CS246—Assignment 1 

CS246—Assignment 1 
Questions 1 and 2 are due on Due Date 1;
the remainder of the assignment is due on Due Date 2.
1. Provide a Linux command line to accomplish each of the following tasks. Your answer in each
subquestion should consist of a single command or pipeline of commands, with no separating
semicolons (;). (Please verify before submitting that your solution consists of a single line.
Use wc for this.) Before beginning this question, familiarize yourself with the commands
outlined on the Linux handout. Keep in mind that some commands have options not listed
on the sheet, so you may need to examine some man pages. Note that some tasks refer to a
file myfile.txt. No myfile.txt is given. You should create your own for testing.
(a) Print the 10th through 25th words (including the 10th and 25th words) in /usr/share/dict/words.
You may take advantage of the fact that the words in this file are each on a separate
line. Place your command pipeline in the file a1q1a.txt.
(b) Print the (non-hidden) contents of the current directory in reverse order. Place your
command pipeline in the file a1q1b.txt.
(c) Print the number of lines in the text file myfile.txt that do not contain the string
cs246. Place your command pipeline in the file a1q1c.txt.
(d) Print the first line that contains the string cs246 from the text file myfile.txt. Place
your command pipeline in the file a1q1d.txt.
(e) Print the number of lines in the text file myfile.txt that contain the string
linux.student.cs.uwaterloo.ca where each letter could be either uppercase or lowercase. Place your command pipeline in the file a1q1e.txt.
(f) Print all (non-hidden) files in any subdirectory of the current directory that end with .c
(immediate subdirectories only, not subdirectories of subdirectories). Do not use find.
Place your command pipeline in the file a1q1f.txt.
(g) Out of the first 20 lines of myfile.txt, how many contain at least one digit? Place the
command pipeline that prints this number in the file a1q1g.txt.
(h) Print all (non-hidden) files in the current directory that start with a, contain at least
one b, and end with .c. Place your command pipeline in the file a1q1h.txt.
(i) Print a listing, in long form, of all non-hidden entries (files, directories, etc.) in the
current directory that are executable by at least one of owner, group, other (the other
permission bits could be anything). Do not attempt to solve this problem with find.
Place your command pipeline in the file a1q1i.txt.
1
(j) Before attempting this subquestion, do some reading (either skim the man page or have
a look on the Web) on the awk utility. In particular, be sure you understand the effect
of the command
awk ’{print $1}’ < myfile.txt
Give a Linux pipeline that gives a sorted, duplicate-free list of userids currently signed
on to the (school) machine the command is running on.
Place your command pipeline in the file a1q1j.txt.
2. For each of the following text search criteria, provide a regular expression that matches the
criterion, suitable for use with egrep. Your answer in each case should be a text file that
contains just the regular expression, on a single line (again, use wc to verify this). If your
pattern contains special characters, enclose it in quotes.
(a) Lines that contain both cs246 and cs247.
Place your answer in the file a1q2a.txt.
(b) Lines that contain nothing but a single occurrence of laughter, where laughter is defined
as a string of the form Hahahahahahahahahahahahahaha!, with arbitrarily many ha’s.
Place your answer in the file a1q2b.txt.
(c) Lines that contain nothing but a single occurrence of generalized laughter, which is
like ordinary laughter, except that there can be arbitrarily many (but at least one) a’s
between each pair of consecutive h’s. (For example: Haahahaaaa!) Place your answer
in the file a1q2c.txt.
(d) Lines that contain at least one a and at least two b’s.
Place your answer in the file a1q2d.txt.
(e) Lines consisting of a definition of a single C variable of type int, without initialization,
optionally preceded by unsigned, and optionally followed by any single line // comment.
Example:
int varname; // comment
You may assume that all of the whitespace in the line consists of space characters (no
tabs). You may also assume that varname will not be a C keyword (i.e., you do not
have to try to check for this with your regular expression). Place your answer in the file
a1q2e.txt.
