Due: Mon Feb 4 11:59 pm
Ontime bonus 5%. Grace period for late submissions until Wed Feb 6 11:59 pm
Assignment by Julie Zelenski and Michael Chang, with modifications by Nick Troccoli
This assignment delves into pointers, arrays, and dynamic memory. You will be building your skills with:
- manipulating pointers & arrays
- managing memory allocation on the stack and heap
- using the C library I/O functions
- using tools such as Valgrind to help track down memory errors and leaks
In this assignment, you will implement simplified versions of the unix commands
uniq command will pull out unique lines from the provided input and print out their frequencies. Your
tail command will print out the last
n lines of the provided input.
To get started on the assignment, clone the starter project using the command
git clone /afs/ir/class/cs107/repos/assign3/$USER assign3
The starter project contains the following:
Makefile: three files that you modify, and their Makefile for compiling
custom_tests: the file where you can (but are not required to on this assignment) add custom tests for your programs
mycat.c: a program that calls your
util.cfor testing purposes.
util_grader: a program to help with output comparison. You can ignore this.
samples: a symbolic link to the shared directory for this assignment. It contains:
sanity.py: files to configure and run Sanity Check. You can ignore these.
mycat_soln: executable solutions for the programs you will write.
names: sample text files for testing your unix filters
tools: contains symbolic links to the
submitprograms for testing and submitting your work.
Background: Unix Filter Commands
tail unix commands are part of a special category of commands called "filters". A filter is a program that reads input (typically line by line) from the user or from a file, transforms the input in some way, and produces some output. Some other example unix filters are
cat is the simplest of all filters; the output just repeats the input. Invoking
cat filename will print the contents of the named file. The man page for
cat indicates that if no filename argument is given, the command will read from standard input. This means the program will read lines typed on the terminal by the user. Almost every unix filter (e.g.
sort and more) conforms to this same interface, either reading from a named file or standard input. Reading from standard input can be handy for running a quick test without having to first create a file.
Try this out for yourself:
cat Makefileto view the contents of the file.
- Invoke with the
-bflag to number the non-blank lines:
cat -b Makefile.
- Now invoke
cat -bwithout a filename argument. The program will wait for you to type input.
- Type a line of text and hit enter. It echoes the line with its line number. Type a few more lines and see they are also echoed back and numbered.
- When you have finished typing your input, press control-d on a line by itself. This signals the end of input (EOF for "end of file"). The filter program exits.
Below are a few other unix filters to explore. For each, skim the manual page to learn about its options. Run the filter on an existing file and then again without a filename to operate on standard input. Remember that control-d is used to signal EOF.
headoutputs the first lines of the input
tailoutputs the last lines of the input
wcoutputs the count of lines, words, and characters in the input
grepoutputs lines from the input that match the search pattern
uniqfilters out adjacent duplicate lines from the input
sortoutputs the input rearranged into sorted order
When running on standard input, you will notice that some filters immediately respond to each line you enter, e.g
grep echoes the line if it matches the pattern. Others (e.g.
sort) don't output anything until you signal EOF. The depends on whether the filter operation needs to see all of the input before it can determine what to output.
The true power of filters is that they can be combined into "pipelines". A pipeline is a sequence of commands where the output of one command is passed to (or "piped into") the input of the next command. The pipe character
| feeds the output of one command as input to the next. Filters are especially useful in pipelines because they can be used at the beginning, in the middle, or at the end. You can read more about piping input in this guide on the Resources page.
Pipelines allow you to build fancier operations out of the basic filters. Some examples:
grep include util.c | wc -lfinds lines in
util.ccontaining "include", and feeds them to
wc, which outputs the number of lines in its received input. This pipeline thus reports the count of matches.
cat -b util.c | tailoutputs the lines of
util.clabeled with line numbers, and feeds them to
tail, which outputs only the last 10 lines and numbers.
- The standard
uniqcommand only filters out duplicate lines if they are adjacent in the input, while your version will filter out duplicate lines appearing anywhere in the input. For this reason, with the standard
uniqcommand, if instead your data has duplicate lines scattered throughout, you need to first rearrange them so duplicates are adjacent for
uniqto detect them. We can use a three-stage pipeline to output the count of distinct lines in a file:
sort samples/names | uniq | wc -l. First,
sortoutputs the sorted lines in
samples/namesand feeds them to
uniq, which outputs out all the lines, omitting adjacent duplicates, to
wc, which outputs the number of lines in its received input.
