Utilities Unleashed

Entire Assignment due 2024-02-07 23:59
Graded files:
  • env.c
  • time.c

Learning Objectives

The learning objectives for Utilities Unleashed are:

  • Fork, Exec, Wait
  • Environment Variables
  • Writing a C Program
  • Using argv, argc
  • Introduction to core utils

Overview #

In this lab, you will be implementing the following C utilities:


  • Do not worry about flags or features that we do not mention.
  • Do not print any of your debug information out for your final submission.
  • All printing (except env vars) should be handled with format.h.
  • A common issue is double printouts. If this happens to you, try flushing stdout before you fork/exec. If this solves your issue, ask yourself why.


If you fork bomb on any autograder run, you will receive a zero on this assignment.

To prevent you from fork bombing your own VM, we recommend looking into ulimit. This will allow you to set a limit for various operations on your system, including how many files you can open concurrently or how many times you can fork. The side effect is that a poorly chosen limit may throttle your system’s operations (e.g. setting the fork limit too low may make your terminal unable to execute commands!).

format.c and .h #

Since this lab requires your programs to print messages to stdout and stderr, we have provided you with format.c and format.h. You should not be printing out to stdout and stderr at all. Instead, you should be using the provided functions. You can find documentation for each function in format.h. Please read the documentation in format.h multiple times to determine when each function should be used. This is our way of ensuring that you do not lose points for formatting issues, but it also means that you are responsible for handling any errors mentioned in format.c and format.h.

It is common for students to fail certain test cases on this assignment with seemingly functional code, it is almost always because of improper usage of format.h.

time #

In this lab, you will be implementing time.

time – run a program and report how long it took

So if a user enters:

./time sleep 2

then time will run sleep with the argument 2 and print how long it took in seconds:

sleep 2 took 2.002345 seconds

For more examples, you can play with Linux’s builtin time command by typing time YOURCOMMAND (time ls -l, for example) in your terminal. Be sure to add ./ to the beginning (or use the full path to your time executable file if you are in another directory), otherwise the builtin time will be called.

We’ve also provided a test executable to run basic tests on your time implementation. Note that although these tests are similar to those that will be run on the autograder they are not identical, so passing locally does not guarantee you will receive full credit. It is still your responsibility to ensure you have functional code.

Note that we only care about wall-clock time, and we recommend using clock_gettime with CLOCK_MONOTONIC.

Pro tip: 1 second == 1,000,000,000 nanoseconds.

Nota bene:

  • You may not use the existing time program.
  • You must use fork, exec, and wait (no other solutions will be accepted).
  • If the child process does not terminate successfully (where its exit status is non-zero), you should exit with status 1 without printing the time.
  • We will only run time with one program.
  • The commands we will run can take any number of arguments.
  • Do your time computations with double-precision floating pointer numbers (double) rather that single-precision (float).
  • We have provided functions in format.h that we expect you to use wherever appropriate.

Useful Resources

env #

In this lab, you will be implementing a special version of env.

env – run a program in modified environments


./env [key=val1] [key2=val1] ... -- cmd [args] ..

Please re-read this section multiple times before starting:

  • Each variable is in the form of NAME=v1, separated by spaces.
  • Values may contain references to environment variables in the form %NAME, including variables that were set earlier. As a result, variables should be processed from left to right.
  • Each reference should be replaced with its value.
  • The names of variables (both in key and in value) only contain letters, numbers, or underscore characters.
  • For each environment variable key/value pair, env will assign value to key in the child environment.
  • Each execution must be done with fork, exec, and wait.
  • The last variable/value(s) pairing is followed by a --.
  • Everything following the -- is the command and any arguments that will be executed by env.
  • Invalid input should result in the usage being printed. Invalid usage includes:
    • Cannot find -- in arguments
    • Cannot find = in an variable argument
    • Cannot find cmd after --

This is the canonical example and a practical use case:

$ ./env TZ=EST5EDT -- date
Sat Sep  9 19:19:42 EDT 2017


$ ./env TZ=EST5EDT -- date
Sat 09 Sep 2017 07:19:42 PM EDT

Example of using references to other variables:

$ ./env TEMP=EST5EDT TZ=%TEMP -- date
Sat Sep  9 19:19:42 EDT 2017


$ ./env TEMP=EST5EDT TZ=%TEMP -- date
Sat 09 Sep 2017 07:19:42 PM EDT

This has the exact same behavior as before, because TEMP is first set to EST5EDT, and then when TZ is set to %TEMP, the value of EST5EDT is retrieved and then TZ is set to that. Notice that the variables are set sequentially, or else it wouldn’t work.

We have provided you with a reference executable env-reference for you to test your understanding of env’s expected behavior. You can also use it to see if your env’s output matches the expected output.

Again like time, you can play with Linux’s builtin env command by typing env <var-list> <command-name> (env MYVAR=CS341 printenv, for example) in your terminal. Again, remember to add ./ to the beginning (or the full path to your env executable file if you are in another directory), otherwise the builtin env will be called. Do not use the built-in env, or you will immediately fail the assignment

In addition, keep in mind that the builtin env uses $ instead of % to denote environment variables. In practice, it can be very useful to change some environment variables when running certain commands.

Extra: Why Env?

For example, you may notice people write #!/usr/bin/env python on the first line of their Python script. This line ensures the Python interpreter used is the first one on user’s environment $PATH. However, users may want to use another version of Python, and it may not be the first one on $PATH. Say, your desired location is /usr/local/bin for instance.

One way to solve this is by exporting $PATH to the correct position in your terminal, however, this may mess up other commands or executable under the same session.

An alternative and better way is to use our env, and enter:

./env PATH=/usr/local/bin -- ./XXX.py

then it runs the script with the desired Python interpreter.

Nota bene

  • You may not use the existing env program. (Our specification is different than the existing env program.)
  • You may not replace % with $ or use wordexp(3).
  • You may not use execvpe, execve, or execle.
  • All changes in environment variables and execution must happen only in the child process.
  • You must use fork/exec/wait.
  • If a variable doesn’t exist, interpret its value as a zero-length string.
  • Do not fork bomb the autograder! You will fail if you forkbomb the AG. (See the warning.)

Useful Resources