CSE 130 Assignment 1: Command-line Memory

Due: January 20, 2023 at 11:59 PM

Goal This project serves as a refresher for building software in ‘C’. In particular, the project will involve review of linux system calls, buffering file I/O, memory management, and c-string parsing.

Assignment Details

For this assignment, you will write a program, memory, that provides a get/set memory abstraction for files in a linux directory. Your program should take a command from stdin and carry out the command in the current working directory. Below, we provide examples of how memory should work, outline the commands formally, describes how to handle invalid commands, and provide a notes about other functionality and limitations on memory.

Examples

Here, we illustrate how your server should operate with four examples. In each example, suppose that the user starts memory from a directory that includes the file foo.txt, with content “Hello from foo”, and the file bar.txt, with content “Hello from bar”.

Example 1. In this example, the user passes the string “get foo.txt \n” to stdin. memory outputs the contents of foo.txt (“Hello from foo”) to stdout and then exit with a return code of 0.
Example 2. In this example, the user passes the string “set baz.txt\nHello World!” to stdin. memory should create a new file, named baz.txt, in the current directory, write “Hello World!” to baz.txt, write the message OK to stdout, and then exit with a return code of 0.
Example 3. In this example, the user passes the string “set foo.txt\nHello World!” to stdin. memory should change the contents of foo to be “Hello World!”, write the message OK to stdout, and then exit with a return code of 0.
Example 4. In this example, a user passes the string “invalid foo.txt” to stdin. memory should output the message “Invalid Command\n” to stderr, and exit with a return code of 1.

Commands

Your program will receive a command from stdin as input. Each valid command with be either a get or set command.

get. A valid get command is formatted as “get \\n” and indicates that memory should output the contents of the file named location. If there is a file named location in the current directory, then memory should write the contents of location to stdout and exit with a return code of 0.
set. A valid set command is formatted as “set \\n\” and indicates that memory should assign contents to a file named location in the current directory. If location is a valid filename (see Hints section for some tips on how to check for this), then memory should assign the content of location to be content (overwriting location’s current contents if applicable), and exit with return code 0.

Invalid Commands

If memory receives an invalid command, then it should should output the text “Invalid Command\n” to stderr and exit with a return code of 1. An invalid command is one that meets any of the following:

The command does not start with either get or set (these are case-sensitive).
The command starts with get or set, but stdin was closed before the user provided a location.
The command starts with get or set, includes a location, but the location is not a valid filename (see Hints section at the end of this document).
The command is a get command, includes a valid location, but there is no file located at location.
The command is a get command, includes a valid location that exists, but also include additional
input (e.g., “get foo.txt ExtraStuff\n” or “get foo.txt\nExtraStuff”).

Notes

In addition to the functionality described above, you memory should meet the following functionality and limitations.

If memory detects any other error (e.g., if it cannot write all of the requested content to a file for any reason), the program should produce the text Operation Failed to stderr and exit with a return code of 1.
memory must be reasonably time efficient: it should buffer file input and output.
memory must be reasonably space efficient: it should use at most 1MB of memory regardless of input.
memory must not leak any memory (i.e., it should free all memory that it allocates).
memory must not leak any file descriptors (i.e., it should close all files that it opens).
memory should not crash (e.g., it should never segfault).
memory must be written using the ‘C’ programming language (not C++!).
memory cannot use the following functions from the ‘C’ stdio.h library: fwrite, fread, variants of put (i.e., fptuc, putc, putc unlocked, putchar, putchar unlocked, and putw), and get (i.e, fgetc, getc,getc unlocked, getchar, getchar unlocked, and getw).
memory cannot use functions, like system(3), that execute external programs.

Rubric

We will use the following rubric for this assignment:
Category

Makefile Your repository includes a Makefile with th rules all and memory, which produce the memory binary, and the rule clean, which removes all .o and binary files. Additionally, your Makefile should use clang (i.e., it should set CC=clang), and should use the -Wall, -Wextra, -Werror, and -pedantic flags (i.e., it should set CFLAGS=-Wall -Wextra -Werror -pedantic).

Clang-Format All .c and .h files in your repository are formatted in accordance with the .clang-format file included in your repository.

Files The following files are included in your repository: memory.c, Makefile, and README.md. Your repository also should not include binary files nor any object files (i.e., .o files). To make it easier for you to maintain tests, we added an exception to the binary file rule: you can includes binary files in any directory whose name starts with the phrase test.

Functionality Your memory program performs the functionality described in the Assignment Details sec- tion.

How to submit

Submit a 40 character commit ID hash on the Canvas assignment to identify which commit you want us to grade. We will grade the last hash that you submit to the assignment on Canvas and will use the timestamp of your last upload to determine grace days. For example, if you post a commit hash 36 hours after the deadline, we will subtract 2 grace days from your total.

Resources

Here are some resources to help you:

Testing We provided you with two resources to test your own code:

An autograder, which is run each time you push code to gitlab, will show you the points that you will
receive for your Makefile, Clang-Format, and Files.
A set of test scripts in the resources repository to check your functionality. You can use the tests to see if your functionality is correct by running them on your Ubuntu 22.04 virtual machine. We provided you with a subset of the tests that we will run, but, I bet you can figure the other ones out by adapting what we have given you :-)

Hints

Check out the document “Setting up Ubuntu” on Canvas. It has some instructions for getting an Ubuntu 22.04 virtual machine up and running.
To format a file, named foo.c, using the included .clang-format, execute the command clang-format -i -style=file foo.c. If you include the .clang-format file in the root directory of your repository (as it was when we created your repositories), you can instead execute clang-format -i foo.c.
System calls that allow you to interact with standard input and output (e.g, read and write) will be helpful to achieve the functionality listed above. If you aren’t sure how to use them, look up a function’s manual page by typing man \ in your terminal. Or, use google!
To check the return code of your program, run your program until completion and then type echo $? in your terminal.
You can check your program for memory leaks using valgrind.
memory should allow a user to get or set data from any filename that is valid on a linux filesystem. In linux, the only restrictions on filenames are that they cannot be longer than PATH MAX from the "" header and that they cannot include null characters (i.e., \0). To make your life easier, we’re simplifying further: none of our tests that send input to your memory program and specify a valid method (i.e., either get or set) will include a filename with non-ascii characters.
This assignment requires that you perform some basic string parsing. There are many ways to parse a cstring; the most straightforward for this assignment is probably strtok, but you could also check out regexec or sscanf