Contents |
The kdb command provide ways to manipulate the Elektra keys database keys.
The subcommands implemented are very similar to regular UNIX commands like ls, and rm, specially in their output and options. Subcommands
Get the value from the specified key. Accepts options: -d, -l, -f, -s
Set the value to the specified key. Accepts options: -c, -t, -m, -b
As the ls(1) command, list key names for the specified key, or children keys, if specified a folder key. The -v argument will make it show also the values of each key. The -d (descriptive) will make it show the comment, key name and its value, as you are watching a plain text file. Accepts options: -x, -d, -l, -f, -v, -R
Creates a key that is a symbolic links to another key.
Move, or renames a key. Currently it can't move keys across different filesystems.
As the rm(1) command, removes the key specified.
A very powerfull subcommand that lets you edit an XML representation of the keys. The parameters it accepts is usually a parent key, so its child keys will be gathered. Can be used with the -R flag to work recursively. The editor used is the one set in the $EDITOR environment variable, or vi. After editing the keys, kdb edit will analyze them and commit only the changed keys, remove the keys removed, and add the keys added.
Export a subtree of keys to XML. If no subtree is defined right after the export command, system and current user trees will be exported. Output is written to standard output. The output encoding will allways be UTF-8, regardeless of your system encoding. UTF-8 is the most universal charset you can get when exchanging data between multiple systems. Accepts -f. import, load,
Import an XML representation of keys and save it to the keys database. If no filename is passed right after the import command, standard input is used.
Monitor a key for some value change. It will block your command line until a change in the key value is detected, then return its new value.
Causes to work recursively. In ls, will list recursively.
Makes ls output an XML representation of the keys, instead of an ls-compatible output.
Causes to display long results. With ls, will generate lists similar to ls -l. With get, will show also the key name.
Causes ls to display also inactive keys. Generate lists similar to ls -a. Inactive keys are keys which basename begins with a '.' (dot). An example of inactive key: system/sw/XFree/current/Monitor/.Monitor1
Causes to work with full key names. A full key name makes sense only on user/* keys, and differentiate from the regular key names in specifying the owner user. If the current user is someuser, the user/some/key full name is user:someuser/some/key. Makes effect in ls, export and get subcommands.
Causes get to work descriptivelly. When requesting a key it will show the comment, key name and its value in a fancy format
Causes get to be more friendly to Shell scripts. For example, when requesting user/env/env2/PATH, the output will be PATH="the value", that is, only the basename of the key will be showed and the value will be surrounded by ' " '.
When setting a key's value, you can specify the type with this switch. Currently accepted types are string for plain text, bin for binary as-is values, dir to create folder keys and link to create symbolic links between keys. Plain text are always stored as UTF-8(7) in Elektra, regardeless of your current encoding ($LANG). If you want to force a value to be stored without the UTF-8(7) encoding (a bad idea), you can set it as binary. Binary values should be avoided, because they are black boxes for system administrators.
Set the key value as the content of file filename. This option is more usefull when setting binary keys.
For the set command. Will set the key access permission to mode, which must be an octal number as for chmod(1).
Create the key with uid user ID. It can be a user name or a uid number.
Create the key with gid group ID. It can be a group name or a gid number
When setting keys, you can use this argument to set a descriptive comment for it. This comment is exactly as a comment in a plain text configuration file. The comment is stored as UTF-8(7) regardeless of your current encoding ($LANG).
With the ls subcommand, will make it show also the value stored in the key.
With the set subcommand, everything after it will be considered the value, even text with dashes (-).
bash$ kdb set -c "My first key" user/example/key "Some nice value"
bash$ kdb set user:luciana/example/key -- "Some - nice - value with dashes"
bash# KDB_ROOT=user:http/sw/httpd kdb set -u nobody -g http key "Some value"
bash$ kdb set -b image.png -t bin user/example/binaryKey
bash$ kdb set -b file.txt user/example/regularKey
bash# kdb set -t link system/sw/XFree/current system/sw/XFree/handmade
bash$ KDB_ROOT=user/example kdb get key
bash$ KDB_BACKEND=gconf kdb get user/sw/gnome-terminal/global/active_encodings
bash$ kdb ls -laR user:valeria
bash$ kdb ls -lR system/sw/xorg/current
bash$ KDB_ROOT=system/sw kdb ls -lR xorg
bash$ KDB_BACKEND=fstab kdb ls -Rv system/filesystems
bash# kdb ln system/sw/xorg/handmade system/sw/xorg/current
bash# kdb mv system/sw/xorg/current system/sw/xorg/old
bash# kdb rm system/inittab/rc4
bash$ KDB_BACKEND=gconf kdb rm user/gconfKey
bash# kdb export user/sw/app | sed -e 's|/app/|/app2/|g' | kdb import
bash# KDB_ROOT=system/sw kdb export myapp > myappconf.xml
bash# kdb import myappconf.xml
bash$ KDB_BACKEND=gconf kdb export user/sw
The kdb edit command provides an easy way to edit a group of keys using your prefered editor. It will dump the selected keys into an XML representation, and run the editor to let the user edit them. After user modifications, it will update, remove or include the keys based on the modified XML file. This real screenshot shows me using kate to modify the XFree configuration keys related to the monitor. If I didn't force it with the $EDITOR environment, vi would be used.
But the de-facto administration tool is kdbedit.
