Quick Incremental Copy

Implementation of "cp -I" linux command

Goals/Tasks

The goal of this project is to copy directories/files quicker than the simple cp command.

Summary

The programm has two arguments: the source (A) file/directory and the destination (T) file/directory. The programm traverses through A, reads the info of the i-nodes and for every a i-node in A it searces the corresponding t i-node in T. We distinguish 4 cases:

1. The t doesn't exist in T: in this case the t has to be created in the right place inside T and the a to be copied in the t. If the i-node refers to a file, we also have to copy the data of the file.
2. The t exists in T and is the same as the a in A: in this case we do nothing. This is the case where the quick incremental copy "wins" the simple cp.
3. The t exists in T, but there is no corresponding in a in A: in this case the file from A was deleted and we have to update the T.
4. The t exists in T and it's not the same as the a in A: in this case we have to copy the content of the a i-node to t i-node.

When 2 files are not the same:

• obviously, we don't read the content of the files, because this is time-consuming.
• if a refers to a file and t refers to a directory
• if a and t refer to directories, we check their content recursively
• if a and t refer to files with different size
• if a and t refer to files with the same size but the t is older than a

If the i-nodes a and t are referring to directories, we have to recursively do the above steps.

Compilation

./quic -v -d origindir destdir

• origindir is the source directory
• destdir is the destination directory
• v is a flag and prints info about choices/actions of the programm regarding to copying/deleting files.
• d is a flag that defines that the files that were deleted from the source directory must not exist in the destination folder.