Describe the feature

Problem

Currently, prerender crawler is allowed to check only html and json links: https://github.com/unjs/nitro/blob/fd15c212b197bb035d675ce05c9722e1099bcf33/src/prerender.ts#L18

const allowedExtensions = new Set(["", ".json"])

But sometimes it is also needed to prerender binary files with different extentions (eg pdf, png).

My case

My case requires prerender pdf files, and it is quite hard to create routes manually. Moreover in Nuxt 3 I also can't add prerender routes generated from Nuxt Content.

Possible solution

Solution here can be adding new extentions to allowed or even deprecating filtration: https://github.com/unjs/nitro/blob/fd15c212b197bb035d675ce05c9722e1099bcf33/src/prerender.ts#L258

if (crawlLinks) {
  _links.push(
    ...[...html.matchAll(LINK_REGEX)]
      .map((m) => m[1])
      .filter((link) => allowedExtensions.has(getExtension(link)))
  );
}

Extend functionality

Now all links are crowled with Regex detecting href="somelink. But it also can be useful to detect sources of images: it also can be generated. In this case it is needed to change Regex: https://github.com/unjs/nitro/blob/fd15c212b197bb035d675ce05c9722e1099bcf33/src/prerender.ts#L242

const LINK_REGEX = /href=["']?([^"'>]+)/g;

Additional information

[X] Would you be willing to help implement this feature?

unjs / nitro

Allow prerender crawler find more types of links #1067