows don't show anything!

emerge "< flex-2.5.33"
emerge spim
emerge flex

ssh s_xxxxx@i08fs1.ira.uni-karlsruhe.de

connecting to a remote host using ssh

ssh user@remotehost or
ssh -l user remotehost

invoking a command

ssh user@remotehost "ls -l /tmp/*"
ssh -t user@remotehost mutt (use -t for programs which will acquire the screen control)

secure copy

scp file.tar.gz user@remotehost:/home/foobar/

Options like -r and -p are available here, too.

First, generate a key using ssh-keygen which will create two files in your ~/.ssh directory: identity and identity.pub. The former key is your private key and the latter one your public key! Then copy the public key to the remote host home directory and add the contents of this public key to the .ssh/authorized_keys file: mkdir -p .ssh && cat identity.pub >> .ssh/authorized_keys. You should delete now the public key in the home directory (remote host). Everything is set up now, just connect to the remote host. If there are problems try to set the correct permissions: chmod go-w ~ chmod 700 ~/.ssh (remote host).

Available facilities:

auth auth-priv cron daemon ftp kern lpr mail
mark news syslog user uucp local0 - local7

Available priorities:

debug info notice warn err crit alert emerg

Simple /etc/syslog-ng/syslog-ng.conf

options {
        long_hostnames(off);
        sync(0);

        # The default action of syslog-ng 1.6.0 is to log a STATS line
        # to the file every 10 minutes.  That's pretty ugly after a while.
        # Change it to every 12 hours so you get a nice daily update of
        # how many messages syslog-ng missed (0).
        stats(43200);
};

#
## messages
#
source messages_src { unix-stream("/dev/log"); internal(); };
destination messages_dest { file("/var/log/messages" perm(0644)); };
log { source(messages_src); destination(messages_dest); };

#
## kernel related logs to /var/log/kernel
#
source kernel_src { file("/proc/kmsg"); };
filter kernel_filter { facility(kern) and level(info,notice,warn,err,crit,alert,emerg); };
destination kernel_dest { file("/var/log/kernel"); };
log { source(kernel_src); filter(kernel_filter); destination(kernel_dest); };

# kernel debug output in /var/log/debug
destination debug { file("/var/log/debug"); };
filter f_debug { not facility(auth, authpriv, news, mail); };
log { source(src); filter(f_debug); destination(debug); };

Resources:
[1] Gentoo Doc about syslog-ng

Sometimes I need a secure connection to the university. Just install vpnc and visit this site if you have an encrypted group password.

Afterwards, edit the configuration file /etc/vpnc.conf:

IPSec gateway rz-ipsec.rz.uni-karlsruhe.de
IPSec ID unika
IPSec secret ********
Xauth username username

Be sure that you have selected the kernel module tun (Device Drivers - Network - Universal TUN/TAP) - otherwise you will get an error message saying you do not have /dev/net/tun i.e.

Finally, invoke vpnc and have fun! To check whether you're using the VPN, look at the output of route -n !

echo "export MANPAGER=more" >> ~/.bashrc

If you wanna work in a large code project, you have to know ctags in combination with vim! In your source code directory enter exuberate-ctags *.c *.h which will create the tags file. After that, you can browse through the source code and jump through the definitions of constants, variables and functions using CTRL+]. If you have read enough switch back to your starting point using CTRL+T.

Source Files can grow very rapidly and for a better structure and overview vim has the facility to fold and unfold parts of your code. To make a fold for a visual selection, just press zf. Using zo opens a fold. If you want to close it again, use zc. zO opens all folds while zC closes all. zi toggles all folds.

Foldings are not saved within the .viminfo, but you can save foldings with :mkview and load them with :loadview!

In .gvimrc

color colorscheme

First of all the great IDE-like winmanager plugin. Just extract the winmanager.zip in your ~/.vim directory. The help-file says you also need bufexplorer: then do it again for that one .

emerge vim-spell-de vim-spell-en

After that open vim (or gvim) and read :help spell. Here's a short summary:

:setlocal spell spelllang=de (or de_19 (before the German reform), de_20 (the new one), en_us)
zg store a marked word as a good word into the spell file
zw store a marked word as a wrong word into the spell file
]s go to the next misspelled word
[s go to the previous misspelled word
z= get a list of suggestions for a word marked as wrong

Of course this is only efficient if you do not have to type in these cryptic commands. So here some mappings:

map  :set spelllang=de:set spell!:echo "German spell check: " . strpart("OffOn", 3 ASTERISK &spell, 3)
map  :set spelllang=en_us:set spell!:echo "English spell check: " . strpart("OffOn", 3 ASTERISK &spell, 3)

Don't forget to use CTRL-] and CTRL-t to jump within the help files, the words are marked specially.

