lindeman.nu

I’ve wrote this shell script that uses an SMS to post to Twitter. It is ment to be used as an Eventhander for SMS Server Tools, but it’s simple enough for stand alone usage with simple text files.

If used as an Eventhandler the arguments are “RECEIVED” and the path to the SMS file, add “eventhandler=/path/to/sms2twitter.sh” to your smsd.conf. Without SMS Server Tools you can also use a simple file for posting to twitter:
Create a file starting like this:

Then call this script with the path to your file:

You will need to create a sqlite3 DB with this schema:

For every Phone that you allow to post Tweets thru SMS you need to run:

Lees meer…

For Postmasters: easy way to read a mailbox from your XMail spool. If you have a CDB file (see this post), it is also capable of looking up aliases.

This scripts creates a Constant DataBase file. We use this to interact with the qpsmtpd plugin check_validrcptto_cdb.

Here is a bash script that uses kdialog to present a password prompt and tries to mount a crypt device. It interacts with the (k)ubuntu “cryptsetup” package.

quick walk-through for compiling PostgreSQL with tsearch2 full text extension and Dutch Snowball stemmer on Debian Etch:

If everything went well you are now in a PostgreSQL prompt.

We will now update ans insert some stuff:

select to_tsvector('dutch', 'ik ga naar school');    to_tsvector
------------------
'ga':2 'schol':4
(1 row)

I’ve created php5 and mysql5 documentation packages for Quanta. They include an installer and come in several languages. I’m working on PHP:PEAR Documentation. Download here …

Today I had this problem with a MySQL/php application:

Illegal mix of collations (latin1_bin,IMPLICIT) and (latin1_swedish_ci,IMPLICIT)

I’ve created this loop to solve the problem, because the server contains several database with lots of tables:

Often I want to copy/paste my WAN IP address, here is a shellscript for it (warning: kde only, requires wget and sed, no error detecting whatsoever!)

We use Spamassassin as anti spam tool with Bayes learning. If I want to feed spam massages to the sa-learn tool, I have to get individual messages from my Thunderbird Mailfolder which stores all mails together in a text file.

First I move alle spam massages to the Thunderbird Junk mail. Thunderbird has a build in learning tool to fight spam, so most spam messages are moved to the junk folder automagically. Make sure there are no messages in this folder that are already marked as spam by Spamassassin!
Split a Thundebird mailfolder into individual Mails:

Sometimes you have to write lots of CD’s/DVD’s of even more files. This scripts loops through files in a target directory and creates new “virtual disks” and places symbolic links in it to the found file. If a “virtual disk” is full it will create another one. You can define your own pattern for the disknames.

#!/bin/bash

# sizes in Megabytes
CDSIZE=750
DVDSIZE=4380

if [ "$(basename $0)" == "make-dvd" ]; then
let "DISKSIZE=$DVDSIZE*1024*1024"
else
let "DISKSIZE=$CDSIZE*1024*1024"
fi

VERBOSE=1
usage ()
{
echo "usage: $0 [OPTIONS] source target, where options can be:"
echo "  -s START      number of 1st disk"
echo "  -v            print usefull messages"
echo "  -V            print usefull and less usefull messages"
echo "  -d BYTES      disksize of the writable media (default $DISKSIZE)"
echo "  -r            puts all symlinks in corresponding directory"
echo "                without this option all symbolic links are placed directly under disk"
echo "  -p            pattern to use for diskname (default DISK_%03d), see man printf"
echo "                example pattern: \"MY_CD_%03d\" for MY_CD_001, MY_CD_002, etc ..."
if [ "$1" != "" ]; then
echo "ERROR: $1"
exit 9
fi
exit 8
}

myprint ()
{
if [ "$VERBOSE" == "$1" ]; then
echo $2;
fi
}

integertest ()
{
stringTest=$(echo $1 | sed 's/[0-9]+//g');
if [ ${#stringTest} -gt 0 ]; then
usage "argument for $2 must be an integer"
fi

}

if [ $# -lt 2 ]; then usage; fi

START=1
VERBOSE=0
INTREE=0
PATTERN="DISK_%03d"

while getopts ":s:t:p:rvVd:" Option; do
OPTARG=$(echo $OPTARG | sed 's/^W+//')
case $Option in
s ) START=$OPTARG;;
v ) VERBOSE=1;;
V ) VERBOSE=2;;
r ) INTREE=1;;
d ) DISKSIZE=$OPTARG;;
p ) PATTERN=$OPTARG;;
* ) usage "unkown option";;
esac
done
shift $(($OPTIND - 1))

SOURCE=$(cd "$1" && pwd)
TARGET=$(cd "$2" && pwd)
integertest $START "-s"
integertest $DISKSIZE "-d"

DISKNAME=$(printf $PATTERN $START)
if [ $? -gt 0 ]; then usage "wrong argument for pattern"; fi

if [ ! -d "$SOURCE" ]; then usage "source \"$SOURCE\" is not a directory"; fi
if [ ! -d "$TARGET" ]; then usage "target \"$TARGET\" is not a directory"; fi
if [ ! -w "$TARGET" ]; then usage "target directory \"$TARGET\" is not writable"; fi

OLDIFS=$IFS
IFS=:
CURRENTSIZE=0

FIRSTRUN=1
echo -n "fetching filelist from $SOURCE ... "
for file in $(find -L $SOURCE -type f -printf "%P$IFS"); do
if [ "$FIRSTRUN" == "1" ]; then
echo "done"
myprint 1 "making 1st disk $DISKNAME"
if [ -d "$TARGET/$DISKNAME" ]; then usage "can not make disk $TARGET/$DISKNAME"; fi
mkdir -p $TARGET/$DISKNAME
FIRSTRUN=0
fi

filesize=$(du -b "$SOURCE/$file" 2> /dev/null | sed 's/W+.*//')
myprint 2 $(printf "%s [%10d - %10d] %s" $DISKNAME $CURRENTSIZE $filesize $file 2>/dev/null)
if [ "$filesize" == "" ]; then echo "can not get filesize of file $SOURCE/$file, skipping it"; continue; fi
let "newsize=$CURRENTSIZE+$filesize"
if [ $newsize -gt $DISKSIZE ]; then
let "START=START+1"
DISKNAME=$(printf $PATTERN $START)
myprint 1 "disk full, making new one: $DISKNAME"
CURRENTSIZE=$filesize
if [ -d "$TARGET/$DISKNAME" ]; then usage "can not make disk $TARGET/$DISKNAME"; fi
mkdir -p $TARGET/$DISKNAME
else
let "CURRENTSIZE=$CURRENTSIZE+$filesize"
fi
DIRNAME=$(dirname "$file")
FILENAME=$(basename "$file")
if [ "$INTREE" == "1" ]; then
if [ ! -d "$TARGET/$DISKNAME/$DIRNAME" ]; then mkdir -p "$TARGET/$DISKNAME/$DIRNAME"; fi
LINKTARGET="$TARGET/$DISKNAME/$DIRNAME/$FILENAME"
else
LINKTARGET="$TARGET/$DISKNAME/$FILENAME"
fi
ln -s "$SOURCE/$file" "$LINKTARGET"
if [ $? -gt 0 ]; then
let "CURRENTSIZE=$CURRENTSIZE-$filesize"
fi
done
IFS=$OLDIFS