dktools 1.2.29

SourceForge.net Logo
 

Usage

DkApp-based applications

Options for all dkapp-based applications

The dkapp module in the dklibs library set (see http://dklibs.sourceforge.net) provides preferences managment, localization, logging and other features.
These features a described in detail in http://dklibs.sourceforge.net/dklibs.pdf
In a short summary, the programs below marked as dkapp-based provide the following command line options:

  • -h
    --help
    prints a help text.
  • -v
    --version
    prints version information.
  • -c options
    --configure options
    sets options as permanent options, stored in the preferences system.
  • -C
    --show-configuration
    prints a summary information about permanent options.
  • -r options
    --reset options
    overrides (ignores) the permanent options for this one program invokation. The program starts based on compiled-in defaults and processes options given explicitely on the command line. This is intended for program usage in scripts where we want to have a well-defined startup.
  • -u
    --unconfigure
    removes all permanent options.

kls - ls clone

The kls program lists files and/or directories. Special features are configurable output column order, size calculation for directories and checksum calculation.
The program is invoked by

kls [<options>] [<path(s)>]

The following options can be used:

  • -p string
    --print=string
    specifies the order of output columns. The string can contain the following characters:
    n  file name
    s file size
    t file type
    p permissions
    x message digest (checksum), see the -m option
    l number of links
    c file creation timestamp
    m last modification timestamp
    a last access timestamp
    u file owner UID
    g file group GID
    d device number
    r relative device number
    i inode number
  • -m string
    --message-digest=string
    selects a message digest type, string can be one of
    MD5  RSA Data Security, Inc. MD5 Message-Digest
    SHA-1  SHA-1 message digest as published in FIPS PUB 180-1.
  • -r
    -r-
    --recursive=boolean
    activates/deactivates recursive directory traversal.
  • -f string
    --type=string
    chooses the file types to create output for. The string can contain the following characters:
    f  regular files
    d directories
    p pipes
    c character special devices
    b block special devices
    s sockets
    o all other file types
  • -l number
    --links=number
    -l unlimited
    --links=unlimited
    limits recursive directory traversal to certain depth of symbolic links.
  • -d string
    --directory=string
    specifies what to do for directory names specified in command line arguments. The string can consist of the following characters:
    d  show only information about the directory itself
    c show information about the directory contents
  • -w number
    --width=number
    specifies the width of the size column in characters. Use 0 for automatic selection, -1 for no fixed width or positive numbers to specify a width.
  • -f
    -f-
    --stay-on-filesystem=boolean
    can be used to limit the kls program to stay on the original file system.
  • -s
    -s-
    --summary=boolean
    enables or disables printing of a summary.

ksort - sort clone

The ksort program reads input from a file (or standard input), sorts the lines and prints output to a file (or standard output).
Lines can be sorted by alphabetical comparisons, decimal number comparisons or floating point number comparisons.
The following options are available:

  • -o string
    specifies the output file name.
  • -l number
    --line-length=number
    specifies the maximum line length to expect in input.
  • -b
    -b-
    --ignore-blanks=boolean
    turns ignoring of leading spaces on or off.
  • -d
    -d-
    --decimal=boolean
    specifies whether or not to compare by leading decimal numbers.
  • -f
    -f-
    --float=boolean
    specifies whether or not to compare by leading floating point number.
    You can use either -d or -f, not both at once.
  • -i
    -i-
    --inverted=boolean
    enables or disables inverted output.
  • -m
    -m-
    --merge=boolean
    enables or disables merging of unique lines.

createp - password suggestor

The createp program suggests passwords. Different account types can be prepared (i.e. for administrators and normal users), passwords for different account types have different complexity.
Account types can be set up by

createp -c -t <type> -o default -l <range> -d <range> -s <range> -k <range> -a <prng> [-p]
createp -c -t <type> -o <mode> -l <range> -a <prng> [-p]