If you need help, just type :h word / shortcut / option. For example to get help on the hls option invoke :h 'hls' or to know something about the shortcut CTRL-K type in :h C-K

If you don't know exactly what you're searching for you can type in :h word and then instead of completing the ex-commando with Return try CTRL-D. A list will appear with help sections similar to your request.

Another way to search for inexact patterns is to use some regular expressions: :helpg pattern is your friend!

The names of the keycodes are listed in :h keycodes.

You can open an imported file in a new buffer using CTRL-W-F. The same can be done with gf, but the file is shown in the same buffer as the old file.

Jump to local / global declaration of a variable using gd / gD. You can also jump to system wide declarations using [<Tab>. An alternative is to use CTRL-W-D.

A way to find all occurences of a word under the cursor is to invoke [I - it will open a small buffer with all found lines in files.

Compiling a program using a make command can be done with :make and the use of quickfix: :copen, :cnext, :cclose, :clist

Vim is also good for writing mails - some standard values here should be: tw=0 or 78, wrap, linebreak. And if you want to reformat a text block, then you should probably try gqap - the output is a nice reformatted text block.

Vim can be started with a folder as a parameter - it will list all files within this folder. Type i to get several views of the directory listings. Another command line parameters are -R and -M with open a file in read-only mode respectively modifiable-off mode.

If you really work heavy on some files with foldings, marks etc. and you want to have these things set after a restart of mutt, you have to know these commands: for working with viminfo files use :wviminfo [!] [file] and :rviminfo [!] file. The problem is that with viminfo not everything is stored, e.g. no foldings are saved: to get them use these: :mkview [!] [number] and :ldview number. Even more information is stored with sessions: :mksession [!] [datei]. To load these on vim startup you can pass some arguments: vim -i infofile -S sessionfile.

Another possibily is to export the current settings and mappings into a vimrc file using :mkvimrc [!] [file]. It can be loaded with vim -u file. A special startup is vim -u NONE.

Really helpful is to get the last visual selection with gv.

Ever tried to capitalize a word? Visual select it, type u for lower and U for upper leters - ~ in normal mode reverts a character.

Sometimes marks are a nice thing - you can show the existing ones with :marks, create one with mLETTER, jump to the mark with `LETTER and jump back with ``.

Some shortcuts / things to know:
ZZ save and close buffer

"a3dw delete three words and store them in register a

I vs. gI: insert at the beginning of a line vs. insert at column 0

78i-<ESC> will produce a line of 78 -

i_CTRL-D / i_CTRL-T make one indentation less / more in insertion mode

i_CTRL-V ASCII/UTF-CODE / ga type in CODE / show CODE information for character under cursor

i_CTRL-W / i_CTRL-U delete last word / line you typed in

^ / $ jump to beginning / end of current line

( ) { } jump to previous / next sentence / paragraph

% jump to the next parenthesis

J join the current and the next line into one line

q/ q? popup search history in forward / backward direction

window movements see :help CTRL-W and :help window-moving

i_CTRL-P / i_CTRL-N for completing words

xp swap current with next character - ideal for typos which happen more than often.

Installation is really simple: you need a functional ghc with docs (USE flag doc) and then of course haskellmode.

Configuration is a bit tricky: be sure that you have the following lines in your .vimrc:

 			filetype plugin on
			filetype indent on
			au BufEnter *.hs compiler ghc
			let g:haddock_browser="/usr/bin/firefox"
			let g:haddock_docdir="/mnt/storage/code/haskell/haskell.org/"
		

Open a Haskell file and type in :ExportDocIndex. After that :DocSettings should be possible which indicates that your installation was successful.

Now let's show a few features (see [3] for a full list):

GHC compiler intregration
    _si is used to show information
    _t is used to show the type
    _T is used to add a type declaration
    _ct creates a tags file in the currenct directory
    i__CTRL-X_CTRL-O is used for insertion-mode completion based on GHCi browsing
    _opt shows a list of common ghc options


Haddock integration:
    _? looks up identifier in haddock
    :IDoc {identifier} looks up identifier in haddock
    :MDoc {module} looks up module in haddock
    _. fully qualifies an unqualified identifier
    _i adds import statement for the identifier under the cursor
    _im adds import module statement
    i_CTRL-X_CTRL-U user-defined insert mode name completion based on all names known to the Haddock index, including package names


HPaste integration:
    :HpasteIndex makes a list of current entries from hpaste.org
    :HpastePostNew takes the current buffer and makes a new hpaste.org entry out of it


Resources:

[1] Claus Reinke's Vim Haskell page
[2] Claus Reinke's graphical tour
[3] Claus Reinke's Haskellmode manual

Just install it, create a directory for the output and run webalizer -T logfile inside this new directory. The output is really create and helps finding bottlenecks and failures in your web presence!

