Functional requirements

In this assignment you will create a program capable of archiving the contents of a directory into a single file, much like tar does. You will write a makefile that compiles your program into an executable named arch, which has usage

arch

action	-c to create a new archive file -e to expand an existing archive file
source	Specifies the path-name of a directory that contains the files to be archived (-c) or of the archive file (-e) being expanded.
target	Specifies the path-name of the archive file that is to be created (-e) or of the directory to be created to hold the contents of the archive file.

Based on the command line arguments does one of the following:

Create: Create a new archive file (target) that, for each file in the directory specified by source, contains the data required to restore the file's name and data. The source directory is unchanged.

Expand: Create a new directory (target), which will contain all of the files and directories in the archive file, in their directory structure

Recurse the entire sub-tree rooted by (source) on create or expand

Proper care (error reporting / program termination) when opening and creating files and directories is required.

Sample Runs:

% ./arch -c . my_archive

archiving . 

Done  my_archive as been created

% ./arch -e ./expand_test my_archive

expanding to ./expand_test 

Done  my_archive as been expanded

Non-functional requirements

You must use basic POSIX I/O to implement both create and expand. You may not use the system() system call. You must implement functions for create and expand--it probably makes most sense to have one recursive function and one function that actually writes to the file for each command. These functions should have their own .c and .c files, while the application should be in main.c.

Archive file format

Each file will be saved in consecutive bytes in the archive file as where

pathname size: a 4 byte unsigned integer

data size: a 4 byte unsigned integer

pathname: ASCII string containing the relative path from the source directory to the file; it is not null terminated

data: the contents of the file

Files may be stored in any order.

So when expanding a file you will need to repeatedly:

Read the first 4 bytes, treat it as an unsigned int, this is the pathname size

Read the next 4 bytes as an unsigned int; this is the data size

Read the pathname into a buffer of exactly the right size

Using the pathname, open a file for writing; you may need to create directories first

Repeatedly read from the archive into a buffer and write the data to the new file

Empty directories

The scheme described above does not allow for empty directories, as only files are stored. A special file entry is used to force the creation of an empty directory. If a file's pathname begins with a "!", no file is created. Instead the archiver simply ensures all of the directories in the path (ignoring the leading !) have been created. When creating an archive, the archiver writes one such entry for each empty directory.