Test RSS feeds

[zaknesler]

Here's a list of RSS feeds with weird behavior that we want to make sure we handle nicely.

https://astro.build/rss.xml

• Good example of not containing article content (Astro straight up just can't render article content yet).
• Contains only an svg favicon.
• Want to scrape article content, see Scrape site content if article doesn't contain HTML #7, and save to a separate field, so we can at least render the article content.

https://code.visualstudio.com/feed.xml

• This was handled in fix fallback and debounce calls #27 already, but this feed doesn't give published_at dates, only updated_at.
• This was why I didn't want to make published_at a required field. We should just allow both to be set and COALESCE(published_at, updated_at) in the DB calls and published_at || updated_at on the UI.

https://petapixel.com/feed
https://devblogs.microsoft.com/typescript/feed

• Messy summaries, we want to only keep raw text, <i>, <em>, <b>, <strong>, and maybe links. See Sanitize summary content #16.
• Want to sanitize HTML and strip out images, <p>, etc.

https://www.youtube.com/feeds/videos.xml?channel_id=UC0intLFzLaudFG-xAvUEO-A

• YouTube RSS feeds don't contain any text content but do have a description that we can use as the article content, maybe.

• Further, it points out that RSS feed content can come in various mimetypes (this is handled by feed-rs already):

  <media:group>
    <media:title>Amsterdam Closed This Bridge to Cars (but not bikes ofc)</media:title>
    <media:content url="https://www.youtube.com/v/g0F_hTGYa0Y?version=3" type="application/x-shockwave-flash" width="640" height="390"/>
    <media:thumbnail url="https://i4.ytimg.com/vi/g0F_hTGYa0Y/hqdefault.jpg" width="480" height="360"/>
    <media:description>[...]</media:description>
    <media:community>
        <media:starRating count="23326" average="5.00" min="1" max="5"/>
        <media:statistics views="421127"/>
    </media:community>
  </media:group>
  

• So we might want to have various content types in the feed model and the DB model for the specific mimetypes. Then the frontend can render it in the best way for that type.

https://feeds.megaphone.fm/TPC3838283892

• This one we might not be able to do anything about w/o breaking other feeds, but the summary here is multi-line with links galore. Ideally we'd just show a single line.
• We'll probably not do anything about this but I wanted to mention it.
• I guess we should actually convert plaintext content into html, so maybe run it through a markdown -> html converter.
• Also R.I.P. the sunny podcast :(( those stupid motherfuckers lost sight big time

https://www.snapsbyfox.com/blog?format=rss
https://store.steampowered.com/feeds/news/app/2198150

• See Logic to treat summary_html as content_html #42
• The article body is in the summary tag.

https://gamedev.rs/rss.xml

• The images are relative links but are relative to the blog post itself, e.g. an image URL of "image.jpg" on the page https://gamedev.rs/news/050 should result in the URL replaced by https://gamedev.rs/news/050/image.jpg not https://gamedev.rs/image.jpg.
• See Fix relative URLs #48