Now it's your turn! How could we write a pipeline that prints out the three most common names in
samples/names? (Hint: review the manpage for
uniq to see what flags it has, and keep in mind that the same command can appear multiple times in a pipeline.)
The Unix philosophy is to move away from the construction of monolithic do-everything programs and instead encourage a proliferation of smaller commands, each focused and streamlined to perform one task well. Those small, discrete tools are combined to accomplish larger tasks. You'll be writing two of these tools in this assignment:
Review and Comment Starter Code
myuniq programs rely on reading files. File reading works similarly, but has different syntax, across different languages. Your first task is to study the provided program code, along with our guide to the C Standard Library (or your favorite C reference) to get an overview of how file-reading works in C. The test program
mycat.c is given to you fully implemented. The starter files for
mytail.c contain code to handle the command-line arguments and file setup, and you will implement the rest. You should read and understand the given code, work out how to change/extend it to suit your needs, and finally add comments to document your strategy.
Some thoughts to consider when reviewing the given code:
- What are the possible options for the mode argument to
fopen? What happens if you attempt to open a file that doesn't exist or you don't have permissions for?
- What is the consequence of failing to
- How does a program switch to reading standard input when there is no filename argument?
- Are the operations that read from a file also used to read standard input or is different code required for each?
- The default behavior of the
mycatprogram corresponds to invoking which command-line flag on standard
mytailsupports a command-line argument of
-numberto control the number of lines printed. How does the given code process that optional argument?
- What range of values is accepted as an argument for
- Do you see anything unexpected or erroneous? We intend for our code to be bug-free; if you find otherwise, please let us know!
As per usual, the code we provide has been stripped of its comments and it will be your job to provide the missing documentation in the
util.c files. You are not required to comment the
The standard library function
fgets is used to read text from a file.
char *fgets(char *buf, int bufsize, FILE *stream);
Like most library functions,
fgets does not do any allocation; instead it depends on the client to provide the buffer to write into. The client must pre-allocate memory based on an expectation of what size will be "large enough" to store the line. If the client's choice of size is too small, a line-reading function might overflow the buffer (if you use
gets) or truncate the read operation (if you use
A more client-friendly design would be for the read function to handle allocation internally and make a line of the appropriate size. You are going to implement a
read_line function that operates in this manner:
char *read_line(FILE *fp);
read_line function reads the next line from a file. The return value is a dynamically-allocated and null-terminated string (or
NULL if at EOF, "end of file"). This function performs like a snazzier version of
fgets (in fact, internally you should be implementing it using
fgets!) in that it reads the next line but this version also allocates the storage, making life easy for the client.
Here are some specific details of the function's operation:
- The function should read the next line from the file. A line is considered to end after the first newline character or EOF, whichever comes first.
- The returned string should include the characters read up to the newline character, but should not include the newline character itself. If the next line consists of just a newline character, return an empty string.
read_lineimmediately hits EOF, meaning there are no characters to be read at all, the function should return
NULL. Note that the EOF condition cannot be checked until after you have called
fgets, because it is the process of trying to read past the end that triggers the EOF condition. This case should clean up any initial allocation that turned out to not be needed after all and return
- If an error occurs in
fgets, the function should return whatever has been read up to that point, or NULL if it has not read anything (hint: your function may end up doing this without you explicitly considering this case). You may assume in this case that if an error occurs during a call to
fgets, no characters were read that time.
- To allocate memory for the returned string, the function should make an initial allocation and successively enlarge it if needed. To be more specific, it should first malloc a buffer of minimum size (32) and read the first part of the line into the buffer, using a single call to
fgets. If there is more to read, it should realloc the buffer to double its current size (64), and make another call to
fgetsto read the rest of the line. It should continue like this, doubling the allocation and reading more, until finally reaching the newline or EOF that signals the end of this line.
- If it is unable to allocate sufficient memory,
read_lineshould raise a fatal assert. As a habit, you should
assertthe result from every allocation request. Allocation failures are rare but deadly.
read_lineshould return the address of a dynamically-allocated piece of memory storing the next line. The client is responsible for deallocating this memory with
freewhen no longer needed.
