fix: match hugo's (bizzare) handling of short markdown strings

CloudCannon · Feb 2, 2022 · f14261f · f14261f
1 parent b13997d
commit f14261f
Show file tree

Hide file tree

Showing 3 changed files with 44 additions and 2 deletions.
diff --git a/javascript-modules/engines/hugo-engine/hugo-renderer/helpers/bookshop_modified_content.go b/javascript-modules/engines/hugo-engine/hugo-renderer/helpers/bookshop_modified_content.go
@@ -18,13 +18,21 @@
 package helpers
 
 import (
+	"bytes"
 	"html/template"
 	"strings"
 	"unicode"
 
 	bp "github.com/gohugoio/hugo/bufferpool"
 )
 
+var (
+	openingPTag        = []byte("<p>")
+	closingPTag        = []byte("</p>")
+	paragraphIndicator = []byte("<p")
+	closingIndicator   = []byte("</")
+)
+
 var stripHTMLReplacer = strings.NewReplacer("\n", " ", "</p>", "\n", "<br>", "\n", "<br />", "\n")
 
 // StripHTML accepts a string, strips out all HTML tags and returns it.
@@ -68,3 +76,23 @@ func StripHTML(s string) string {
 func BytesToHTML(b []byte) template.HTML {
 	return template.HTML(string(b))
 }
+
+// TrimShortHTML removes the <p>/</p> tags from HTML input in the situation
+// where said tags are the only <p> tags in the input and enclose the content
+// of the input (whitespace excluded).
+// Bookshop: Changed this to _not_ be a method of *ContentSpec
+func TrimShortHTML(input []byte) []byte {
+	firstOpeningP := bytes.Index(input, paragraphIndicator)
+	lastOpeningP := bytes.LastIndex(input, paragraphIndicator)
+
+	lastClosingP := bytes.LastIndex(input, closingPTag)
+	lastClosing := bytes.LastIndex(input, closingIndicator)
+
+	if firstOpeningP == lastOpeningP && lastClosingP == lastClosing {
+		input = bytes.TrimSpace(input)
+		input = bytes.TrimPrefix(input, openingPTag)
+		input = bytes.TrimSuffix(input, closingPTag)
+		input = bytes.TrimSpace(input)
+	}
+	return input
+}
diff --git a/javascript-modules/engines/hugo-engine/hugo-renderer/tpl/transform/transform.go b/javascript-modules/engines/hugo-engine/hugo-renderer/tpl/transform/transform.go
@@ -99,6 +99,9 @@ func (ns *Namespace) Markdownify(s interface{}) (template.HTML, error) {
 	// This includes using a different (smaller) markdown implementation.
 	output := markdown.ToHTML([]byte(ss), nil, nil)
 
+	// Strip if this is a short inline type of text.
+	output = helpers.TrimShortHTML(output)
+
 	return helpers.BytesToHTML(output), nil
 }
 

diff --git a/...t-modules/integration-tests/features/hugo/live_editing/hugo_bookshop_live_filters.feature b/...t-modules/integration-tests/features/hugo/live_editing/hugo_bookshop_live_filters.feature
@@ -18,11 +18,12 @@ Feature: Hugo Bookshop CloudCannon Live Editing Filters and Functions
   Scenario: Bookshop live renders markdown
     Given a component-lib/components/beetroot/beetroot.hugo.html file containing:
       """
-      <div><code>{{ .md | markdownify }}</code></div>
+      <div>{{ .md | markdownify }}</div>
       """
     Given [front_matter]:
       """
       md: title
+      multi_md: title
       """
     And a site/content/_index.md file containing:
       """
@@ -36,15 +37,25 @@ Feature: Hugo Bookshop CloudCannon Live Editing Filters and Functions
       <body>
       {{ partial "bookshop_bindings" `(dict "md" .Params.md )` }}
       {{ partial "bookshop" (slice "beetroot" (dict "md" .Params.md )) }}
+
+      {{ partial "bookshop_bindings" `(dict "md" .Params.multi_md )` }}
+      {{ partial "bookshop" (slice "beetroot" (dict "md" .Params.multi_md )) }}
       </body>
       </html>
       """
     And 🌐 I have loaded my site in CloudCannon
     When 🌐 CloudCannon pushes new yaml:
       """
       md: "**bold** title"
+      multi_md: >-
+        # Hello  
+
+        World
       """
     Then 🌐 There should be no errors
     *    🌐 There should be no logs
-    *    🌐 The selector code should match "<p><strong>bold</strong> title</p>"
+    # Hugo strips single <p> tags
+    *    🌐 The selector div:nth-of-type(1) should match "<strong>bold</strong> title"
+    *    🌐 The selector div:nth-of-type(2) should match "<h1>Hello</h1>"
+    *    🌐 The selector div:nth-of-type(2) should match "<p>World</p>"