jsxtreme-html

[davidtheclark]

@jfirebaugh needs some jsxtreme-html, for all the same reasons somebody might want jsxtreme-markdown plus wanting to write the source code in HTML, instead of Markdown.

The code for jsxtreme-html would be only slightly different from that for jsxtreme-markdown — I wonder how much they could share. So I think we could accomplish this without too much fuss.

[davidtheclark]

Thought about this some more and I'm changing my mind. I don't think jsxtreme-html would fit the use case, actually. Here's what's needed:

• Create pages that are HTML partials with front matter.
• In that front matter they specify a wrapper React component, just like Markdown pages.
• The wrapper React component receives the HTML as a string (not as JSX) and does what it will with that string.

There may actually be a much simpler way to address this need. I'm going to try this:

• If a file in the pages directory is HTML without front matter, leave it as-is (which means it will be copied into the corresponding place in the output directory).
• If a file in the pages directory is HTML with front matter, use gray-matter to parse it. Pass the result of the parsing — which includes the body of the file as a string — to the designated wrapper component.
• That component that then
  • Manipulate the string (e.g. use regex to interpolate dynamic values).
  • Render it as an HTML code string.
  • Use dangerouslySetInnerHTML to render it onto the page.

[davidtheclark]

Started work on this and realized how much code it will add (and need for examples and documentation and tests). I'm wondering if this use case will be common enough to justify that. Because the simpler alternative exists to create a JS page and use the raw-loader to import an HTML file — the only downside being that you create two files instead of one.

[jfirebaugh]

That's totally reasonable, the only other cases I know of that might need something like this are Mapbox.js (if its docs ever get converted to batfish) or something like http://turfjs.org/docs/.

I'll take a look at the raw-loader approach.

[davidtheclark]

Ok 👍 Closing this issue in favor of https://github.com/mapbox/batfish/pull/195, which introduces ignoreWithinPagesDirectory so you can colocate HTML files with the pages that import them.