- You should need to use only the
fgetsfile I/O function, and not any other functions such as
Write your implementation of
read_line in the
util.c file. You can test it using our provided
mycat.c program. The
mycat program is integrated with sanitycheck. You will later use the
read_line function to write the
What does the
tail command do?
tail is a unix filter that prints the final N lines of the input. The value of N defaults to 10, but can also be set as a command-line flag, e.g.
tail -4. Consider the following sample use of
myth$ cat samples/colors red green green yellow blue blue blue blue myth$ tail -3 samples/colors blue blue blue
If the input contains fewer than N total lines, all lines from the input are printed.
mytail program you are to write is similar in operation to the standard
tail with these differences:
mytailsupports only one command-line flag,
-N, where N is the number of lines to output (standard
tailhas a number of other flags)
mytailonly reads one file: either the named file (if specified) or standard input (if not) (standard
tailcan read from any number of file arguments)
mytail is processing its input, it will need to keep track of N lines. For this you must use an array of N entries. At any given point, this array holds the most recent N lines read. The array is a "sliding window" of N lines that is moving its way through the input. When
mytail reaches the end of the input, the window will contain the final N lines.
mytail starts processing the input, it will read the first N lines into the array using your
read_line function. Upon reading the N+1st line, it should not enlarge the array, but instead discard the 1st line of the input and use that array slot for the N+1st. The N+2nd overwrites the second line and so on. In this way, an array is used as a circular queue.
Since the array size is known in advance and doesn't change over its lifetime, this array is good opportunity to use stack allocation. The challenges in implementing this program are the logic for the circular queue, as well as being careful with the strings being allocated and deallocated.
What does the
uniq command do?
uniq is a unix filter that outputs the provided input with adjacent duplicate lines removed. When
uniq is invoked with the
-c flag, it counts the number of occurrences for each line. Consider the following sample use of
myth$ cat samples/colors red green green yellow blue blue blue blue myth$ uniq -c samples/colors 1 red 2 green 1 yellow 4 blue
myuniq program you are to write is similar in operation to the standard
uniq with a few differences:
myuniqfilters out all duplicate lines, no matter where they appear in the input (standard
uniqonly detects repeated lines when they are immediately adjacent to one another in the input)
- Only one copy of each unique line is printed. The unique lines are printed in the same order they appeared in the input.
myuniqalways prefixes each line of output with its count of occurrences in the input (this is the behavior of standard
uniqwhen invoked with the
myuniqsupports no command-line flags (standard
uniqhas a number of flags)
In order to detect repeated lines anywhere in the input,
myuniq must keep track of all of the lines it has seen as it moves through the input. It will also need to count the frequency for each line. An array of structs to store this information seems like just the tool for the job, but we have a conundrum. How many entries do we need for this array? Unlike
mytail which knew in advance the exact number of entries to allocate,
myuniq is the dark. It might need an array of 10 entries or 10,0000. Any arbitrary fixed choice will be too large for some inputs and too small for others. Sounds like an opportunity for resize-as-you-go!
The strategy we require for
myuniq is to allocate the array for 100 entries as the initial estimate, and if/when that fills up, resize the array to add another 100 entries. Then repeat this as needed.
To match the output format of the sample solution, when printing out the line count number you should use the
%7d instead of just
%d. This will print the number with width 7, meaning the output will look something like:
1 red 5000 blue 12413 green 1253473 yellow
Most of the challenge in the
myuniq program concerns proper memory handling. By the time you are done with this assignment, you will be a heap master!
This assignment provides a lot of targeted practice in managing memory. You'll be using both the stack and heap for various purposes and need to take care to appropriately allocate and properly deallocate each. Here are some general memory guidelines as well as a few recommendations specific to this program:
- Stack allocation is as simple as declaring a variable. The stack is convenient (auto-allocation/deallocation), type-safe, and very efficient.
- Dynamic allocation is via malloc/realloc/free. The heap gives the programmer control over lifetime (explicit allocation/deallocation) and flexibility (storage can be re-sized/re-purposed), but comes at a cost in safety and efficiency.
- Take care to understand proper use of stack versus heap allocation and when each is appropriate. As a general rule of thumb, unless a situation requires dynamic allocation, stack allocation is preferred. Often both techniques are used together in a program. Heap allocation is a necessity when:
- you have a very large allocation that could blow out the stack
- you need to control the memory lifetime, or memory must persist outside of a function call
- you need to resize memory after its initial allocation
It is essential that you are using Valgrind throughout. Check out our Valgrind guide, and/or revisit the Valgrind exercises in labs 2 and 3 to be sure you know how to put Valgrind to work for you. In particular, let Valgrind help you with:
- Memory errors. This is by far the most important reason to have Valgrind on the job. If Valgrind is reporting errors, they need your immediate attention.