The type argument is the account type name to set up, the -l option specifies the password length, the -d option the number of digits in the password, the -s argument the number of special characters and the -k argument the number of uppercase characters.
The -a argument chooses a PRNG, choose either ``all'' or one or multiple (separated by comma) from the list ``openssl'', ``random'', ``rand48'', ``rand''.
The -o argument specified the output mode, choose one from ``default'', ``hex'', ``ascii85'' or ``charset''. The default mode allows to specifiy the number of digits, special and uppercase characters explicitly. The hex mode produces hexadecimal output containing digits and the characters ``a''...``f''. The ascii85 mode creates password containing characters in the range 33...127 (decimal). The charset mode requires an additional argument to specify the set of usable characters. In this argument any lowercase character represents all lowercase characters, any uppercase character represents all uppercase characters and any digit represents all digits.
 
Examples:

createp -c -t admin -o ascii85 -l 8-10 -a openssl
createp -c -t wpa-key-1 -o hex -l 64 -a all -p
createp -c -t wpa-key-2 -o ascii85 -l 32 -a all -p
createp -c -t test -o charset a0 -l 8-10 -a all
createp -c -t user -o default -l 6 -d 2 -s 0 -k 0 -a all

creates the account type ``admin'', ``wpa-key-1'', ``wpa-key-2'', ``test'' and ``user''. The ``user'' type was the last type configured, so it is the default type.
 
Run

createp <username(s)>
createp -t <type> <username(s)>

i.e.

createp joe jim jack
createp -t admin richard

to create passwords.
In dktools-1.2.26 and above createp can use several PRNGs depending on the system and the libraries installed on the system. The program checks which of the PRNGs ``openssl'', ``random'', ``rand48'' and ``rand'' are allowed by the -a option (in the order shown here). It uses the first allowed PRNG which can be seeded.
The recommended PRNG is the PRNG provided by the OpenSSL library, the other PRNGs are more predictable.

dksavepw - save password to file

The dksavepw application reads a password from terminal and saves it to file. Keyboard echo is disabled while reading the password.
If the program fails to disable keyboard echo a message is printed and the user is asked whether or not to continue reading the password. 

dksavepw [<options>] <outputfile>

The following options can be used:

  • -f
    forces operation. If the program fails to disable keyboard echo no warning is printed and no question asked.
  • -n
    adds a newline after the password in the file. Some programs need password files without trailing newlines, other programs read password files linewise and remove the last character simply relying on having a newline to delete.

dklogger - logger clone

The dklogger program is logger clone for Windows. One can send messages by typing

dklogger [-p <feature>.<priority> ] <message>

i.e.

dklogger -p auth.notice LOGIN jim

Syslog hosts are specified in preferences, the names are

/syslog/host/<nnn>/destination

where nnn is a three-digit number in the range 000-015. An example for a preference name is "/syslog/host/000/destination".
The preference value is a string consisting of host and port number, separated by a colon. The host can be specified as host name or IP address.
The available features are "auth", "authpriv", "cron", "daemon", "ftp", "kern", "lpr", "mail", "mark", "news", "security", "syslog", "user", "uucp", "local0", "local1", "local2", "local3", "local4", "local5", "local6", and "local7".
The priorities are "alert", "crit", "debug", "emerg", "error", "info", "none", "notice", "panic" and "warning".

showhex - show binary files in hexadecimal notation

The showhex program shows the contents of a binary file in hexadecimal notation. It can be invoked by

showhex [ <options> ] [ <file(s)> ]

If no filename is specified the program runs as filter.
The following options are available:

  • -t
    -t-
    --text=boolean
    enables/disable printing of text beside the hexadecimal data.
  • -a
    -a-
    --address=boolean
    enables/disables printing addresses at the beginning of each line.
  • -b number
    --buffer=number
    specifies the buffer size to use for the input buffer.

rndbytes - random byte block generator

The rndbytes program writes pseudo random data to a file. The syntax is
 
  rndbytes options outputfile
 
The following options can be used:

  • -b size
    specifies the block size.
  • -n number
    specifies the number of blocks. Use 0 for an ``endless'' sequence of blocks.
  • -a prngs
    specifies the list of allowed PRNGs. Use either ``all'' or one or multiple (comma-separated) from ``openssl'', ``random'', ``rand48'', ``rand''.

