Go to the first, previous, next, last section, table of contents.

Clusters

E2compr works by conceptually dividing a file into sections of n blocks (where n is constant for a given file). We call these sections `clusters', and n the `cluster size'. E2compr replaces non-hole data in a cluster with blocks of compressed data followed by one or more holes. Suppose we have a file that looks like this (11 blocks of data), and that the cluster size is 4 blocks:

    block#
    ------

     [0]        data
     [1]        data
     [2]        data
     [3]        data
     [4]        data
     [5]        data
     [6]        data
     [7]        data
     [8]        data
     [9]        data
     [10]       data

If we want to compress the file, we cut the file into clusters and compress every cluster individually:

    block#
    ------

     [0]        data              <- cluster #0
     [1]        data
     [2]        data
     [3]        data

     [4]        data              <- cluster #1
     [5]        data
     [6]        data
     [7]        data

     [8]        data              <- cluster #2
     [9]        data
     [10]       data

Suppose the cluster #0 compressed into 1 block, the cluster #1 into 2 blocks and the cluster #3 into 2 blocks. We just replace the uncompressed blocks with the compressed one, at the same place, and remove the unneeded blocks. The file becomes like this:

    block#          bytes
    ------          -----

     [0]        compressed data   <- cluster #0
     [1]        missing block
     [2]        missing block
     [3]        missing block

     [4]        compressed data   <- cluster #1
     [5]        compressed data
     [6]        missing block
     [7]        missing block

     [8]        compressed data   <- cluster #2
     [9]        compressed data
     [10]       missing block

The file now uses only 5 blocks on the disk.

Go to the first, previous, next, last section, table of contents.