- Memory leaks. If Valgrind reports leaks, you have some heap allocations without their matching deallocation. Given that leaks generally do not cause bugs, and incorrect deallocation can, we recommend that you do not worry about freeing memory until your program is finished. Then, you can go back and deallocate your memory as appropriate, ensuring correctness at each step.
We will be looking at the memory and runtime efficiency of your programs. For memory efficiency, Valgrind can be a huge help, as the Valgrind report includes stats on your total heap usage. Look for a line at the end of a Valgrind report in this format:
total heap usage: 10 allocs, 10 frees, 888 bytes allocated
This is the aggregate total of all heap allocations during the program's lifetime. The count of allocs should always equal the count of frees. The count of allocations and total size allocated is a measure of your program's memory "footprint".
Runtime efficiency (how fast your program runs) is also important. Prefix a command with
time to execute and report on the time used to complete it. For example,
time ./mytail will measure your program's runtime, and
time samples/mytail_soln will measure the sample solution's runtime. The time labeled "user" is the one of interest.
% time ./mytail -1 samples/dictionary zygote real 0m0.010s user 0m0.006s sys 0m0.000s
% time samples/mytail_soln -1 samples/dictionary zygote real 0m0.009s user 0m0.007s sys 0m0.000s
The best strategy for ensuring appropriate memory and runtime efficiency is to measure the sample executable on a given input and compare to the measurement of your program running on the same input. If your performance is in the same ballpark as ours (say within a factor of 2 or 3), you're golden, but you should aim for closer to validate your usage of memory and time is correct. When considering tradeoffs of space versus time, the sample executable shows you the balance we are looking for. If your program uses noticeably more memory or time, it indicates you have some unnecessary or redundant work that should be eliminated. Be sure to note that very small inputs are, well, too small for meaningful measurement; instead also measure on heftier input to see the program operating at scale.
Once you are finished working and have saved all your changes, check out the guide to working on assignments for how to submit your work. We recommend you do a trial submit in advance of the deadline to allow time to work through any snags. You may submit as many times as you would like; we will grade the latest submission. Submitting a stable but unpolished/unfinished is like an insurance policy. If the unexpected happens and you miss the deadline to submit your final version, this previous submit will earn points. Without a submission, we cannot grade your work.
Here is the tentative point breakdown:
Functionality (90 points)
- Sanity cases (25 points) Correct results on the default sanity check tests.
- Comprehensive/stress cases (40 points) Correct results for additional test cases with broad, comprehensive coverage and larger, more complex inputs.
- Clean compile (2 points) Compiles cleanly with no warnings.
- Clean runs under valgrind (15 points) Clean memory reports for all programs when run under valgrind. Memory errors (invalid read/write, use of freed memory, etc) are significant deductions. Memory leaks are a minor deduction. Every normal execution path is expected to run cleanly with no memory errors nor leaks reported. We will not test exceptional/error cases under Valgrind except for
fgetsencountering an error, which will be tested.
- Reasonable efficiency (8 points) We expect programs to be reasonably efficient in use of time and space. Full points are awarded for being on par (2-3x) with the sample program; deductions will apply for immoderate use of memory/time. There is no bonus for outperforming the benchmark (and extreme efforts to do so could detract from your code quality).
Code review (buckets together weighted to contribute ~20 points)
- Use of pointers and memory. We expect you to show proficiency in handling pointers/memory, as demonstrated by appropriate use of stack versus heap allocation, no unnecessary levels of indirection, correct use of pointee types and typecasts, const correctness, and so on.
- Program design. We expect your code to show thoughtful design and appropriate decomposition. Data should be logically structured and accessed. Control flow should be clear and direct. When you need the same code in more than one place, you should unify, not copy and paste. If the C library provides functionality needed for a task, you should leverage these library functions rather than re-implement that functionality.
- Style and readability. We expect your code to be clean and readable. We will look for descriptive names, defined constants (not magic numbers!), and consistent layout. Be sure to use the most clear and direct C syntax and constructs available to you.
