Fix dependencies; add test build to ci

[why-not-try-calmer]

No description provided.

[suricactus]

Thanks for the speedy reaction! It would be great to have the smoke test in place!

I would either rename the workflow or even better add a separate file.

BTW, this is one of the reasons I prefer having all deps in the requirements.txt.

[why-not-try-calmer]

Thanks for the speedy reaction! It would be great to have the smoke test in place!

I would either rename the workflow or even better add a separate file.

BTW, this is one of the reasons I prefer having all deps in the requirements.txt.

pre-commit.yml -> lint-test.yml would be ok? With different files you need to write more logic to avoid proceeding with a workflow that should not occur when the test / build fails.

[suricactus]

Thanks for the speedy reaction! It would be great to have the smoke test in place!
I would either rename the workflow or even better add a separate file.
BTW, this is one of the reasons I prefer having all deps in the requirements.txt.

pre-commit.yml -> lint-test.yml would be ok? With different files you need to write more logic to avoid proceeding with a workflow that should not occur when the test / build fails.

Yes, it is ok. Or lint-and-test, lint-and-build, even continuous-integration to keep it generic.

Also change the name of the workflow in line 1: name: pre-commit.

[why-not-try-calmer]

Not sure what this about though

[why-not-try-calmer]

Ah, branch protection rules I guess. I don't have the rights to update that.

[suricactus]

Still fails, interesting. It works fine locally.

[why-not-try-calmer]

... and remotely! https://github.com/opengisch/QField-docs/actions/runs/6628670609/job/18006277873

A true mistery.

[why-not-try-calmer]

Ah I bet you 10 lev that it's material-insiders replacing the correct dependencies with incorrect ones.

[suricactus]

Ok, locally I managed to reproduce, at least for me it uses the wrong mkdocs command. We should revert installing the extensions module too, see the first paragraph https://pypi.org/project/mkdocs-material-extensions/ .

Will debug later tonight.

[why-not-try-calmer]

Ok, locally I managed to reproduce, at least for me it uses the wrong mkdocs command. We should revert installing the extensions module too, see the first paragraph https://pypi.org/project/mkdocs-material-extensions/ .

Will debug later tonight.

Okay but if it uses the wrong mkdocs command, I wonder why https://github.com/opengisch/QField-docs/actions/runs/6628670609/job/18006277873 did not fail? For me the reasoning is:

• the version that declares mkdocs-material-extensions https://github.com/opengisch/QField-docs/actions/runs/6628670609/job/18006277873 does not fail (remotely or locally);
• the version that didn't declare mkdocs-material-extensions failed (remotely and locally);
• therefore it's fine declaring mkdocs-material-extensions;
• however mkdocs-insiders re-installs some dependencies that might not match in versions with the combination of versions that work;
• therefore Adjust mkdocs-insiders version #390 might be worth trying out.