To erase an entire disk containing confidential data run

rndbytes -b 4096 -n 0 -a all | dd of=/dev/sda bs=4096

multiple times.
Run

dd if=/dev/zero of=/dev/sda bs=512

once, so a software installation program or a disk formatting program finds a ``new'' empty disk.
Note: The /dev/sda device name was used as an example. Make sure to provide the correct device name of the disk you want to erase.
Note: Overwriting an entire disk using pseudo-random data may take a very long time. The edbp program can be used to fill an entire disk with static patterns. Even if you run edbp multiple times to use different patterns it is faster than using rndbytes once for pseudo-random patterns.

text2lat - text to LaTeX source converter

The text2lat program converts a text file into a corresponding LaTeX source file. The program can also run as a filter.
The syntax is

text2lat [<options>] [<inputfile> [<outputfile>]]

The following options can be used:

  • -e encoding
    specifies the input encoding.
    The following encodings can be used:
    • latin-1
      Normal ISO-LATIN-8859-1 encoding.
    • utf-8
      UTF-8 encoded text.
    • automatic (default)
      the program has to inspect the LANG environment variable for an appearance of ``.UTF-8''.
  • -n
    -n- (default)
    to enable/disable the conversion of newlines to \\ sequences.
  • -s
    -s- (default)
    to enable/disable the conversion of spaces to ~.

A typical usage scenario is to write software documentation using Vim or other vi clones. Imagine you want to typeset text like

Install the library to
C:\Program Files\My Software
and use the function
mylib_do_something()
to do something.

Normally you would type

Install the library to
C:\textbackslash{}Program Files\textbackslash{}My Software
and use the function
mylib\textunderscore{}do\textunderscore{}something()
to do something.

in the editor.
Using Vim or another vi clone you can type the original text, find the line numbers where the text section starts and ends and run

:11,15!text2lat

This removes the contents of lines 11...15 from the editor buffer, feeds the text as input to text2lat and inserts the programs output into the editor buffer.

text2htm - text to HTML converter

The text2htm program converts text (plain text or UTF-8 encoded text) to HTML. The program can either convert files or run as a filter.
The syntax is

text2htm [<options>] [<inputfile> [<outputfile>]]

The following options can be used:

  • -e encoding
    specifies the input encoding, one of:
    • latin-1
      ISO-LATIN-8859-1 encoding
    • utf-8
      UTF-8 encoding
    • automatic
      Inspect the LANG environment variable whether or not it ends on ``.UTF-8''.
  • -t title
    specifies the title for the HTML document.
  • -s style
    specifies the URL of the CSS style sheet the doucment uses.
  • -f
    -f-
    decides whether to produce a full HTML document or to just recode the characters.
  • -g
    -g-
    decides whether or not to add template comments.
  • -b
    -b-
    specifies whether or not to convert newlines to <br> tags.

Simple programs

tape - backup media management

The tape program can be used to manage backup tapes (or other backup media). It uses a ten-tape-rotation-scheme using each media 20 times for a 200 day period (200 backups).
A tape set is a set of 10 tapes. This set is to used a backup set (group of directories backed up together).
When a backup is to be done we have to ask for the current tape set first. The next step is to ask for the tape number within the set.
After doing the backup we have to confirm the tape and the tape set.
The following program invokations can be used:

  • tape sn tape-set-file
    asks for the current set name and prints this name to standard output.
    The tape-set-file must be prepared, see below.
  • tape tn tape-number-file
    asks for the current tape number in the set and prints this to standard output. The tape-number-file is created automatically by the tape program, it contains information about the last tape used and the timestamps for all the tapes.
  • tape tc tape-number-file
    confirms that the current tape was written successfully.
  • type sc tape-set-file
    confirms that the current set was backed up sucessfully.