- Documentation. You should document both the code you wrote and what we provided. We expect program overview and per-function comments that explain the overall design along with sparing use of inline comments to draw attention to noteworthy details or shed light on a dense or obscure passage. The audience for the comments is your C-savvy peer.
For more information about style guidelines to use for assignments, take a look at the CS107 Style Guide, linked to from the Assignments page.
On-time Bonus (+5%) The on-time bonus for this assignment is 5%. Submissions received by the due date earn the on-time bonus. If you miss the due date, late work may be submitted during the grace period without penalty. No submissions will be accepted after the grace period ends, please plan accordingly! The bonus is calculated as a percentage of the point score earned by the submission. See the General Information handout for more information about the late policy.
How did the assignment go for you? We encourage you to take a moment to reflect on how far you've come and what new knowledge and skills you have to take forward. Once you finish this assignment, you will have written your own implementation of two Unix filter commands and an improved version of a C I/O function, using pointers, arrays and dynamic memory. That's a pretty darn impressive accomplishment!
To help you gauge your progress, for each assignment/lab, we identify some of its takeaways and offer a few thought questions you can use as a self-check on your post-task understanding. If you find the responses don't come easily, it may be a sign a little extra review is warranted. These questions are not to be handed in or graded. You're encouraged to freely discuss these with your peers and course staff to solidify any gaps in you understanding before moving on from a task. They could also be useful as review before the exams.
scan_tokenfunction you wrote for assignment 2 required that the client supply the memory to write the scanned token. The
read_linefunction of assignment 3 instead allocates memory on the client's behalf. What are the advantages and disadvantages between the two approaches?
- It is stated that a correct program should have a one-to-one correspondence between
freecalls. How does adding a
realloccall change the needed number of
freecalls? Briefly explain your reasoning.
- Which is the bigger problem: memory errors or memory leaks? Why? How can you tell which is which in a Valgrind report?
- How can you gauge whether your program has acceptable performance in terms of runtime or memory efficiency?
- When it is appropriate to use
We would also appreciate if you filled out this homework survey to tell us what you think. We appreciate your feedback!
Frequently Asked Questions
Sanity check is reporting a mismatch about my program outputting an extra blank line at the end of the
mycat Makefile test. None of the other tests complain about this. Any ideas what the issue might be?
We configured this specific test to report discrepancies in how your
read_line handles empty lines and/or EOF. If
read_line is called when the next line consists of just a newline, it should return an empty string. If called where there are no more characters to read, it should return NULL. Returning empty string instead of NULL or vice versa will produce extra or missing blank lines in the output. The usual behavior of sanitycheck is to ignore whitespace when comparing the output. This specific test uses a strict comparison that only accepts an exact match so as to alert you that there is a problem with your
read_line. (Hint: make sure you're properly detecting and handling the EOF condition).
What is assert? How do I use it?
assert macro from
<assert.h> is a mechanism for fatal errors. You assert a condition that must be true to continue:
assert(num < 1000);
If the condition fails, assert prints the error and halts the program. Liberal use of asserts is a defensive programming practice used by a programmer to communicate with themselves (or other programmers) about fatal situations that should "never happen" and from which there is no possibility of recovery. For this assignment, you should verify any allocation request was successful by asserting the result was non-NULL. A NULL result can mean that the heap is exhausted or the internal malloc data structures have been damaged. In either case, there is no point in continuing. The alternative of blundering on with a NULL pointer will eventually lead to a crash, but far removed from the original error. Better to assert right away!
One other point to make about
assert is that they can be enabled/disabled by a compile-time setting. Asserts are often disabled when finalizing the production version of a program. As such, it's not a good idea to put an expression with side effects or an essential statement inside of an assert. Instead, the assert expression should merely be a test that you can do without. For example, consider these examples:
assert(ptr = malloc(size)); // if assert disabled, no malloc either! ptr = malloc(size); assert(ptr); // if assert disabled, nothing critical skipped
How do I use gdb/valgrind on my program running in a pipeline?
A simple approach is to instead write a file with the input you want to test and then gdb/valgrind your program running on that file. To prepare the input file, you can use redirection
> to capture the output of the command. Check out the guide on piping and redirection on the Resources page for more information.
How do I use standard boolean values like
false in my program?
C does not have booleans as a built-in type; instead they suggest non-zero and zero values instead, respectively. However, if you have an explicit boolean value, it's better to import
stdbool.h, which provides a
bool variable type and