Internal links point to page.html, but the true location is just page (Page Not Found)

[daviewales]

Thank you for reporting an issue.

I think this is a broader issue, but for example, on the following page:
https://search.opentofu.org/provider/hashicorp/azurerm/latest/docs/resources/linux_virtual_machine

If I click any of the internal links (within the registry), I get page not found.
This is because the internal links point (for example) to:
https://search.opentofu.org/provider/hashicorp/azurerm/latest/docs/resources/virtual_machine.html

But the correct location is:
https://search.opentofu.org/provider/hashicorp/azurerm/latest/docs/resources/virtual_machine

This pattern is repeated across other links too.

There are also broken links for which the trick above doesn't work.
For example, the page above links to:
https://search.opentofu.org/docs/state/sensitive-data.html

But the correct location is:
https://opentofu.org/docs/language/state/sensitive-data/

OpenTofu Version

Not a CLI issue.

OpenTofu Configuration Files

Not a CLI configuration issue.

[abstractionfactory]

Hi @daviewales thank you for this issue, this is one of the problems we still need to solve as these links come from the original Markdown files created by HashiCorp.

[abstractionfactory]

Hey folks, we would like to either implement this as the core team, or we are happy to accept a community contribution for this.

The way we would like to implement this is by creating a worker that converts the original links into links pointing to the correct markdown file in the API, and then the frontend should turn these links into the frontend's path understanding.

[daviewales]

Looking at the markdown, there appear to be two types of links:

• Relative links such as: virtual_machine.html. This already gets the correct path, and just needs to lose the .html suffix.
• Absolute links such as: /docs/state/sensitive-data.html. This needs to map to https://opentofu.org/docs/language/state/sensitive-data/, which is slightly more complicated.

[abstractionfactory]

@daviewales the situation is a bit more complicated. For example, a markdown document could have a relative link to /r/some_resource.md, which we need to map to the new storage scheme we use in our API. The source path schemes are documented here. The target schemes are in the API.