Remark parses <svelte:head> as a link

[babakfp]

Hi 👋

These get parsed as a link:

• <aa:0>
• <aa:a>
• <aa:->
• etc

And these don't:

• </aa:a>
• <a:a>
• <aa:a >
• < aa:a>

Is there a reason for this? I mean, is it a valid thing to parse something that starts with < and ends with > as a link?

If it doesn't hurt the current valid behavior of Markdown, it would be great to get it parsed as html.

[ChristianMurphy]

Hey @babakfp!
We follow the autolink rules commonmark https://spec.commonmark.org/0.31.2/#autolinks

---

The characters < and > are used in Markdown to define autolinks, which are absolute URIs and email addresses enclosed within those characters. The specification defines the rules for what qualifies as a valid autolink.

According to the specification, a URI autolink must start with <, followed by an absolute URI (a scheme, colon, and then characters other than ASCII control characters, space, <, and >) and end with >. Similarly, an email autolink must start with <, followed by an email address (matching a specific regex pattern), and end with >.

The examples you provided that get parsed as links (<aa:0>, <aa:a>, <aa:->) follow this pattern of starting with <, containing a valid URI-like sequence (even though the aa is not a real scheme), and ending with >. Therefore, they are interpreted as autolinks according to the specification.

On the other hand, the examples that don't get parsed as links (</aa:a>, <a:a>, <aa:a >, < aa:a>) violate the rules for autolinks in some way. For instance, </aa:a> starts with an invalid character sequence </, <a:a> is too short to be a scheme, <aa:a > has a space before the closing >, and < aa:a> has a space after the opening <.

Regarding your question about parsing these as HTML, the specification doesn't mention parsing arbitrary strings within <> as HTML tags. It's specific to autolinks, which have a strict definition. Parsing arbitrary <> sequences as HTML tags could potentially cause conflicts with the existing Markdown syntax and might not be desirable in a Markdown context.

[babakfp]

Thank you for the awesome explanation 👍
That makes total sense.
Thanks.