Due: Monday, June 24th, 2024 by 11:00 pm.
Summary |
For homework #0, you will fetch a source distribution that we have prepared. Next, you will trivially modify a simple hello world C application, use "make" to compile it, and run it to learn the magic code. You will also run a tool that checks your code for memory leaks and another that checks for potential style problems. Finally, you will package up and submit your code.
As you can probably tell, the real goal of this homework is to make sure that all of the course infrastructure is working for you.
Part A: The Code |
Overview
Assignments will be distributed and collected using your CSE 333 GitLab repository. Starter code will be added to your repository by the course staff and you will tag your repository when you are done with an assignment to indicate which revision should be evaluated and graded by the course staff. More information about git and GitLab is given below and in other writeups on the course web site. We will demonstrate and discuss the process in sections during the first week of class.
Clone your GitLab repository
If you have not done so already, change to a directory where you want to store the local repository for your CSE 333 projects, then follow the instructions in the CSE 333 Gitlab Guide to clone your CSE 333 repository. The Gitlab Guide is also linked to the CSE 333 resources web page, and there are additional links there to the CSE GitLab site and other git resources.
Before cloning the repo, be sure you have run the
git config ...
commands described in the CSE 333 Gitlab Guide.
That will set options that will allow you to create and change files
in your gitlab repo without receiving various alarming-looking messages each time.
You can use the CSE GitLab web interface to browse your
files and find out information about your repository, but you must
clone a copy to a Linux machine to do your work.
(Hint: In the Gitlab web interface for your repository there is a little
"clone" button in the top right corner. Click on it and find the string
labeled "ssl". Make a copy of the long git@gitlab...
string and use that in the git clone
command below to save
some typing.)
You might have additional personal projects and repositories in GitLab for your own work, but be sure to use the repository provided by us for your CSE 333 course projects.
Once you have cloned the repository, change into that directory (it
will be named something like cse333-24su-xyzzy
,
where xyzzy
is your userid) then enter a
ls -a
command. You should see something like this:
bash% git clone git@gitlab.cs.washington.edu:cse333-24su-students/cse333-24su-xyzzy.git -- git output appears here bash% cd cse333-24su-xyzzy bash% ls -a .gitignore exercises hw0(you may have additional files and/or directories in your output, but these three are the ones to verify)
The .gitignore
file lists file types that should not
normally be saved in the repository. These are typically things like
editor backup files (names ending in ~
) and binary object code files
(ending in .o
). They are files that are generated and used
locally while you are working but are recreated as needed from files
that are in the repository rather than archived permanently.
The exercises
directory is mostly empty. It is a place where you can
store your code for cse333 exercise problems. It is especially useful if you want
to discuss an exercise with a TA during online office hours. It will save time if you
push your code in its current state to this folder before you connect with the TA online.
That way you and the TA
can browse your code easily while the two of you are discussing it.
If you originally cloned your repository before the staff added the
hw0
directory to it, enter a git pull
command
to bring your local copy up to date. Once you see the hw0
directory, enter cd hw0
to change into that directory to
work on the assignment.
Part B: Edit, Compile, and Run hello_world |
In the hw0 directory, type make. This command will use the instructions in file Makefile to compile the hello_world application using the gcc compiler. Run hello_world by typing the command ./hello_world. You should see one line of output that looks like:
bash% ./hello_world The magic code is: <xxxxx> bash%
for some <xxxxx>. Now, edit hello_world.c using your favorite Unix-based editor, and change the printf so that it instead says:
bash% ./hello_world The magic word is: <xxxxx> bash%
Execute make to rebuild, and then re-run hello_world to make sure it does what you expect.
Part C: Experience the gdb Debugger
|
You're unlikely to have any runtime bugs in this homework, but let's use the debugger so that you know what it can do. Here is a session that launches the debugger, sets a breakpoint on source line 19 of hello_world.c, prints the values of a couple of variables and an expression, prints a "backtrace" of the stack (showing the sequence of procedure calls that led to executing line 19 of hello_world.c), and then continues execution. Type the italicized lines. (You may see some differences in gdb or Linux version numbers, or exact addresses of variables and code, but the data values should be basically the same.)
bash$ gdb ./hello_world GNU gdb (GDB) Red Hat Enterprise Linux 10.2-10.el9 Copyright (C) 2021 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html> This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law. Type "show copying" and "show warranty" for details. This GDB was configured as "x86_64-redhat-linux-gnu". Type "show configuration" for configuration details. For bug reporting instructions, please see: ... Reading symbols ./hello_world... (gdb) break 19 Breakpoint 1 at 0x401143: file hello_world.c, line 19. (gdb) run Starting program: /home/auser/cse333/hw0/hello_world Breakpoint 1, main (argc=1, argv=0x7fffffffd668) at hello_world.c:19 19 printf("The magic code is: %X\n", a + b); (gdb) print argv[0] $1 = 0x7fffffffe394 "/home/auser/cse333/hw0/hello_world" (gdb) print a $2 = -889262067 (gdb) print /x a $3 = 0xcafef00d (gdb) print a+b $4 = -559038737 (gdb) backtrace #0 main (argc=1, argv=0x7fffffffd668) at hello_world.c:19 (gdb) continue Continuing. The magic code is: <xxxxx> [Inferior 1 (process 6760) exited normally] (gdb) quit bash$Notes: For gdb to be useful, you must use the -g option for gcc when compiling. The makefile we distributed does that. If you have already modified your file for part B you will see the new output message after "Continuing." Most gdb commands can be abbreviated, for instance using
p
instead of print
or bt
instead of backtrace
.
