![[Blue Ribbon Campaign icon]](http://br.eff.org/br.png)
last modified: Tuesday, 23-Feb-2010 15:42:11 CET
Lists IP-address(es) of the given or else all available network devices.
#!/bin/bash
/sbin/ifconfig $1 | grep inet | awk '{print $2}' | sed 's/^addr://g'
At least since Mozilla RC3, starting mozilla again if an instance of it is already running,
will result in the Profile Manager's window asking you to choose a profile or create a new one.
Just trying to use the current one will not work.
This is especially a quite boring behaviour if the browser is started from another application
(for example by double-clicking on a link within my mail client Sylpheed).
I wrote the following script which uses the -remote option previously known from Netscape 4.x in order to get around this problem. Just always use this script to start mozilla ... and everything's fine. :)
#!/bin/sh
#
# /opt/mozilla_start-this_max
#
# assuming we get a URL as parameter $1, we either start mozilla and open the URL,
# or we just open a new window using an already running mozilla-process.
#
# for further details check out
# /opt/mozilla/mozilla --help
# and
# http://www.mozilla.org/unix/remote.html
#
if /bin/ps ax | /bin/grep mozilla-bin | /bin/grep -v grep > /dev/null; then
# an instance of mozilla is already running, just open a new window
/opt/mozilla/mozilla -remote "openurl($1, new-window)"
else
# no instance of mozilla present, start totally new.
/opt/mozilla/mozilla $1
fi
A cooler startup-script by August Mayer:
#!/bin/bash
if mozilla -remote "ping()"; then
mozilla -remote "xfeDoCommand(openBrowser)";
if [ "x$*" != "x" ]; then
mozilla -remote "openURL($*)"
fi
else
mozilla -splash $* &
fi
/root/.fetchmailrc # as one line: poll localhost port 10110 protocol pop3 username pop-server-username is local-username password "xxxxxxx" fetchall preconnect "/usr/bin/ssh -f -C -x -L 10110:pop.server.net:110 ssh-server-username@ssh-server /bin/sleep 10" /etc/crontab [...] */10 * * * * root /usr/bin/test -f /home/max/nomails || /usr/bin/fetchmail -s -f /root/.fetchmailrc
As I continuously forget how to build for-loops using the BASH, here are two ways to do so...
for i in `/usr/bin/seq 1 10`
do
/bin/echo $i
done
for (( i=11; i<=20; i++ ))
do
/bin/echo $i
done
As can be found in the BASH Programming - Introduction HOWTO, here are the redirection rules:
Redirect stdout to file: command > stdout.txt Redirect stderr to file: command 2> stderr.txt Redirect both stdout and stderr to file: command &> stdout-and-stderr.txt Redirect stdout to stderr: command 1>&2 Redirect stderr to stdout: command 2>&1
Thus, you can easily grep on the stderr by redirecting it to stdout and performing a normal grep, something like: command 2>&1 | grep special-errors.
If you are placed behind an HTTP-proxy (with or without user-authentication), you can not simply
use ssh to connect to an outside server. Everything else but HTTP is locked.
If for some reason you desperately need to get out, do the following:
server (outside the proxy): apt-get install httptunnel hts -F localhost:22 1234 client (inside the proxy): apt-get install httptunnel htc -F 5678 -A proxyuser:proxypwd -P proxy-server remote-host:1234 ssh -p 5678 localhost
With this, the local port "5678" is tunnelled to "remote-host's" port "22" via an HTTP-protocol-based
"connection" between the HTTP-Tunnel-Client and the HTTP-Tunnel-Server (port "1234").
SSH connects to localhost:5678 and reaches remote-host:22 (in this configuration).
From time to time, security-updates are necessary. Once installed, every program started
afterwards uses the new binaries.
If the security-update covers libraries, it is necessary to restart all programs which use
these libraries. Just doing an "apt-get update && apt-get upgrade -u" might
not suffice in these cases!
Especially with servers (which tend to be running 24/7, or at least SHOULD do so :) ...),
it is important to know if and which programs/services are still using such old (and already
replaced) libraries. And that's exactly what
check-old-libraries.sh does: find these programs.
