Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Run the test suite as a GitHub Action #208

Draft
wants to merge 8 commits into
base: main
Choose a base branch
from

Conversation

EmperorArthur
Copy link
Collaborator

No description provided.

@EmperorArthur EmperorArthur marked this pull request as draft August 4, 2024 05:35
@EmperorArthur EmperorArthur force-pushed the ci branch 3 times, most recently from 3fbdfa8 to 207e9dc Compare August 4, 2024 05:51
scad_file_path: Path
openscad_binary_path: Path
image_folder_base: Path
parameters: Optional[dict]
'''If set, a temporary parameter file is created, and used with these variables'''

WINDOWS_DEFAULT_PATH = 'C:\\Program Files\\OpenSCAD\\openscad.exe'
LINUX_DEFAULT_PATH = Path('/bin/openscad')
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This could be just openscad, no need to add the /bin/ prefix as Linux resolves the path by whatever you put in env.PATH

Copy link
Collaborator Author

@EmperorArthur EmperorArthur Aug 9, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Does it in cases where you're not running through a shell?

Edit: I also don't think the automatic expansion works when doing the exists() check.

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I am a fan of the absolute path as it won't depend on the users environment. If something is by default on Linux it does not mean we need to depend on it when it is not necessary. Defaults van change, users can do weird shit. We already have a big dependency on openscad. If they change the install location, it fails everywhere and we can change it.

- name: Install OpenSCAD
run: |
sudo apt-get update
sudo apt-get install openscad
Copy link

@pimlie pimlie Aug 8, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This most likely installs the v2021 version, which is quite old. It might be beneficial to either install/use a nightly (appimage) version or even better use openscad's docker container for these tests.

Copy link
Collaborator

@Ruudjhuu Ruudjhuu Aug 8, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I do agree it is very old but it is the latest stable afaik.

Maybe it is possible to use a fixed "nightly" version? I hate the ci/cd breaking due to the nightly being updated and unstable. I also don't like an openscad binary in this repo.

tests/openscad_runner.py Show resolved Hide resolved
tests/openscad_runner.py Show resolved Hide resolved
@pimlie
Copy link

pimlie commented Aug 8, 2024

FWIW, I had also been working on adding a github action. You can see my WIP branch for that here: https://github.com/kennetek/gridfinity-rebuilt-openscad/compare/main...pimlie:test-use-bash-github-yaml?expand=1 It's still a WIP as I was testing the workflow with act but ran into an issue with the work folder.

Not sure if it's worth continuing that branch, but to explain my choices:

  • I used bash and not python so we could run tests directly using the openscad docker image once built-in EGL support lands (see remark on https://github.com/openscad/docker-openscad under 'Rendering a PNG).
  • The Python tests contain a lot of repetition, but due to way the tests are setup using the parameterized .scad files, we basically only ever need to change the scad parameters and maybe camera angle. Therefore I opted to use yaml files to configure the tests as that has a much lower threshold for adding new tests
  • I was planning to adding support for image comparison similar to openscad does

But I will hold off on working on my changes for now as to now 'compete' with the work from this pr.

@EmperorArthur
Copy link
Collaborator Author

@pimlie I suspect it's partly also a methodological approach.

  • I chose Python's unittest because it's a simple test framework that's easy to install and run. I'm not opposed to other test frameworks at all, but feel that trying to homebrew a testing framework just to avoid a dependency causes far too many problems.
  • Do your tests run under Windows? It is very common for git bash to be installed, but it's not required for someone to use git.
  • You might notice the Python tests can also use scad parameter files for some tests. The difference is instead of one file per test, the parameter being tested is modified.
  • This is explicitly doing unit tests at the library level. Hence all the pictures of hole negatives.
  • Image comparison would be nice, but I wonder if we could use csg export. No major feelings either way.

Future Work

  • Realistically, the hole tests should be changed to do a matrix setup instead of brute force copy/paste,
  • I use Python's unittest framework not because it's the best test framework but because it's part of Python, and that is the easiest cross platform tool I am familiar with.

@pimlie
Copy link

pimlie commented Aug 9, 2024

@EmperorArthur

  • RE: Do your tests run under Windows? -> They should under WSL. Native Windows support itself for tests should not be a strict requirement imho, too many possible caveats/limitations. The main env (and thus first priority) that will run the tests should be the github workflow which normally runs directly under Linux, if we can manage to run the tests under Windows too then that could be a nice-to-have but requiring Windows test support could limit the project.
  • RE: You might notice the Python tests can also use scad parameter files for some tests -> I did indeed and thought that having a default parameter file and then overwrite individual parameters on the cli was an excellent choice by you for the tests! Hence why I tried to copy that behaviour, I just used yaml to make that 'overwrite' behaviour super clear and simple to do :)

Btw, I looked at quite a bit of other SCAD projects and I didn't find any other projects yet that have fully automated OpenSCAD tests. Have you found any? Cause it would also be nice if we wouldn't have to (re-)invent the wheel for that...

Anyway, don't let my remarks/suggestions keep you from implementing the tests the way you feel is best :)

@ntindle
Copy link

ntindle commented Oct 9, 2024

Just checking into what’s blocking this. Imo the comments are addressed and I’m always in favor of running tests in CI (even in a decreased capacity). Time is cheaper spent in CI than debugging after a break

@EmperorArthur
Copy link
Collaborator Author

Just checking into what’s blocking this.

Merging requires at least one approving reviewer. Which has not happened. There are also some merge conflicts I need to address, and I have left it in draft status for the moment.

Right now, I'm focusing on fixing some bugs in the code itself, and getting CI/CD working again.

@ntindle
Copy link

ntindle commented Oct 13, 2024

Gotcha, no pressure just was trying to see if there was a place to help if I had free time

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants