Skip to content

Latest commit

 

History

History
120 lines (77 loc) · 9.03 KB

README.md

File metadata and controls

120 lines (77 loc) · 9.03 KB
Raw

The Selfie Autograder

Selfie comes with an autograder implemented in Python that features compiler and systems assignments, located in the assignments subdirectory. The autograder is publicly available and meant to be used by students for self-grading before submitting solutions of assignments.

In order to use the autograder install selfie first.

Running the Autograder

The autograder reports grades in the range from 2 to 5 where 2 is the best and 4 is the worst passing grade, and 5 is failed. Grade 1 is reserved for submitting solutions that autograde to 2 but also match the code quality and conventions of selfie as detailed below.

The autograder is invoked as follows:

  1. Change directory to the root directory of your selfie installation.
  2. Invoke the autograder without any command line options: ./grader/self.py

The autograder responds with its command line options and the list of supported assignments. If it does not respond, your version of Python is probably outdated. Make sure you have at least Python 3.

For Students

Whenever you changed selfie code you can check if selfie still self-compiles:

  1. Change directory to the root directory of your selfie installation.
  2. Invoke the autograder with the self-compile assignment: ./grader/self.py self-compile

Selfie still self-compiles if the autograder responds with grade 2. In order to find out about all other assignments supported by the autograder browse the assignments.

For Teachers

The autograder supports bulk grading of an <assignment> as follows:

  1. Change directory to the root directory of your selfie installation.
  2. Create a text file commit-links.txt that contains GitHub commit links to which you have read access, one per line and student.
  3. Create an empty directory student-repos.
  4. Invoke the autograder in bulk mode: ./grader/self.py -b commit-links.txt -d student-repos <assignment>

The autograder clones, for each commit link in commit-links.txt, the corresponding repository from GitHub into student-repos, grades it for the given <assignment>, and reports the grade as well as details on which part of the assignment passes and which does not for each commit link. To suppress the details and just see the grades use option -q.

If you subsequently invoke the autograder in bulk mode on student-repos again, it only clones repositories for commit links in commit-links.txt that are not yet in student-repos. Repositories that are already in student-repos are just updated from GitHub.

Submitting Solutions for Grading

Solutions of assignments must be submitted as git commit links to a private clone of the selfie repository hosted on GitHub and exclusively shared with your teacher such as the GitHub user ckirsch and the teaching assistant.

Creating a private clone of selfie

First install selfie either in the cloud or locally on your machine.

Then, on the web:

  1. Create an account on github.com unless you already have one.
  2. Set up SSH-key authentication for GitHub if you haven't done it yet.
  3. Create a new, empty repository, name it myselfie, and set it to private.
  4. Invite your teacher, for example, the GitHub user ckirsch, to the repository as your only collaborator.

And then, in a terminal where your selfie installation is:

  1. Change directory to the root directory of your selfie installation (from https://github.com/cksystemsteaching/selfie).
  2. Change the origin remote name to upstream: git remote rename origin upstream
  3. Add your myselfie repository on GitHub as origin: git remote add origin git@github.com:<yourusername>/myselfie.git
  4. Update your installation from upstream: git fetch upstream (or, initially git fetch --unshallow upstream if you are on repl.it)
  5. Mirror your installation to your myselfie repository on GitHub: git push --mirror origin
  6. Set up the main branch of your installation to push to your myselfie repository: git branch --set-upstream-to=origin/main main

Your selfie installation as well as your myselfie repository on GitHub are successfully set up and ready for submitting solutions of assignments.

Before working on an assignment

Update your selfie installation to the latest version of the official selfie repository:

  1. Change directory to the root directory of your selfie installation.
  2. Make sure that the official selfie repository is set as upstream: git remote add upstream https://github.com/cksystemsteaching/selfie.git
  3. Fetch the latest commits from the official selfie repository (upstream): git fetch upstream main
  4. Make sure the main branch of your selfie installation is checked out: git checkout main
  5. Merge the updated main branch of the official selfie repository into your main branch: git merge upstream/main
  6. Depending on how complex the changes made in your selfie installation or the official selfie repository are, you may need to resolve merge conflicts by hand. Please make sure to include both your changes as well as the changes in the official selfie repository.
  7. Push your updated main branch to your myselfie repository on GitHub: git push origin

You have successfully pulled changes from the official selfie repository into your selfie installation as well as your myselfie repository on GitHub.

Note that whenever you need to update your selfie installation but your working tree is not clean according to git status, you first need to commit your changes using git add and git commit. You may also stash your changes using git stash and later apply them again using git stash pop.

Working on an assignment

Create a development branch in your selfie installation and work on your solution of an <assignment> (regularly committing, pushing, updating, merging):

  1. Change directory to the root directory of your selfie installation.
  2. Make sure the main branch in your selfie installation is checked out: git checkout main
  3. Create the development branch off the main branch and check it out: git checkout -b <developmentbranch>
  4. Work on your solution and use the autograder for feedback. Do not change the autograder in any way! If you discover a bug, please report it to your teacher.
  5. Commit your changes regularly using git add and git commit with the commit messages formatted as "message [assignment]" which triggers the autograder on the <assignment> as GitHub Action on the next push.
  6. Push your changes to your myselfie repository on GitHub for backup: git push -u origin <developmentbranch>
  7. Update your selfie installation to the latest version of the official selfie repository regularly using the above instructions.
  8. If you fetched and merged updates go back to the <developmentbranch>: git checkout <developmentbranch>
  9. Merge the updates into your <developmentbranch>: git merge main

Submitting your solution

In a terminal where your selfie installation is:

  1. Change directory to the root directory of your selfie installation.
  2. Make sure the main branch of your selfie installation is checked out: git checkout main
  3. Merge the <developmentbranch> with your solution into the main branch of your selfie installation: git merge --squash <developmentbranch>
  4. Commit the merged changes: git commit
  5. Push your updated main branch to your myselfie repository on GitHub: git push origin

Note that the option --squash in step 3 is important. It makes sure that all commits on the <developmentbranch> are squashed into one commit which improves readability of your solution.

Finally, on the web:

With your solution committed into the main branch of your myselfie repository on GibHub in accordance to the above instructions:

  1. Go to your myselfie repository on GitHub in a browser and click on the Latest commit link.
  2. Copy the link and make sure that the link is in the form https://github.com/<yourusername>/myselfie/commit/<commithash>.
  3. Submit the link along with the grade reported by the autograder to your teacher as discussed in class.

If the autograder reports grade 2, you may submit grade 1 provided that, in a code review, you are able to demonstrate that your code matches the code quality and conventions of selfie.

For each assignment we provide a summary of the results to all students. If you have not received any individual feedback from us by that time, your grade is the grade you submitted!

Solutions of assignments must be submitted before the deadline stated in class or online. Moreover, submitted commit links must refer to a private repository that is accessible to you and us (teacher and teaching assistant) and no one else throughout the semester. Late submissions or submissions that violate these rules will be failed.