Example:
Imagine a typical NFS server. We want to back up user data very often, this is backup set "A". For disaster recovery we want to back up the "/etc" directory and some other data, this is backup set "B".
We write a file "/etc/backup/sets", contents is

A A A A B

This means we want to back up "A" four times a week, "B" only once.
The backup script may look like

set BCKDIR = /etc/backup
set theset = `tape sn $BCKDIR/sets`
set thetape = `tape tn $BCKDR/tape.$theset`
echo Starting backup to $theset $thetape .
if ( { write_backup_to_tape $theset } ) then
  tape tc $BCKDIR/tape.$theset
  tape sc $BCKDIR/sets
endif
set theset = `tape sn $BCKDIR/sets`
set thetape = `tape tn $BCKDR/tape.$theset`
echo Please insert tape $theset $thetape for next backup.

uid - search for free user ID in a given range

The uid program searches for the first free (unused) user ID in a given range.
The syntax is

uid <min> <max>

i.e.

uid 3111 4000

searches for the first free UID in the range 3111...4000.

blksize - change block size of data stream

Blksize changes the block size of a data stream, input is read from standard input, output is written to standard output. The syntax is

blksize
blksize <size>

The intended use is restoring disk partitions (file systems) from compressed images.
To save a partition use

dd if=/dev/hda3 bs=512 | gzip -9> image.dat

To restore the image use

gzip -dc image.dat | blksize 512 | dd of=/dev/hda3 bs=512

You can use more complicated mechanisms, i.e. use netcat to save the compressed image on a different computer.

filtmsql - mysqldump output formatter

The filtmsql application reformats output produced by mysqldump to avoid overlong lines. Line breaks are inserted before the second table entry and all following entries of each table.
Use

filtmsql
filtmsql <inputfile>
filtmsql <inputfile> <outputfile>

to run the program. Standard input and standard output are used if no file names are specified.

cona - correct file names

Windows operating systems allow to use spaces and ligatures (i.e. german umlauts) in file names.
Samba also accepts these names.
When using scripts for doing backup spaces and umlauts may cause problems. The cona program corrects all filenames in a given directory and all subdirectories to have no spaces and no umlauts.
The program invokation syntax is

cona <directory>

i.e.

cona /export/home

echo2lat - text to LaTeX source converter

The echo2lat program converts the command line arguments to the corresponding LaTeX encoding.
Imagine you want to write a documentation for some software. You want to typeset ``Install the software to C:\Program Files\My Software'' in LaTeX, you would have to write ``Install the software to C:\textbackslash{}Program Files\textbackslash{}My Software''. To avoid typing all the ``\textbackslash{}'' manually, choose an option like ``insert a programs output'' from the editors menu. In the dialog asking for a command, type

echo2lat C:\Program Files\My Software

As an example, in Vim (or other vi clones) use the command

:r!echo2lat C:\Program Files\My Software

klammern - paranthesis counter for source code

The klammern program is a simple paranthesis counter for C code. It counts opening and closing brackets, curly brackets... ("()[]{}") and prints a report. This can be used to track down compiler errors in large source files.
Modern editors - i.e. Vim - allow to find the matching paranthesis of a pair by simply pressing one key. The klammern program is for those without such an editor.
The program invokation syntax is:

klammern <inputfile>

i.e.

klammern mysource.c

addctrld - add finishing Ctrl-D to print job

The addctrld program simply pipes input to output and adds a Ctrl-D at the end.
This is usefull to indicate the end of a PostScript print job.
The program invokation syntax is

addctrld [ <inputfile> [ <outputfile> ] ]

tabrep - replace tabulators by sequences of spaces

Source code often contains tabulators. When inserting source code into a "verbatim"-environment in LaTeX, tabulators are ignored. So we must replace tabulators by an appropriate sequence of spaces.
The tabrep program does this, the syntax is

tabrep [ <input-file> [ <output-file> ] ]

Instead of using tabrep you might decide to use the listings LaTeX package to include program source code in *.tex files.

dkpwd - pwd clone (for windows)

The dkpwd program is a pwd clone for Windows users. It prints the current working directory to the standard output.
Syntax is:

dkpwd [ -b ] [ -p ]

