🎨 Improve HTML clipping siyuan-note/siyuan#13353

h2m.go

@@ -358,16 +358,27 @@ func (lute *Lute) genASTByDOM(n *html.Node, tree *parse.Tree) {
 			return
 		}
 
-		if atom.Div == firstc.DataAtom && nil != firstc.NextSibling && atom.Code == firstc.NextSibling.DataAtom {
-			firstc = firstc.NextSibling
-			n.FirstChild.Unlink()
+		codes := util.DomChildrenByType(n, atom.Code)
+		if 0 < len(codes) {
+			// 删除第一个 code 之前的标签
+			unlinks := []*html.Node{}
+			for prev := codes[0].PrevSibling; nil != prev; prev = prev.PrevSibling {
+				unlinks = append(unlinks, prev)
+			}
+			for _, unlink := range unlinks {
+				unlink.Unlink()
+			}
+			firstc = n.FirstChild
+			if nil == firstc {
+				return
+			}
 		}
 
 		if atom.Em == firstc.DataAtom && nil != firstc.NextSibling && atom.Em == firstc.NextSibling.DataAtom {
 			// pre.em,em,code 的情况，这两个 em 是“复制代码”和“隐藏代码” https://github.com/siyuan-note/siyuan/issues/13026
-			code := util.DomChildrenByType(n, atom.Code)
-			if 0 < len(code) {
-				firstc = code[0]
+
+			if 0 < len(codes) {
+				firstc = codes[0]
 				if nil != firstc {
 					unlinks := []*html.Node{}
 					for prev := firstc.PrevSibling; nil != prev; prev = prev.PrevSibling {
@@ -381,7 +392,6 @@ func (lute *Lute) genASTByDOM(n *html.Node, tree *parse.Tree) {
 		}
 
 		if atom.Div == firstc.DataAtom && nil == firstc.NextSibling {
-			codes := util.DomChildrenByType(n, atom.Code)
 			if 1 == len(codes) {
 				code := codes[0]
 				// pre 下只有一个 div，且 div 下只有一个 code，那么将 pre.div 替换为 pre.code https://github.com/siyuan-note/siyuan/issues/11131

test/h2m_test.go

@@ -18,6 +18,8 @@ import (
 
 var html2MdTests = []parseTest{
 
+	{"211", "<pre><textarea></textarea><span class=\"copyTooltip\" style=\"display: inline;\"><img src=\"https://learnblockchain.cn/css/default/copy.svg\"></span><code class=\"language-go hljs\"><span class=\"hljs-keyword\">type</span> PurchaseOrder <span class=\"hljs-keyword\">struct</span> {  \n     Id      <span class=\"hljs-type\">uint</span>      <span class=\"hljs-string\">`json:\"id\" gorm:\"primaryKey\"`</span>  \n     UserId  <span class=\"hljs-type\">string</span>    <span class=\"hljs-string\">`json:\"userId\"`</span>  \n     Product []Product  <span class=\"hljs-string\">`json:\"product\" gorm:\"foreignKey:Id\"`</span>  \n     Date    time.Time <span class=\"hljs-string\">`json:\"date\"`</span>  \n}\n</code></pre>", "```go\ntype PurchaseOrder struct {  \n     Id      uint      `json:\"id\" gorm:\"primaryKey\"`  \n     UserId  string    `json:\"userId\"`  \n     Product []Product  `json:\"product\" gorm:\"foreignKey:Id\"`  \n     Date    time.Time `json:\"date\"`  \n}\n```\n"},
+	{"210", "<pre><textarea></textarea><code class=\"language-go hljs\">this is code</code></pre>", "```go\nthis is code\n```\n"},
 	{"209", "<figure> <img alt=\"c03f001.eps\" class=\"center\" src=\"https://www.w3schools.com/tags/pic_trulli.jpg\"> <figcaption><p><a id=\"figure3-1\" href=\"#figureanchor3-1\"><b>Figure 3-1:</b></a> If you try to design an automobile that pleases every possible driver, you end up with a car with every possible feature that pleases nobody. Software today is too often designed to please too many users, resulting in low user satisfaction. <a id=\"figureanchor3-2\" href=\"#figure3-2\">Figure 3-2</a> provides an alternative approach.</p></figcaption> </figure>", "![c03f001.eps](https://www.w3schools.com/tags/pic_trulli.jpg \"Figure 3-1: If you try to design an automobile that pleases every possible driver, you end up with a car with every possible feature that pleases nobody. Software today is too often designed to please too many users, resulting in low user satisfaction. Figure 3-2 provides an alternative approach.\")\n"},
 	{"208", "foo<h1>bar</h1>", "foo\n# bar\n"},
 	{"207", "foo<ul><li>bar<ul><li>baz</li></ul></li></ul>", "foo\n* bar\n  * baz\n"},