Non-conforming HTML elements got us stuck

[dprothero]

Getting this error:

The parser has encountered a text content tag (`<style>`) in the context where it is ambiguous whether this tag should be ignored or not. And, thus, is is unclear is consequent content should be parsed as raw text or HTML markup.

This error occurs due to the limited capabilities of the streaming parsing. However, almost all of the cases of this error are caused by a non-conforming markup (e.g. a `<script>` element in `<select>` element).

Indeed, it is from a <style> tag that is nested under a <select> tag, but there's nothing we can do about it because it's generated by a third-party component library we need to use.

I tried adding data-pagefind-ignore="all" on the container with the offending HTML, but that didn't seem to help.

I've attached an example HTML that throws this error. index_html.txt

Any ideas for a workaround?

[dprothero]

I also tried adding data-pagefind-body to an element after the offending HTML. No luck. :(

index2_html.txt

[bglw]

Unfortunately there is no current way around that — Pagefind parses HTML using the https://github.com/cloudflare/lol-html library, and that error there is a parsing error passed through directly from that library. Since the HTML document itself is failing to parse, it isn't possible for Pagefind to select down to a data-pagefind-body or anything.

Immediately after typing that I realized that lol-html can be initialized with strict: false to continue in the face of these parsing ambiguities. Since indexing pages for search isn't a security risk, I'll disable strict mode for Pagefind 🙌

[bglw]

Tracking as #154

[bglw]

Released as v0.10.1 🎉

[dprothero]

Awesome 🎉