For each of the following problems, I'd like to see not only the
command or alias you came up with, but also one or more tests. To show
me your tests, I suggest starting a shell session whose output goes to
both the screen and a file (bash | tee logfile), typing
the various commands you want to demonstrate, then exit and
e-mail me the log file.
Imagine that you're building a big Web site with lots of pages.
Some of the pages are in .html files, some in
.shtml files, and some in .php files;
they're all in the same directory.
Write a Unix command to list all the files in a
specified directory whose names end in .html,
.shtml, or .php.
Let's change the problem slightly: now the pages are all scattered throughout a main directory and its subdirectories, sub-subdirectories, and so on.
Write a Unix command to list all the files in a specified
file hierarchy whose names end in .html,
.shtml, or .php.
Write a bash alias, shell function, or shell
script (your choice) named pages for this
command. It should take in the directory as its first and only
parameter.
Imagine that many of these Web pages refer to a page named
snark.html, which you've decided to rename as
boojum.html. Naturally, all the references will need to be
fixed. We could do this with vi, but we don't want to waste
time looking through the hundreds of files that don't even mention
snark.html.
Write a Unix command to edit (using vi)
all the files in a specified file hierarchy whose names end in
.html, .shtml, or .php
and whose contents contain the name
snark.html. (Hint: first figure out a Unix
command to simply list the pathnames of the relevant files.)
Write a bash alias, shell function, or shell
script (your choice) named vipages for this command.
It should take in the directory name as its first parameter, and the
name of the file you're looking for (snark.html in the above
example) as its second parameter.
As the number of Web pages in the site grows, editing all this stuff by hand with vi gets pretty tiresome.
Write a Unix command that automatically
goes through all the HTML, SHTML, and PHP files in the file hierarchy
that refer to
snark.html, and replaces each such reference with
boojum.html.
Write a bash alias, shell function, or shell
script (your choice) named substpages for this command.
It should take in the
directory name as its first parameter, the old name
(snark.html in this example) as the second parameter,
and the new name
(boojum.html in this example) as the third parameter.
Write a Unix command to show all the lines in a
specified file that do not contain the string
"-option".
(This can be done either with sed or with grep.
Try both approaches.)
Write an alias, shell function, or shell script
to extract the "useful information"
from an online movie listing page like this
one and present it in a useful plain-text form. Ideally, there would be a
line for each showing of each movie, containing all the essential information
about that movie and showing.
Also write a second alias, shell
function, or shell script to format this information nicely for a human
reader (perhaps multiple lines per showing). You should be able to (for example)
pipe the first of these into grep to find all the movies starting at
4:15, then pipe that into the second to get a nicely-formatted list.
Write a bash alias, shell function, or shell script (your
choice) named freq
which reads words from stdin, counts how often each one
occurs, and produces output in which each line is a word followed by how
many times it occurred in the input. The lines of output should be
sorted in decreasing order of frequency. (You should be able to do this
using only standard Unix commands, not writing any C or C++ code.)
Hint:
It may be easier to do this as a shell script or a shell function,
rather than an alias for a pipe.
You can do it as a shell script using fairly simple filters;
you can do it as a single pipe using awk.
For those of you who have taken Dr. Siegfried's courses, another name
for this might be "concordance"....