#!/bin/sh
#
# README:
# Library-Security-Updates via "apt-get update && apt-get upgrade"
# replace the old libraries with the new ones.
# Programs which continue running (e.g. services) still use the
# old ones. They can be tracked down by this script, a mail is
# sent to admin@subnet.at to have him realize the situation.
# Sample output:
# Restart "sshd", it uses the old library "/usr/lib/i586/libcrypto.so.0.9.6.dpkg-new".
#
# INSTALL:
# *) Save this script as /root/bin/check-old-libraries.sh
# *) Create a cron-job to run it once a day
#
# LICENSING:
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version.
# This program is distributed in the hope that it will be useful, but
# WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
# See the GNU General Public License for more details:
# http://www.fsf.org/licensing/licenses/gpl.txt
# If you cannot access the GNU General Public License via the link
# above, write to the Free Software Foundation, Inc., 59 Temple Place,
# Suite 330, Boston, MA 02111-1307 USA
#
# Copyright 2003 Markus Amersdorfer
lsof +L1 | grep dpkg-new$ | awk '{print $1 " " $9}' | sort | uniq | awk '{print "Restart \"" $1 "\", it uses the old library \"" $2 "\"."}'
I've recently patched action.c from Sylpheed-0.9.0:
Creating the diff file: diff --unified --recursive --new-file sylpheed-0.9.0_ORIGINAL/ sylpheed-0.9.0 > actions-patch_sylpheed-0.9.0 Patching the original code (with "./sylpheed-0.9.0/" being the original code now): patch -p0 < actions-patch_sylpheed-0.9.0
search and replace:
Once in current line only: s/pattern/newpattern/
All occurences in current line: s/pattern/newpattern/g
Only first occurence in every line in the whole file: %s/pattern/newpattern/
All occurences in the whole file: %s/pattern/newpattern/g
Add-on: % equals 1,$ which is the range "line 1 to end of file"
Once you got the source-code (e.g. from one of the mirrors of kernel.org), unpack it to /usr/src/ and create a symlink /usr/src/linux -> /usr/src/linux-2.4.20 (or similar). If you run Debian GNU/Linux, best use kernel-package. Otherwise, run:
cd /usr/src/linux make xconfig make dep clean bzImage modules modules_install cp System.map /boot/System.map /boot/System.map-2.4.20 cp arch/i386/boot/bzImage /boot/vmlinuz-2.4.20 cd /boot/ ln -ivs System.map-2.4.20 System.map ln -ivs vmlinuz-2.4.20 vmlinuz
Edit your /etc/lilo.conf accordingly and run lilo (or edit /boot/grub/menu.lst if you use GRUB).
In order to get a human-readable ASCII-printout of the network traffic to a certain machine,
use the call "tcpdump -n -X -s 1500 host 10.0.0.2".
(-X enables ASCII, -s 1500 sets the maximum length of a packet's data to be shown.
With the default of 68 you won't see the packet's content/payload.)
To print a file's content starting with a specific line, e.g. the 200th, run:
awk '(FNR >= 16) { print }' $file
Suppose you have a file "my-files.txt" holding a list of files, each of them in a single line. To execute a command for each of them (e.g. to untar all those .gz files):
Using list of .tar.gz-files in "my-files.txt": cat my-files.txt | xargs -n 1 tar xzvf Directly in the directory (with "my-files.txt"): echo *.tar.gz | xargs -n 1 tar xzvf
If you want to use that list of files as an indicator for "tar" to tell it which files to extract or archive, use tar's option "-T":
tar cvzf my-tarball.tar.gz -T my-files.txt
In order to rename all *.xxx files to *.yyy, you may either run "rename s/\.xxx$/\.yyy/ *", or you
may experiment a little with sed and pipes: "ls *.xxx | sed 's/\(.*\)\.xxx$/mv & \1.yyy/' | sh".
(Run "man sed" and "/\\1" to search for the explanation of "& \1" in the last command...)