3. Write a Bash script called findGreater that is executed as: ./findGreater myword file1 file2
where myword is a sequence of non-whitespace characters, and file1 and file2 are names
of text files in the current directory. The script prints the name of the file that contains a
larger number of lines with the occurrence of the word myword. If the files have the same
number of lines that contain occurrences of the provided word, the script prints file1 followed by a single space and then file2. In all cases, the script produces a single line of
output to standard output. You may assume that the user will call this script correctly; no
error checking is needed. Using the provided file1.txt and file2.txt, the following shows
some example runs of the script (lines not starting with a $ are the output produced by the
command executed in the previous line):
$ ./findGreater thousand file1.txt file2.txt
file2.txt
$ ./findGreater be file1.txt file2.txt
file1.txt file2.txt
$ ./findGreater Hello file1.txt file2.txt
file1.txt file2.txt
$ ./findGreater thought file1.txt file2.txt
file1.txt
$ ./findGreater thought file2.txt file2.txt
file2.txt file2.txt
Testing tools
Note: the scripts you write in the following questions will be useful every time
you write a program. Be sure to complete them! In this course, you will be responsible for your own testing. As you fix bugs and refine your code, you will very often need
to rerun old tests, to check that existing bugs have been fixed, and to ensure that no new
bugs have been introduced. This task is greatly simplified if you take the time to create a
formal test suite, and build tools to automate your testing. In the following questions, you
will implement such tools as Bash scripts.
4. Write a Bash script called produceOutputs that is invoked as follows:
./produceOutputs suite-file program
The argument suite-file is the name of a file containing a list of filename stems (more
details below), and the argument program is the name of a program to be run.
The produceOutputs script runs program on each test in the test suite and, for each test,
creates a file that contains the output produced for that test.
The file suite-file contains a list of stems, from which we construct the names of files
containing the command line arguments used by each test. Stems will not contain spaces.
For example, suppose our suite file is called suite.txt and contains the following entries:
test1 test2
reallyBigTest
Then our test suite consists of three tests. The first one (test1) will use the file test1.args.
The second one (test2) will use the file test2.args. The last one (reallyBigTest) will use
the file reallyBigTest.args.
A sample run of produceOutputs would be as follows:
./produceOutputs suite.txt ./myprogram
The script will then run ./myprogram three times, once for each test specified in suite.txt:
• The first time, it will run ./myprogram with command line arguments provided to the
program from test1.args. The results, captured from standard output, will be stored
in test1.out.
• The second time, it will run ./myprogram with command line arguments provided to the
program from test2.args. The results, captured from standard output, will be stored
in test2.out.
• The third time, it will run ./myprogram with command line arguments provided to the
program from reallyBigTest.args. The results, captured from standard output, will
be stored in reallyBigTest.out.
Note that if the test suite contains a stem but a corresponding .args file is not present, the
program is run without providing any command line arguments.
Your script must also check for incorrect number of command line arguments to produceOutputs.
If such an error condition arises, print an informative error message to standard error (any
informative format is acceptable) and abort the script with a nonzero exit status.
Note on purpose of this script: This script will be useful in situations where we provide
you with a binary version of a program (but not its source code) that you must implement.
By creating your own test cases (.in files) and then using this script to produce the intended
output you will have something to compare with when you implement your own solution (see
the next question for how to automate the comparisons).
5. Create a Bash script called runSuite that is invoked as follows:
./runSuite suite-file program
The argument suite-file is the name of a file containing a list of filename stems (more
details below), and the argument program is the name of the program to be run.
In summary, the runSuite script runs program on each test in the test suite (as specified by
suite-file) and reports on any tests whose output does not match the expected output.
The file suite-file contains a list of stems, from which we construct the names of files
containing the command line arguments and expected output of each test. Stems will not
contain spaces. For example, suppose our suite file is called suite.txt and contains the
following entries:
test1 test2
reallyBigTest
Then our test suite consists of three tests. The first one (test1) will use the file test1.args to
hold its command line arguments, and test1.out to hold its expected output. The second one
(test2) will use the file test2.args to hold its command line arguments, and test2.out to
hold its expected output. The last one (reallyBigTest) will use the file reallyBigTest.args
to hold its command line arguments, and reallyBigTest.out to hold its expected output.
