fix a whitespace handling bug

fixes #74
ckruse · Jan 13, 2024 · 65c2524 · 65c2524
1 parent 7538400
commit 65c2524
Show file tree

Hide file tree

Showing 4 changed files with 552 additions and 12 deletions.
diff --git a/lib/microformats2.ex b/lib/microformats2.ex
@@ -125,5 +125,16 @@ defmodule Microformats2 do
     |> String.replace(~r/\012/, "&#x0A;")
     |> String.replace(~r/\013/, "&#x0B;")
     |> Floki.parse_document()
+    |> normalize_tag_names()
   end
+
+  defp normalize_tag_names({:ok, tree}) do
+    {:ok,
+     Floki.traverse_and_update(tree, fn
+       {tag, attrs, children} -> {String.trim(tag), attrs, children}
+       other -> other
+     end)}
+  end
+
+  defp normalize_tag_names(other), do: other
 end
diff --git a/test/documents/real_world_nesting.html b/test/documents/real_world_nesting.html
@@ -0,0 +1,339 @@
+<!DOCTYPE html>
+<html lang="en-US">
+  <head prefix="og: http://ogp.me/ns#">
+    <meta charset="utf-8" />
+    <meta http-equiv="X-UA-Compatible" content="IE=edge" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <meta name="mobile-web-app-capable" content="yes" />
+    <meta name="apple-mobile-web-app-capable" content="yes" />
+    <meta name="msapplication-TileColor" content="#434279" />
+    <meta name="description" content="Web developer, video gamer, tabletop gamer, and comic reader" />
+
+    <title>A boost of &#39;gilest.org: Make the indie web easier&#39;</title>
+    <!-- <link rel="favicon" href="/favicon.ico" /> -->
+    <link rel="icon favicon" href="/images/pint.svg" />
+    <link rel="stylesheet" href="/css/base.css" />
+    <link rel="stylesheet" href="/css/pub.css" />
+    <link rel="stylesheet" href="/css/utils.css" />
+    <link rel="stylesheet" href="/css/scroll.css" />
+    <link rel="stylesheet" href="/css/print.css" media="print" />
+
+      <link rel="stylesheet" href="/css/note.css" />
+
+
+    <link rel="manifest" href="/manifest.json" />
+    <!-- <link rel="search" href="/opensearch.xml" title="Adhoc Systems" type="application/opensearchdescription+xml" /> -->
+    <link rel="index" href="/" />
+    <link rel="micropub" href="/micropub" />
+    <link rel="microsub" href="https://aperture.p3k.io/microsub/150" />
+    <!-- <link rel="webmention" href="https://webmention.io/adhoc.systems/webmention" /> -->
+    <link rel="webmention" href="https://adhoc.systems/webmention" />
+    <link rel="pingback" href="https://webmention.io/adhoc.systems/xmlrpc" />
+
+      <link rel="me" href="https://twitter.com/0x1C3B00DA" />
+
+      <link rel="me" href="https://toot.cafe/@zack" />
+
+      <link rel="me" href="https://huffduffer.com/0x1C3B00DA" />
+
+    <!--
+    <link rel="authorization_endpoint" href="https://indieauth.com/auth" />
+    <link rel="token_endpoint" href="https://tokens.indieauth.com/token" />
+    -->
+    <link rel="authorization_endpoint" href="https://indieauth.com/auth" />
+    <link rel="token_endpoint" href="https://tokens.indieauth.com/token" />
+
+    <link rel="alternate" href="https://adhoc.systems/boosts/51052d4f-a968-4a72-9150-225134e90423" type="application/activity+json" />
+
+<link
+  rel="alternate"
