From 16427b0aea03d09e6c3ef41b41f0144fe8b741ff Mon Sep 17 00:00:00 2001
From: Jason Frey <fryguy9@gmail.com>
Date: Thu, 9 Jul 2020 10:34:04 -0400
Subject: [PATCH] Add support for canonical URLs

Fixes #844
---
 lib/miq/ref_page.rb               |  1 +
 lib/miq/ref_versions.rb           |  7 ++++++-
 site/_includes/head.html          |  4 ++--
 site/_plugins/canonical_url.rb    | 20 ++++++++++++++++++++
 test/lib/miq/ref_versions_test.rb | 13 +++++++++++++
 5 files changed, 42 insertions(+), 3 deletions(-)
 create mode 100644 site/_plugins/canonical_url.rb

diff --git a/lib/miq/ref_page.rb b/lib/miq/ref_page.rb
index a06a1bac3..d1175fb6b 100644
--- a/lib/miq/ref_page.rb
+++ b/lib/miq/ref_page.rb
@@ -31,6 +31,7 @@ def initialize(site, base, dir, name)
       data['branch'] = @branch || data['doc_branch']
       data['legacy'] = @legacy
       data['branch_paths'] = RefVersions.instance.paths_for(@source)
+      data['canonical_url'] = RefVersions.instance.canonical_path_for(@source)
     end
 
     # Only render with liquid if markdown file
diff --git a/lib/miq/ref_versions.rb b/lib/miq/ref_versions.rb
index 3cab359c2..33ab8c2ce 100644
--- a/lib/miq/ref_versions.rb
+++ b/lib/miq/ref_versions.rb
@@ -21,7 +21,12 @@ def paths_for(path)
 
     def canonical_path_for(path)
       recent_path = self[path].first
-      path_version(recent_path) == "latest" ? recent_path : nil
+
+      # TODO: This case should not happen, but some docs do not conform to versioned docs (e.g. /api/)
+      #       Remove this when all docs conform.
+      recent_path ||= path.to_s.sub(%r{/docs/reference/(?:\w+)/}, "/docs/reference/latest/")
+
+      recent_path
     end
 
     private
diff --git a/site/_includes/head.html b/site/_includes/head.html
index 72592e244..309132509 100644
--- a/site/_includes/head.html
+++ b/site/_includes/head.html
@@ -7,9 +7,9 @@
   <title>{{ site.title | escape }}{% if page.title %} - {{ page.title | escape }}{% endif %}</title>
   <meta name="description" content="{% if page.excerpt %}{{ page.excerpt | strip_html | strip_newlines | truncate: 160 }}{% else %}{{ site.description }}{% endif %}">
   <meta property="og:image" content="http://manageiq.org/assets/images/logo/manageiq-logo-blue.png" />
-  
+
   {% css main %}
-  <link rel="canonical" href="{{ page.url | replace:'index.html','' | prepend: site.baseurl | prepend: site.url }}">
+  <link rel="canonical" href="{{ page.canonical_url | replace:'index.html','' | prepend: site.baseurl | prepend: site.url }}">
   <link rel="alternate" type="application/rss+xml" title="{{ site.title }}" href="{{ "/feed.xml" | prepend: site.baseurl | prepend: site.url }}">
 
   <link href="/assets/images/favicon.ico" rel="shortcut icon">
diff --git a/site/_plugins/canonical_url.rb b/site/_plugins/canonical_url.rb
new file mode 100644
index 000000000..92c528f12
--- /dev/null
+++ b/site/_plugins/canonical_url.rb
@@ -0,0 +1,20 @@
+# Jekyll::Page#url is memoized on first call, but this causes issues with the
+#   paginator.  As such, we get the URL and then remove the memoization.
+def extract_page_url(page)
+  had_url = page.instance_variable_defined?(:@url)
+  page.url.tap do
+    page.instance_variable_set(:@url, nil) unless had_url
+  end
+end
+
+Jekyll::Hooks.register :pages, :post_init do |page|
+  page.data['canonical_url'] ||= extract_page_url(page)
+end
+
+Jekyll::Hooks.register :posts, :post_init do |post|
+  post.data['canonical_url'] ||= extract_page_url(post)
+end
+
+Jekyll::Hooks.register :documents, :post_init do |doc|
+  doc.data['canonical_url'] ||= extract_page_url(doc)
+end
diff --git a/test/lib/miq/ref_versions_test.rb b/test/lib/miq/ref_versions_test.rb
index edd77710e..219d8dfc0 100644
--- a/test/lib/miq/ref_versions_test.rb
+++ b/test/lib/miq/ref_versions_test.rb
@@ -26,4 +26,17 @@ def test_paths_for_with_ref_doc
     assert_equal "latest", paths.keys.first
     assert_equal "euwe",   paths.keys.last
   end
+
+  def test_canonical_path_for_with_ref_doc
+    path = Miq::RefVersions.instance.canonical_path_for("/docs/reference/latest/installing_on_google_compute_engine/index.html")
+    assert_equal "/docs/reference/latest/installing_on_google_compute_engine/index.html", path
+
+    path = Miq::RefVersions.instance.canonical_path_for("/docs/reference/jansa/installing_on_google_compute_engine/index.html")
+    assert_equal "/docs/reference/latest/installing_on_google_compute_engine/index.html", path
+  end
+
+  def test_canonical_path_for_with_legacy_ref_doc
+    path = Miq::RefVersions.instance.canonical_path_for("/docs/reference/ivanchuk/doc-Installing_on_Google_Compute_Engine/miq/index.html")
+    assert_equal "/docs/reference/latest/installing_on_google_compute_engine/index.html", path
+  end
 end