The full versions of commands are used in the example above for clarity, but you will quickly want
to learn the abbreviations to speed up your use of gdb.
Part D: Verify You Have No Leaks |
Throughout the quarter, we'll also be testing whether your code has any memory leaks or other memory errors like reading values of uninitialize variables. We'll be using Valgrind to do this. Try out Valgrind for yourself, to be sure you know how run it:
bash% valgrind --leak-check=full ./hello_world
Note that Valgrind prints out that no memory leaks were found.
Part E: Check for Common Bugs and Style Issues |
Another requirement during the quarter is that your code must
be written in good style. Although there are many opinions about
what constitutes "good style", in this course we will generally follow the
Google C++ Style Guide
for both C and C++ code.
The repository files for this assignment included a Python script cpplint.py
to check C and C++ source files for style issues. Try it yourself to be sure
you can run it:
bash% ../cpplint.py --clint hello_world.c
(The --clint
option tells cpplint
that the source file
contains C code. If this option is omitted, cpplint
assumes the
source code in the file is C++.)
No style-checking tool is perfect, but you should try to clean up any problems that clint detects unless clint flags something that is definitely not a problem in this particular context (but be sure you have very good reasons to ignore any warnings).
Part F: Document Your Work |
Create a README.md
file in directory
hw0 that contains:
gcc -v
command,
which starts with "gcc version 11...." on the CSE lab machines.
You should be able to copy and
paste this from a terminal window to save typing.
If you see an older version number you should be sure that you are using
the current CSE lab machines, attu servers, or the home virtual machine.We are not picky about the format of this file as long as it has all this information. Be sure to add this file and push it to your gitlab repository.
Homework Submission and Grading |
Once you're done, "turning in" the assignment is simple -- create an appropriate tag in your git repository to designate the revision (commit) that the course staff should examine for grading. But there are multiple ways to get this wrong, so you should carefully follow the following steps in this order. The idea is:
1. Tidy up and be sure everything is properly committed. Commit all
of your changes to your repository (see the beginning of the
assignment or the main course web page for links to git
information if you need a refresher on how to do this). Then in the
hw0
directory:
If you see any messages about uncommitted changes or any other indications that the latest version of your code has not been pushed to the GitLab repository, fix those problems and push any unsaved changes before going on. Then repeat the above steps to verify that all is well.bash% git pull bash% make clean bash$ git status On branch master Your branch is up-to-date with 'origin/master'. nothing to commit, working directory clean
2. Tag your repository and push the tag information to GitLab to indicate that the current commit is the version of hw0 that you are submitting for grading:
Do not do this until after you have pushed all parts of yourbash% git tag hw0-final bash% git push --tags
hw0
solution to GitLab in the previous step.
3. Check that everything is properly stored and tagged in your repository. To be sure that you really have updated everything properly, create a brand new, empty directory that is nowhere near your regular working directory, clone the repository into the new location, and verify that everything works as expected. It is really, really, REALLY important that this not be nested anywhere inside your regular, working repository directory. Do this:
Use your own userid instead ofbash% cd <somewhere-completely-different> bash% git clone git@gitlab.cs.washington.edu:cse333-24su-students/cse333-24su-xyzzy.git bash% cd cse333-24su-xyzzy bash% git checkout hw0-final bash% ls exercises hw0
xyzzy
, of course. The
commands after git clone
change to the newly cloned
directory, then cause git to switch to the tagged commit you created
in step 2, above. We will do the same when we examine your files for
grading.
At this point you should see your hw0
directory.
cd
into it, run make
, run any tests you wish
(something that will be crucial on future assignments). If there are
any problems, immediately erase this newly cloned copy of your
repository (rm -rf cse333-24su-xyzzy
) go back to
your regular working repository, and fix whatever is wrong. It may be
as simple as running a missed git push --tags
command if
the tag was not found in the repository. If it requires more
substantive changes, you may need to do a little voodoo to get rid of
the original hw0-final
tag from your repository and
re-tag after making your repairs,
To eliminate the hw0-final
tag,
do this (this should not normally be necessary):
Then make and commit and push your repairs, and only after all the changes are pushed, repeat the tag and tag push commands from step 2. And then repeat this verification step to be sure that the updated version is actually correct.bash% git tag -d hw0-final bash% git push origin :refs/tags/hw0-final
Note: if you discover that repairs are needed when you check your work, it is crucial that you delete the newly cloned copy and make the repairs back your regular working repository. If you modify files in the cloned copy you may wind up pushing changes to GitLab that will leave your repository in a strange state, and files may appear to mysteriously vanish. Please follow the instructions precisely.
For hw0, we'll give you full credit if you learn the magic code and correctly submit your work with all the requested changes and information.