Other relevant commands
Many files have names containing a suffix, or a
sequence of characters at the end of the name and commencing with a
dot. For example, if you have a program written in the language C,
the file in which that program is stored should have a suffix
.c of necessity. Let us suppose you have written a C
program that is stored in file myfile.c in your home
directory /home/ugrad/chris. From the point of view of
the UNIX kernel, it is irrelevant what name this file has. Only
when you attempt to compile and run the program will the suffix
become important, as the UNIX command for compiling a C program
demands that the .c suffix be present, and indeed will
create files with the same base
myfile and different suffix. In this example, a file
myfile.o would be created (the 'o' stands
for 'object code', i.e. binary code for the processor). A standard
POSIX command makes no demands on a file's suffix, although other
utilities may well do so; the manual page for that command will
tell you.
The command dirname takes as argument the name of a
file and strips off the actual filename, leaving only the
directories. Command basename also takes a filename as
argument, but strips off the directory information leaving only the
filename relative to its parent directory. If basename
is given two arguments, and the second argument is a suffix of the
filename, that suffix is also removed. For instance:
$ dirname
/home/ugrad/chris/test.c
/home/ugrad/chris
$ basename /home/ugrad/chris/test.c
test.c
$ basename /home/ugrad/chris/test.c .c
test
The benefit of these two commands will not be apparent at this
stage, but later on, when writing shell scripts to manipulate
files, they are exceptionally useful.
When the command mv is called, the directory in
which its first argument is located is updated so that the file's
absolute name is changed. The inode of the file is not changed if
the new filename is on the same filesystem. This command name is
somewhat misleading, since the file doesn't really move at all.
We also need to know the total amount of space taken up
by our files. Here the command du ('disk usage') comes
to our rescue. With argument of the name of a directory (or the
current directory, if no argument is given), du prints
the total number of kilobytes used to store the data in
the files in that directory. For example,
$ du
12 ./dir1
7 ./dir2
27 .
indicates that directory dir1 takes up 12k (kilobytes) and
dir2 takes 7k, whereas the total amount of storage
used for the current directory is 27k (including dir1
and dir2).
There are some other standard commands that are not required for
simple use of UNIX. Nevertheless, they are included within the
standard, and are included here for completeness.
Suppose you have distributed some text files to a colleague, and
you then make minor alterations to them. You want your colleague to
have updated copies of the files. One possibility is to send them
all the files anew, but this has the disadvantage that a
potentially large volume of data must be transmitted, which may
well incur costs. An arguably preferable method would be to send
your colleague a list of the changes to the files. These changes
can be displayed using diff, but it would be
unreasonable for your colleague to edit all the files by hand to
make the changes. Fortunately, the command patch is
provided to perform the task automatically. It takes a file
containing the changes, as generated by diff, and the
name of a file to which those changes are to be applied, and
carries out the changes.
The commands we have introduced in this chapter will be seen to
perform only simple manipulations of UNIX files, especially when
examining the contents of files. Three programs - Grep, Sed and Awk
- which we discuss later in the book - provide comprehensive
facilities for processing file contents, and obviate the need for
more 'simple' UNIX commands over and above those mentioned in this
chapter.
|