The -b option replaces all backslashes by forward slashes.
The -p option prints the directory without a trailing newline. This can be used to generate prompt signs (i.e. in tcsh on Windows systems).

edbp - byte pattern block generator

The edbp program creates an ``infinite'' sequence of blocks containing a byte pattern. The program runs until stopped by a signal, i.e. SIGPIPE. It can be used to overwrite entire disks.
The command line syntax is:
 
edbp pattern-number block-size
 
The pattern number must be in the range 0...7.
The following patterns are used:
 

pattern numberpattern
00x00
10xFF
20x0F
30xF0
40x33
50xCC
60x55
70xAA

 

To erase the sdb disk for example, run

edbp 7 4096 | dd of=/dev/sdb bs=4096
edbp 6 4096 | dd of=/dev/sdb bs=4096
edbp 5 4096 | dd of=/dev/sdb bs=4096
edbp 4 4096 | dd of=/dev/sdb bs=4096
edbp 3 4096 | dd of=/dev/sdb bs=4096
edbp 2 4096 | dd of=/dev/sdb bs=4096
edbp 1 4096 | dd of=/dev/sdb bs=4096
dd if=/dev/zero of=/dev/sdb bs=4096

For increased security you might want to use random patterns instead of static patterns. For this purpose use the rndbytes program instead.
Note: Using pseudo-random patterns instead of static patterns is probably more secure but takes a significantly longer time.

Unix/Linux daemons

rshdown - remote shutdown on power fail

The rshdown program is a daemon for Unix/Linux. If one UPS is used for powering multiple computers, only one computer (the shutdown master) has a control connection to the UPS. If a power fail shutdown is necessary this computer must notify the other computers (shutdown slaves). The shutdown master uses netcat to send an UDP data packet to the slaves. Multiple UDP packets can be sent to each slave, the first valid packet shuts down the slave, further packets are ignored. On the slaves the rshdown daemon waits for UDP data packets. If a shutdown-related data packet is received the rshdown daemon checks the packet for integrity and shuts down the slave if necessary.

Create a datafile containing up to 1500 bytes of random data (you can use the rndbytes program to do so).
Use a secure mechanism (floppy or scp) to transfer the data packet file between master and slaves. Make sure the data packet file is only readable by user root, other users must not have any permissions to it.

On the shutdown master modify the runlevel scripts as follows: On each boot create an indicator file, i.e. /var/run/rshdown/rshdown.txt. The contents of this file does not matter, only presence is important. On each shutdown check for the presence of the indicator file. If the file exists, use netcat to send the contents of the data packet file as UDP datagrams to the slaves. Make sure netcat binds to a specified local port before sending datagrams, use a free privileged port (1...1023). In the example below we use port 2/UDP. In /etc/motd write some text to remind the user root to remove the indicator file before doing a scheduled (not power-related) shutdown or reboot.

Configure each of the shutdown slaves as follows:

  • Transfer the data packet file from the master to the slave, use -- as example -- file name /etc/rshdown-secret.
  • Check permissions to the data packet file, only user root must have read access to it.
  • Create a fast shutdown script, i.e. /etc/powerfail-fast: 
    /etc/powerfail-full > /dev/null 2>&1 &
    and make it executable. This script must return immediately, it can start longer processes in the background.
  • Create the full shutdown script, i.e. /etc/powerfail-full: 
    /usr/sbin/poweroff
    Make the file executable.
  • Decide, which local port to use. You should use a free (unused) privileged port (1...1023). In the example below we use port 2/UDP.
  • Create the configuration file $(prefix)/etc/rshdown.conf 
    sender-ip	192.168.1.2
sender-port	2
receiver-ip	192.168.1.3
receiver-port	2
file		/etc/rshdown-secret
action		/etc/powerfail-fast
  • Modify the system startup scripts to make sure rshdown is run after each reboot.
  • Optionally create a static ARP entry for the master (relationship between MAC- and IP-address) on each slave.
Dirk Krause
Schmalkalden (Germany), 2008-11-01
http://dktools.sourceforge.net