From e158655f4f28a73358b9e24e82963a193d9fc30e Mon Sep 17 00:00:00 2001
From: Alexander Medvednikov <alexander@medvednikov.com>
Date: Tue, 21 Apr 2026 16:46:11 +0300
Subject: [PATCH] net: fix net.html parser ignoring inner tags in tag.content
 (fixes #14138)

---
 vlib/net/html/README.md   | 13 +++++++++++++
 vlib/net/html/dom.v       | 19 +++++++++++++++++++
 vlib/net/html/html_test.v | 12 +++++++++++-
 vlib/net/html/tag.v       | 23 +++++++++++++++++++++--
 4 files changed, 64 insertions(+), 3 deletions(-)

diff --git a/vlib/net/html/README.md b/vlib/net/html/README.md
index a4f503f44..8aa6c74ce 100644
--- a/vlib/net/html/README.md
+++ b/vlib/net/html/README.md
@@ -15,4 +15,17 @@ fn main() {
 }
 ```
 
+For tags with nested markup, `tag.content` preserves the inner HTML:
+
+```v
+import net.html
+
+fn main() {
+	doc := html.parse('<p>before <span>in between</span> after</p>')
+	tag := doc.get_tags(name: 'p')[0]
+	println(tag.content) // before <span>in between</span> after
+	println(tag.text()) // before in between after
+}
+```
+
 More examples found on [`parser_test.v`](parser_test.v) and [`html_test.v`](html_test.v)
diff --git a/vlib/net/html/dom.v b/vlib/net/html/dom.v
index 435d3f323..028dd3583 100644
--- a/vlib/net/html/dom.v
+++ b/vlib/net/html/dom.v
@@ -1,6 +1,7 @@
 module html
 
 import os
+import strings
 
 // The W3C Document Object Model (DOM) is a platform and language-neutral
 // interface that allows programs and scripts to dynamically access and
@@ -172,6 +173,24 @@ fn (mut dom DocumentObjectModel) construct(tag_list []&Tag) {
 		}
 	} // println(tag_list[root_index]) for debug purposes
 	dom.root = tag_list[0]
+	mut root := dom.root
+	dom.normalize_tag_content(mut root)
+}
+
+fn (mut dom DocumentObjectModel) normalize_tag_content(mut tag Tag) {
+	tag.text_content = tag.content
+	tag.content_is_inner_html = true
+	if tag.children.len == 0 {
+		return
+	}
+	mut inner_html := strings.new_builder(tag.content.len + tag.children.len * 32)
+	inner_html.write_string(tag.text_content)
+	for idx := 0; idx < tag.children.len; idx++ {
+		mut child := tag.children[idx]
+		dom.normalize_tag_content(mut child)
+		inner_html.write_string(child.str())
+	}
+	tag.content = inner_html.str()
 }
 
 // get_root returns the root of the document.
diff --git a/vlib/net/html/html_test.v b/vlib/net/html/html_test.v
index 2f30fe31d..f66927ac3 100644
--- a/vlib/net/html/html_test.v
+++ b/vlib/net/html/html_test.v
@@ -26,7 +26,17 @@ fn test_parse_inline_tags() {
 	assert p_tags.len == 1
 
 	p_tag := p_tags[0]
-	assert p_tag.str() == '<p>before <span>in between</span><text> after</text></p>'
+	assert p_tag.content == 'before <span>in between</span> after'
+	assert p_tag.str() == '<p>before <span>in between</span> after</p>'
 
 	assert p_tag.text() == 'before in between after'
 }
+
+fn test_parse_content_with_nested_tags() {
+	doc := parse('<tr id="line"><td><a href="#">wom<nobr>All</a></td></tr>')
+	td_tag := doc.get_tags(name: 'td')[0]
+	a_tag := td_tag.get_tags('a')[0]
+	assert td_tag.content == '<a href="#" >wom<nobr>All</nobr></a>'
+	assert a_tag.content == 'wom<nobr>All</nobr>'
+	assert td_tag.text() == 'womAll'
+}
diff --git a/vlib/net/html/tag.v b/vlib/net/html/tag.v
index 7059b4f5e..df52e85fe 100644
--- a/vlib/net/html/tag.v
+++ b/vlib/net/html/tag.v
@@ -22,6 +22,9 @@ pub mut:
 	position_in_parent int
 	closed             bool
 	close_type         CloseTagType = .in_name
+mut:
+	text_content          string
+	content_is_inner_html bool
 }
 
 fn (mut tag Tag) add_parent(t &Tag, position int) {
@@ -36,18 +39,34 @@ fn (mut tag Tag) add_child(t &Tag) int {
 
 // text returns the text contents of the tag.
 pub fn (tag &Tag) text() string {
+	if tag.name == 'text' {
+		return tag.leading_text()
+	}
 	if tag.name.len >= 2 && tag.name[..2] == 'br' {
 		return '\n'
 	}
 	mut text_str := strings.new_builder(200)
-	text_str.write_string(tag.content)
+	text_str.write_string(tag.leading_text())
 	for child in tag.children {
 		text_str.write_string(child.text())
 	}
 	return text_str.str()
 }
 
+fn (tag &Tag) leading_text() string {
+	if tag.text_content.len > 0 {
+		return tag.text_content
+	}
+	if !tag.content_is_inner_html || tag.children.len == 0 || tag.name == 'text' {
+		return tag.content
+	}
+	return ''
+}
+
 pub fn (tag &Tag) str() string {
+	if tag.name == 'text' {
+		return tag.leading_text()
+	}
 	mut html_str := strings.new_builder(200)
 	html_str.write_string('<${tag.name}')
 	for key, value in tag.attributes {
@@ -58,7 +77,7 @@ pub fn (tag &Tag) str() string {
 	}
 	html_str.write_string(if tag.closed && tag.close_type == .in_name { '/>' } else { '>' })
 	html_str.write_string(tag.content)
-	if tag.children.len > 0 {
+	if !tag.content_is_inner_html && tag.children.len > 0 {
 		for child in tag.children {
 			html_str.write_string(child.str())
 		}
-- 
2.39.5