Represent SafeString as plain string on schema rendering.

[hashlash]

Note: Before submitting this pull request, please review our contributing guidelines.

Description

Fix #8428.

[hashlash]

I tried running tox -e docs on branch master, and it also failed (with the same error). So the problem shouldn't be from the changes in this PR.

Should we wait for the fix, or can we merge it first?

[hashlash]

The source of error: mkdocs/mkdocs#2799

[hashlash]

The error should be fixed in #8433.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[auvipy]

I don't have much depth regarding yaml. will be waiting for more eyes into this. on a side note, can you add more relevant tests for the changes?

[hashlash]

can you add more relevant tests for the changes?

@auvipy I think the test is sufficient. As long as it can render SafeString, I think it's good enough.

Or could you elaborate on what kind of test I should add?

[auvipy]

I have requested review from another OpenAPI specialist

[hashlash]

Btw, here's the section in PyYAML that discuss about representing custom object: https://pyyaml.org/wiki/PyYAMLDocumentation#constructors-representers-resolvers

[tfranzel]

Yes that works. It is key that the yaml does not contain !! type injections, which are valid in newer YAML versions, but upstream tools mostly choke on them.

But just as a sidenote: there are a bunch more objects (e.g uuid, datetime, etc) that solicit this kind of behavior: https://github.com/tfranzel/drf-spectacular/blob/03677b6bb86b2076a97b70b918ed3b3b528edaa4/drf_spectacular/renderers.py#L52

[hashlash]

Oops, you're right. I just read the discussion #8453, but wasn't aware with the drf-spectacular renderer implementation. Based on the discussion, going forward, the renderer part will also be deprecated from rest_framework?

[tfranzel]

@hashlash, yes. everything that does not belong to the schema interface itself will be retired and removed eventually.