net.html: fix text parsing for inline tags (#18085)

vlang · Apr 30, 2023 · 0898f57 · 0898f57
1 parent d086cc2
commit 0898f57
Show file tree

Hide file tree

Showing 2 changed files with 27 additions and 4 deletions.
diff --git a/vlib/net/html/html_test.v b/vlib/net/html/html_test.v
@@ -13,3 +13,20 @@ fn test_parse() {
 	assert h1_tag.str() == '<h1 class="title" >Hello world!</h1>'
 	// assert h1_tag.str() == '<h1 class="title">Hello world!</h1>'
 }
+
+fn test_parse_inline_tags() {
+	doc := parse('<html><body><p>before <span>in between</span> after</p></body></html>')
+	tags := doc.get_tag('span')
+	assert tags.len == 1
+
+	span_tag := tags[0]
+	assert span_tag.str() == '<span>in between</span>'
+
+	p_tags := doc.get_tag('p')
+	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.text() == 'before in between after'
+}
diff --git a/vlib/net/html/parser.v b/vlib/net/html/parser.v
@@ -14,6 +14,7 @@ mut:
 	opened_code_type string
 	line_count       int
 	outside_tag      bool
+	text_after_tag   bool
 	lexeme_builder   strings.Builder = strings.new_builder(100)
 	code_tags        map[string]bool = {
 		'script': true
@@ -221,10 +222,7 @@ pub fn (mut parser Parser) split_parse(data string) {
 				if parser.lexical_attributes.current_tag.name.len > 1
 					&& parser.lexical_attributes.current_tag.name[0] == 47
 					&& !blank_string(temp_string) {
-					parser.tags << &Tag{
-						name: 'text'
-						content: temp_string
-					}
+					parser.lexical_attributes.text_after_tag = true
 				} else {
 					parser.lexical_attributes.current_tag.content = temp_string // verify later who has this content
 				}
@@ -234,6 +232,14 @@ pub fn (mut parser Parser) split_parse(data string) {
 			parser.generate_tag()
 			parser.lexical_attributes.open_tag = true
 			parser.lexical_attributes.outside_tag = false
+
+			if parser.lexical_attributes.text_after_tag == true {
+				parser.tags << &Tag{
+					name: 'text'
+					content: temp_string
+				}
+				parser.lexical_attributes.text_after_tag = false
+			}
 		} else {
 			parser.lexical_attributes.lexeme_builder.write_u8(chr)
 		}