Accessing page links through a different URL than base-url fails

[monti-python]

What happened?

In our company we access the same Confluence instance through different domains depending on the authentication scheme. However, mark hardcodes the entire URL (including the base-url) whenever a Confluence page from the same instance is referenced via relative paths, creating a hard reference to the base-url that was specified at the time mark was invoked.

This can be fixed by removing the hardcoded reference to base-url as Confluence supports relative links to refer to pages under the same Confluence instance.

This would also make the pages and links independent from the current base-url in general, allowing for more flexibility.

What did you expect to happen?

Links to Confluence pages within the same page should always work regardless of the Confluence's hostname/domain.

How can we reproduce the behavior you experienced?

Steps to reproduce the behavior:

1. Create a page referencing a relative file containing another page, e.g.:

   <!-- Title: PageWithRef -->
   
   See this [page](../path/to/other/page.md)

2. Create a DNS alias for your Confluence instance

3. Deploy the page above using the new alias as base-url

4. Observe how the references above points to the new base-url (DNS alias), even in the regular URL

Further Information (please complete the following information)

• Mark Version (mark --version): 9.12.0

[monti-python]

Created a PR to solve it: #457

[monti-python]

Solved in #457

[monti-python]

Reopening due to #468 #469

[monti-python]

@mrueg @st-karol As a simple mechanism to form relative links we could simply extract the path from the base_url, e.g.:

link = url.Parse(api.BaseURL).Path + page.Links.Full

That should be robust enough to handle both Confluence Cloud (with the /wiki subpath) and Confluence Server (without a subpath), assuming base_url includes the /wiki subpath for Confluence Cloud.

What do you think?

[st-karol]

@monti-python

	if page != nil {
		var path, err = url.Parse(api.BaseURL)
		if nil == err {
			link = path.Path + page.Links.Full
		}
	}

Works perfectly fine in our case (cloud). Correct relative link is generated 💥 Go for it if that works with your setup as well 🙇

[mrueg]

Sure, feel free to propose the suggestion as a PR!

[monti-python]

@st-karol it'd be great if you could have a look at #477 as well 🙏

[mrueg]

Closing as #477 has been merged. Thanks for your contribution!