md_src_plugins_dump_README(3) README


The file format consists of a simple command language with arguments. When an argument is binary or string data the length needs to passed first. Because the size is known in advance, any binary dump is accepted. Terminating characters present no problem. The commands are assembled similar to the ones present in Elektra’s API.

The file starts with the magic word kdbOpen followed by a version number. Processing can be stopped immediately when it is not in Elektra’s dump format at all. A wrong version number most likely indicates that the version of the plugin is too old to recognise all commands in the file. The basic idea of the dump plugin is to write out the way that the KeySet needs to be constructed. The dump plugin interprets such a file. The file also looks similar to Ccode that would create the KeySet. Keys can contain any binary values and arbitrary metadata and are still stored and parsed correctly. The dump plugin can even reconstruct pointers to metadata to save memory. When a pointer to the same region of memory is found, a special command keyCopyMeta is written out that is able to reconstruct the data structure the way it was before. The commands were designed to make parsing of the file an easy task.

Format Examples

The serialised configuration can look like (0 bytes at end of strings are omitted):

    kdbOpen 1               
    ksNew 207               
    keyNew 27 1             
    keyMeta 8 27            
    commentBelow are the mountpoints.                
    keyNew 32 19             
    system/elektra/mountpoints/dbusserialised Backend               
    keyEnd keyNew 39 1              
    keyMeta 8 72            
    commentThis is a configuration for a backend, see subkeys for more information          
    keyNew 53 1             
    keyCopyMeta 59 11               


Export a KeySet using dump:

    kdb export system/example dump > example.ecf

Import a KeySet using dump:

    cat example.ecf | kdb import system/example dump

Using grep/diff or other unix tools on the dump file. Make sure that you treat it as text file, e.g.:

    grep --text mountpoints example.ecf