+  type="application/rss+xml"
+  title="Index (RSS)"
+  href="https://adhoc.systems/feed.xml"
+/>
+
+
+<link
+  rel="alternate"
+  type="application/rss+xml"
+  title="Notes (RSS)"
+  href="https://adhoc.systems/notes/feed.xml"
+/>
+
+
+<link
+  rel="alternate"
+  type="application/rss+xml"
+  title="Articles (RSS)"
+  href="https://adhoc.systems/articles/feed.xml"
+/>
+
+
+<link
+  rel="alternate"
+  type="application/rss+xml"
+  title="Boosts (RSS)"
+  href="https://adhoc.systems/boosts/feed.xml"
+/>
+
+
+<link
+  rel="alternate"
+  type="application/rss+xml"
+  title="Likes (RSS)"
+  href="https://adhoc.systems/likes/feed.xml"
+/>
+
+
+<link
+  rel="alternate"
+  type="application/rss+xml"
+  title="Bookmarks (RSS)"
+  href="https://adhoc.systems/bookmarks/feed.xml"
+/>
+
+
+<link
+  rel="feed"
+  type="text/html"
+  title="Index"
+  href="/"
+/>
+
+
+<link
+  rel="feed"
+  type="text/html"
+  title="Notes - Adhoc Systems"
+  href="/notes"
+/>
+
+
+<link
+  rel="feed"
+  type="text/html"
+  title="Articles - Adhoc Systems"
+  href="/articles"
+/>
+
+
+<link
+  rel="feed"
+  type="text/html"
+  title="Photos - Adhoc Systems"
+  href="/photos"
+/>
+
+
+<link
+  rel="feed"
+  type="text/html"
+  title="Boosts - Adhoc Systems"
+  href="/boosts"
+/>
+
+
+<link
+  rel="feed"
+  type="text/html"
+  title="Likes - Adhoc Systems"
+  href="/likes"
+/>
+
+
+<link
+  rel="feed"
+  type="text/html"
+  title="Bookmarks - Adhoc Systems"
+  href="/bookmarks"
+/>
+
+
+    <meta property="og:site_name" content="Adhoc Systems" />
+    <meta property="og:locale" content="en-US" />
+    <meta property="og:url" content="adhoc.systems" />
+
+      <meta property="og:description" content="" />
+
+  </head>
+  <body class="" >
+    <a class="skip-link" href="#main-content">Skip to main content</a>
+    <div id="top"></div>
+    <div class="scroll-container scroll-container--intersecting"></div>
+    <header class="main-header">
+      <h1 class="p-name" itemprop="name">
+<a class="u-url" href="/" itemprop="url" rel="me" title="Home">Adhoc Systems</a>
+      </h1>
+      <nav class="main-nav flex jc-between">
+        <div>
+<a href="/notes" title="Zachary Dunn&#39;s Notes">Notes</a>
+          <span class="current-color mar-0 pad-0">|</span>
+<a href="/recipes" title="Zachary Dunn&#39;s Recipes">Recipes</a>
+          <span class="current-color mar-0 pad-0">|</span>
+<a href="/about" title="About Zachary Dunn">About Me</a>
+        </div>
+
+        <div class="right">
+
+        </div>
+      </nav>
+    </header>
+    <main class="h-entry" itemscope itemtype="https://schema.org/BlogPosting" >
+
+
+<article class="entry card entry--boost" id="main-content">
+  <header class="note__header">
+    <div class="entry__badge">
+
+
+  <svg class="svg-icon svg-icon--feather">
+    <use xlink:href="/images/lucide-sprite.svg#repeat" />
+  </svg>
+
+
+
+    </div>
+<address class="p-author h-card h-card--mini" itemprop="author" itemscope itemtype="https://schema.org/Person">
+  <a class="u-url" href="/" rel="me" itemprop="url">
+    <img
+      class="u-photo avatar"
+      src="/images/avatar.jpg"
+      alt="Zachary Dunn's avatar."
