fix html not included in pre code block

commonmark.go

@@ -11,7 +11,6 @@ import (
 
 	"github.com/JohannesKaufmann/html-to-markdown/escape"
 	"github.com/PuerkitoBio/goquery"
-	"golang.org/x/net/html"
 )
 
 var multipleSpacesR = regexp.MustCompile(`  +`)
@@ -268,12 +267,7 @@ var commonmark = []Rule{
 	{
 		Filter: []string{"code"},
 		Replacement: func(_ string, selec *goquery.Selection, opt *Options) *string {
-			code, err := selec.Html()
-			if err != nil {
-				return nil
-			}
-			// We don't want the html encoded characters to be displayed as is.
-			code = html.UnescapeString(code)
+			code := getHTML(selec)
 
 			// Newlines in the text aren't great, since this is inline code and not a code block.
 			// Newlines will be stripped anyway in the browser, but it won't be recognized as code
@@ -309,9 +303,9 @@ var commonmark = []Rule{
 			language := codeElement.AttrOr("class", "")
 			language = strings.Replace(language, "language-", "", 1)
 
-			code := codeElement.Text()
+			code := getHTML(codeElement)
 			if codeElement.Length() == 0 {
-				code = selec.Text()
+				code = getHTML(selec)
 			}
 
 			fenceChar, _ := utf8.DecodeRuneInString(opt.Fence)

from.go

@@ -107,6 +107,7 @@ func NewConverter(domain string, enableCommonmark bool, options *Options) *Conve
 
 	conv.before = append(conv.before, func(selec *goquery.Selection) {
 		selec.Find("a[href]").Each(func(i int, s *goquery.Selection) {
+			// TODO: don't hardcode "data-index" and rename it to avoid accidental conflicts
 			s.SetAttr("data-index", strconv.Itoa(i+1))
 		})
 	})

testdata/TestCommonmark/pre_code/goldmark.golden

@@ -29,3 +29,15 @@ totally ~~~~~~ normal
 ~ code
 
 </code></pre>
+<pre><code>
+The &lt;img&gt; tag is used to embed an image.
+
+The &lt;img/&gt; tag is used to embed an image.
+
+</code></pre>
+<pre><code>
+&lt;a href=&quot;#Blabla&quot; data-index=&quot;1&quot;&gt;
+    &lt;img src=&quot;http://bla.bla/img/img.svg&quot; style=&quot;height:auto&quot; width=&quot;200px&quot;/&gt;
+&lt;/a&gt;
+
+</code></pre>

testdata/TestCommonmark/pre_code/input.html

@@ -83,4 +83,19 @@
 Some ~~~
 totally ~~~~~~ normal
 ~ code
-</code></pre>
+</code></pre>
+
+
+<!--code tag with pre and other tags inside-->
+<pre><code>
+The &lt;img&gt; tag is used to embed an image.
+
+The <img> tag is used to embed an image.
+</code></pre>
+
+<!--TODO: make sure that "data-index" does not get added to the link-->
+<pre><code>
+<a href="#Blabla">
+    <img src="http://bla.bla/img/img.svg" style="height:auto" width="200px"/>
+</a>
+</code></pre>

testdata/TestCommonmark/pre_code/output.fenced_backtick.golden

@@ -59,4 +59,20 @@ Some ~~~
 totally ~~~~~~ normal
 ~ code
 
+```
+
+```
+
+The <img> tag is used to embed an image.
+
+The <img/> tag is used to embed an image.
+
+```
+
+```
+
+<a href="#Blabla" data-index="1">
+    <img src="http://bla.bla/img/img.svg" style="height:auto" width="200px"/>
+</a>
+
 ```

testdata/TestCommonmark/pre_code/output.fenced_tilde.golden

@@ -59,4 +59,20 @@ Some ~~~
 totally ~~~~~~ normal
 ~ code
 
-~~~~~~~
+~~~~~~~
+
+~~~
+
+The <img> tag is used to embed an image.
+
+The <img/> tag is used to embed an image.
+
+~~~
+
+~~~
+
+<a href="#Blabla" data-index="1">
+    <img src="http://bla.bla/img/img.svg" style="height:auto" width="200px"/>
+</a>
+
+~~~

testdata/TestCommonmark/pre_code/output.indented.golden

@@ -59,4 +59,20 @@ Some ~~~
 totally ~~~~~~ normal
 ~ code
 
+```
+
+```
+
+The <img> tag is used to embed an image.
+
+The <img/> tag is used to embed an image.
+
+```
+
+```
+
+<a href="#Blabla" data-index="1">
+    <img src="http://bla.bla/img/img.svg" style="height:auto" width="200px"/>
+</a>
+
 ```

utils.go

@@ -259,3 +259,15 @@ func findMax(a []int) (max int) {
 	}
 	return max
 }
+
+// getHTML gets the HTML content and unescapes the encoded characters.
+// Returns "" if there is an error.
+func getHTML(selec *goquery.Selection) string {
+	content, err := selec.Html()
+	if err != nil {
+		return ""
+	}
+
+	// We don't want the html encoded characters to be displayed as is.
+	return html.UnescapeString(content)
+}