A sample run of runSuite would be as follows:
./runSuite suite.txt ./myprogram
The script will then run ./myprogram three times, once for each test specified in suite.txt:
• The first time, it will run ./myprogram with command line arguments provided to the
program from test1.args. The results, captured from standard output, will be compared with test1.out.
• The second time, it will run ./myprogram with command line arguments provided to
the program from test2.args. The results, captured from standard output, will be
compared with test2.out.
• The third time, it will run ./myprogram with command line arguments provided to the
program from reallyBigTest.args. The results, captured from standard output, will
be compared with reallyBigTest.out.
Note that if the test suite contains a stem but a corresponding .args file is not present, the
program is run without providing any command line arguments.
If the output of a given test case differs from the expected output, print the following to
standard output (assuming test test2 failed):
Test failed: test2
Args:
(contents of test2.args, if it exists)
Expected:
(contents of test2.out)
Actual:
(contents of the actual program output)
with the (contents ...) lines replaced with actual file contents, as described. The literal
output Args: should appear, even if the corresponding file does not exist.
Follow these output specifications very carefully. You will lose a lot of marks if
your output does not match them. If you need to create temporary files, create them in
/tmp, and use the mktemp command to prevent name duplications. Also be sure to delete
any temporary files you create in /tmp.
Note: Do NOT attempt to compare outputs by storing them in shell variables, and then
comparing the shell variables. This is a very poor idea, and it does not scale well to programs
that produce large outputs. We reserve the right to deduct marks on all assignments for poor
solutions such as this.
You can get most of the marks for this question by fulfilling the above requirements. For full marks, your script must also check for the following error conditions:
• incorrect number of command line arguments to runSuite
• missing or unreadable .out files (for example, the suite file contains an entry xxx, but
xxx.out doesn’t exist or is unreadable).
If such an error condition arises, print an informative error message to standard error (any
informative format is acceptable) and abort the script with a nonzero exit status.
6. In this question, you will generalize the produceOutputs and runSuite scripts that you
created in problems 4 and 5. As they are currently written, these scripts cannot be used
with programs that take input from standard input. For this problem, you will enhance
produceOutputs and runSuite so that, in addition to (optionally) passing command line
arguments to the program being executed, the program can also be (optionally) provided
input from standard input. The interface to the scripts remains the same:
./produceOutputs suite.txt ./myprogram
./runSuite suite.txt ./myprogram
The format of the suite file remains the same. But now, for each testname in the suite file,
there might be an optional testname.in. If the file testname.in is present, then the script
(produceOutputs or runSuite) will run myprogram with the contents of testname.args
passed on the command line as before and the contents of testname.in used for input on
stdin. If testname.in is not present, then the behaviour is almost identical to problem 4/5
(see below for a difference): myprogram is run with command line arguments coming from
testname.args with nothing supplied as input on stdin.
The output of runSuite is changed to now also show the input provided to a test if the test
failed. Assuming test test2 from Q5 failed, the output generated by the updated runSuite
is as follows:
Test failed: test2
Args:
(contents of test2.args, if it exists)
Input:
(contents of test2.in, if it exists)
Expected:
(contents of test2.out)
Actual:
(contents of the actual program output)
with the (contents ...) lines replaced with actual file contents, as described. The literal
output Args: and Input: should appear, even if the corresponding files do not exist.
All error-checking that was required in problems 4 and 5 is required here as well.
(a) Modify produceOutputs to handle input from standard input
(b) Modify runSuite to handle input from standard input
Note: To get this working should require only very small changes to your solution to problems
4 and 5.
Submission:
The following files are due at Due Date 1: a1q1a.txt, a1q1b.txt, a1q1c.txt, a1q1d.txt, a1q1e.txt,
a1q1f.txt, a1q1g.txt, a1q1h.txt, a1q1i.txt, a1q1j.txt, a1q2a.txt, a1q2b.txt, a1q2c.txt,
a1q2d.txt, a1q2e.txt.
The following files are due at Due Date 2: findGreater, produceOutputs, runSuite, produceOutputs,
runSuite.