+      itemprop="image"
+      loading="lazy"
+    />
+  </a>
+  <div class="col">
+    <h1 class="p-name" itemprop="name">Zachary Dunn</h1>
+    <h4 class="m-0">
+      <a href="https://matrix.to/#/@zack:mozilla.org?via=mozilla.org&via=matrix.org">
+        @<span class="p-nickname" itemprop="alternateName">zack</span>:mozilla.org
+      </a>
+    </h4>
+  </div>
+</address>
+
+<h4 class="entry__timestamp timestamp">
+  Published on
+ <a href="https://adhoc.systems/boosts/51052d4f-a968-4a72-9150-225134e90423" class="u-url" itemprop="url"> 
+    <time class="dt-published" datetime="2024-01-09 17:11:10Z" itemprop="datePublished">
+2024-01-09 17:11:10Z
+    </time>
+ </a> 
+</h4>
+
+
+  </header>
+
+
+    <section
+      class="block__inner u-repost-of h-cite"
+      itemprop="citation "
+      itemscope
+      itemtype="https://schema.org/CreativeWork"
+    >
+
+      <div class="entry__remote-authorship">
+        <a
+          href="https://gilest.org/indie-easy.html"
+          class="u-url entry__link"
+          itemprop="url"
+        >
+
+            <span class="p-name" itemprop="headline">gilest.org: Make the indie web easier</span>
+
+        </a>
+
+      </div>
+    </section>
+
+
+
+<section class="tags mar-1">
+
+    <a href="/tags/indieweb" class="p-category hashtag" title="IndieWeb Tag" itemprop="keywords">
+      <svg class="svg-icon svg-icon--feather svg-icon--tag">
+        <use xlink:href="/images/lucide-sprite.svg#tag" />
+      </svg>
+      <span>IndieWeb</span>
+    </a>
+
+    <a href="/tags/tech" class="p-category hashtag" title="tech Tag" itemprop="keywords">
+      <svg class="svg-icon svg-icon--feather svg-icon--tag">
+        <use xlink:href="/images/lucide-sprite.svg#tag" />
+      </svg>
+      <span>tech</span>
+    </a>
+
+    <a href="/tags/decentralization" class="p-category hashtag" title="decentralization Tag" itemprop="keywords">
+      <svg class="svg-icon svg-icon--feather svg-icon--tag">
+        <use xlink:href="/images/lucide-sprite.svg#tag" />
+      </svg>
+      <span>decentralization</span>
+    </a>
+
+</section>
+
+
+
+</article>
+
+<section class="comments mar-1">
+  <h2 class="border-bottom">Comments</h2>
+
+    <p>
+      There aren't any comments on this post yet. To comment, publish your reply online and send
+      a <a href="https://indieweb.org/Webmention">webmention.</a>
+    <p>
+
+
+</section>
+
+    </main>
+
+    <a
+      href="#top"
+      id="scrolltop"
+      class="back-to-top fab"
+      role="button"
+      aria-label="Back to Top"
+      x-data="scrollProgress"
+    >
+      <svg viewBox="0 0 100 100" width="50px" height="50px" tabindex="-1" role="img" aria-labelledby="title">
+        <title id="title">Back To Top</title>
+        <circle
+          class="progress-ring highlight-boost"
+          fill="transparent"
+          r="47"
+          cx="50"
+          cy="50"
+          x-ref="progressCircle"
+          :stroke-dasharray="circumference"
+          :stroke-dashoffset="circumference - percent / 100 * circumference"
+        />
+        <path d="M 50 30 L 25 55 L 30 60 L 50 40 L 70 60 L 75 55 L 50 30 Z" fill="white" stroke="white" />
+      </svg>
+    </a>
+
+    <footer class="main-footer">
+      <span>&copy; Copyright 2018 by  <a href="/" title="Home">Zachary Dunn</a></span>
+      <div class="current-color mar-0 pad-0"></div>
+
+        <div class="rss">
+          <a href="https://adhoc.systems/boosts/feed.xml" rel="alternate" class="rss__link" title="Boosts Feed (RSS)">
+            <span class="visually-hidden">Boosts Feed (RSS)</span>
+            <img
+              src="/images/feed.svg"
+              class="rss__icon"
+              loading="lazy"
+            />
+          </a>
+          <a href="https://aboutfeeds.com/" class="rss__explainer">Help! What is this?</a>
+        </div>
+
+    </footer>
+    <script type="module" src="/js/app.js"></script>
+    <script type="module" src="/js/scroll.js"></script>
+    <script defer src="https://unpkg.com/[email protected]/dist/cdn.min.js"></script>
+  </body>
+</html>