UNIX FILE SYSTEM

Perhaps the most important thing to obtain a grasp of when dealing with web development, is the prevalence of the UNIX style file system in almost everything you will be doing. The benefits of the unix style file system and directory layout are apparent once you have utilized them for some time, but on first introduction it can be a confusing, arcane system. While it is not compulsory to fully understand how the unix file system works to use your MediaHostNet account, many problems you may encounter will find their cause in issues surrounding your knowledge of the file system. This document is intended as an extended guide and reference - we recommended you familiarize yourself with it for consultation at a later time. Especially if the DOS/windows file system is your only prior file system experience.

There is one factor that can ease your grasp of Unix file system concepts - this is that if you have ever used HTTP or FTP (web browsing or file transfers), you will have already encountered a good deal of unix file system concepts. Much of your knowledge of how URL's (web addresses) are written will apply to the unix file system, as URL's take their format from the same source.

To illustrate this - consider the URL.

Retrieves this file from the server's hard drive:

You'll see during this chapter how the relation between these two addresses is interpreted.

Differences

The unix file system differs from the DOS style in several notable ways that make it considerably more powerful and configurable than dos [and windows]. These same features can also provide numerous headaches for the unprepared.

Names are different: Take note that windows and macintosh systems, in their goal of being user friendly, change the established name of things to a more 'friendly' name. Hence, what would be called a 'folder' under windows, will be called by its true name of 'Directory' in the context of this manual. Other terms will be explained as they are encountered.

Reversed Slash: The \ slash used to separate directories under windows, is not present in unix. Just as URL's (web addresses) use the / slash between directories, so does Unix; this is related to the creation of the web and the URL system on Unix systems initially.

Unix is case sensitive: Firstly, and perhaps most important to remember: the file system is case sensitive - consider the following:

file1.text
File1.TEXT
FILE1.text

These three names are three individual files, each could exist together in the same directory. Under windows, these would all refer to the same file, as DOS-FAT is a case INsensitive file system. Directories ['folders'] are also case sensitive.

While the debate over which is better [case sensitive or insensitive] may well continue for years, the immediate point is to always remember to check the case of the letters in a filename. 95 percent of commonly encountered problems in web development come down to a case of trying to access a a file or directory with the wrong case letters in the name.

The Root Directory: The other major fundamental difference between unix and the DOS style file systems , is the absence of 'drive letter'.
Instead of C: D: , etc. partitions and extra drives are mounted in sub directories under the 'root' file system

Thus, if drive C: is the 'root' file system , drives D: and E: could be [for example] mounted under /mnt/data and /mnt/databases.

e.g.
/usr/bin/pico
/home/jason/projects/report.txt
/usr/local/share/Windowmaker/menu.pl

Features + Concepts

The most important concepts when dealing with a Unix file system are detailed here.

Current Working Directory: As you navigate from place to place in the unix file system, , the 'place' you are currently at is referred to as the current working directory, or cwd for short. important to note this when using relative directories [see down]

Path: There are two meanings for path in the unix file system, depending on context the first usage of 'path' means the 'address' of a file, expressed as a list of directories the file is in e.g. /home/jason/projects/paper.txt the second usage of 'path' refers to the list of directories that will be looked in to find commands entered at the system command line interface. e.g. /bin /usr/bin/ /usr/sbin this is more accurately referred to as the 'working path'.

Absolute Paths: Always start with a leading / to indicate they start in the root directory. e.g. this is an absolute path /usr/bin/pico this is not bin/pico [this would look for a file named pico in the sub directory 'bin' of where ever the current working directly [cwd] is.]

Relative Paths: A relative path has no leading / and it relative to the current working directory [cwd] e.g. if the current working directory is /home/jason projects/paper.txt would refer to the file [as absolute path] /home/jason/projects/paper.txt there are a pair of 'special directories' that indicate certain directories in relative paths:
 

./	'The current working directory'
../	'The parent directory' [the one above the current dir]
~/	When a path begins with this 'your home directory'
some examples of using these special dirs in a path.
./configure	'The file configure, in the current directory' [e.g. as opposed to the configure from /usr/bin]
../paper.txt	'The file paper.txt, in the directory above this one'
../backups/oldpaper.txt	'Go up a directory, then down into the directory "backups" and get oldpaper.txt"
~/projects/paper.txt	'The file paper.txt from the directory 'projects', in my home directory'

Hard Links: Files may by linked, that is to say, a file may appear in two different directories on a file system, yet still be only the one file [the duplicates of the file are not copies, they take up no extra disk space. Changes made to the file in one location will affect it in all locations, and a file will not be deleted until all links to it and the original have been deleted. Symbolic links both directories and files may be symbolically linked.

Symlinks: [as they are more often referred to as] Are really just pointers to another file or directory, much in the way a windows shortcut or a macOS alias is. Unlike a windows shortcut however, a symlink is , for most intents and purposes, effectively the file/directory it points to, however, it differs slightly from a [hard] link in certain ways when it is copied , archived, deleted etc. for example, a symbolic link may point to a destination that does not exist, where with hard linking, so long as one link remains, the file will always be found. Absolute paths An absolute path is a way of entering the location of a file/directory, that is relative only to the root directory, not the current working directory, hence, it is an absolute 'path' the the file.

Access Permissions: File and directories have two other important properties: owner and permissions. a file or directory has three major pieces of information relating to its access permissions. The user that owns the file, the usergroup that owns the file, and the access permissions relating to the owner, the group, and the rest of the system.

permissions are of the form 'rwxrwxrwx' where rwx stands for 'read' , ' write' and 'execute' respectively
the three sets of 'rwx' relate to 'owner' 'group' and 'world' respectively.
rwx rwx rwx
[user] [group] [world]

The own of the file has the first set of permissions, then the usergroup of the file, and then anyone not in this group e.g. a file owned by user 'jason' and the group 'users', with the permissions 'rw-r----' this reads as jason has read and write permissions anyone in the group 'users' has read access anyone else has no access to the file at all. read and write are self-explanatory, but the x permission has some special cases. the x permission in the permissions of a file, determine what users can execute [run] the file. The x in the permissions of a directory determine what users can access that directory.

Summary of Access Permissions

Although not an exhaustive list of possible file permissions, this list covers all major permissions for files, and their numeric modes [to use with the 'chmod' shell command]. Because permissions affect files and directories differently, they have been separated into two lists. Unless you have very specific needs, it is recommended you do not set your files to permissions other than these listed; doing so may result in security problems on your web site, or may deny your access to your own files.

Firstly, the terms used below will be clearly defined.

• Read - Read access allows the viewing of the contents of a file. if this file resides under your public_html or public_ftp folder, this means it may be accessible over the internet [should other necessary permissions be set]. Read access to a directory gives the user the right to retrieve a listing of the files within that directory.
• Write - Write access allows the user to add/remove data from a file, or delete it entirely. Write access to a directory will allow the user to add new files to that directory.
• Run/Execute - Execute permissions on a program or script give the user the ability to active the program on the server [which will then carry out its intended function]. Execute permissions on a directory will allow the user to enter (and access files within) that directory.