Install layman to get access to the haskell overlay. Since it is accessed through darcs which needs haskell to compile and as we do not yet have haskell installed, we download a pre-compiled one and use that (which it is removed after adding the haskell overlay using layman):

emerge subversion layman
layman -a haskell
wget http://www.pps.jussieu.fr/~jch/software/files/darcs-1.0.7-i386-linux.gz
gunzip darcs-1.0.7-i386-linux.gz
mv darcs-1.0.7-i386-linux /usr/bin/darcs
chmod +x usr/bin/darcs
layman -a haskell
rm /usr/bin/darcs

Add these lines to the /etc/make.conf (only if this was your first overlay installation):

#PORTDIR_OVERLAY="/usr/portage/local/layman/haskell"
source /usr/portage/local/layman/make.conf

Now one has to set some USE-flags in /etc/portage/package.use ...

dev-lang/ghc binary
x11-wm/xmonad-darcs extensions

... and some keywords in /etc/portage/package.keywords (if you want to use the latest version, uncomment the darcs lines and comment the appropriate lines):

=dev-lang/ghc-6.6.1
=dev-haskell/mtl-1.1.6.2
dev-haskell/regex-base
dev-haskell/regex-compat
dev-haskell/regex-posix
dev-haskell/cabal
dev-haskell/filepath
dev-util/darcs
dev-haskell/quickcheck
dev-haskell/html
x11-wm/xmonad
#x11-wm/xmonad-darcs
dev-haskell/x11
dev-haskell/x11-extras
#dev-haskell/x11-extras-darcs

Now install ghc and xmonad using emerge -av dev-lang/ghc x11-wm/xmonad-darcs

The main concept: use the us layout and add the special characters to it.

Section "InputDevice"

Identifier "Keyboard1"
Driver "kbd"
# Option "Protocol" "Xqueue"
Option "AutoRepeat" "500 30"
Option "XkbRules" "xorg"
Option "XkbModel" "pc104"
Option "XkbLayout" "us,de"
Option "XkbOptions" "grp:alt_shift_toggle,grp_led:scroll"
Option "XkbVariant" "nodeadkeys"
EndSection

Some notes for this snippet. You can change always between German and US layout by pressing ALT and Shift, but the US layout is the default one. Now we use xmodmap to add our special keys. Put the following in your ~/.xinitrc:

# right alt
xmodmap -e 'keycode 113=Mode_switch'
xmodmap -e 'keycode 38=a A adiaeresis Adiaeresis'
xmodmap -e 'keycode 32=o O odiaeresis Odiaeresis'
xmodmap -e 'keycode 30=u U udiaeresis Udiaeresis'
xmodmap -e 'keycode 26=e E EuroSign EuroSign'
xmodmap -e 'keycode 39=s S ssharp ssharp'

xrdb .Xdefaults

Options:
-d display = server
-DPMS = switches on "Energy Star" functions
+DPMS = switches off "Energy Star"
m acceleration = acceleration for mouse in x
q = informations about current settings

Write the settings in ~/.bashrc

Starting from 2.6.13 config files for this driver are located in /sys/devices/platform/i8042/serio0/serio2.

Press-to-Select allows you to tap the control stick which will simulate a left click.

enable:  # echo -n 1 > /sys/devices/platform/i8042/serio0/serio2/press_to_select
disable: # echo -n 0 > /sys/devices/platform/i8042/serio0/serio2/press_to_select

You can control the sensitivity and the speed of the trackpointer with values between 0 and 255.

# echo -n 120 > /sys/devices/platform/i8042/serio0/serio2/speed 
# echo -n 250 > /sys/devices/platform/i8042/serio0/serio2/sensitivity 

Scrolling while holding mouse button 3: just add these two lines to your existing mouse section in /etc/X11/xorg.conf.

Option          "EmulateWheel"          "on"
Option          "EmulateWheelButton"    "2"

Resources:
[1] ThinkWiki article about trackpoint

Install some Gentoo Xcursor packages like gentoo-xcursors and look into /usr/share/cursors/xorg-x11/ where the themes are installed.

You can choose your theme using ~/.Xdefaults or by creating the /usr/share/cursors/xorg-x11/default/index.theme. The local variant:

Xcursor.theme: THEME_NAME

And the global variant:

Inherits=THEME_NAME

zip -r file.zip dir/*