net.html: fix panic in html.parse() called with empty string, remove replacement of \n in the original content (#17206)

walking devel · web-flow · commit 8cdc554c63aa · 2023-02-03T00:32:03.000+02:00
diff --git a/vlib/net/html/dom.v b/vlib/net/html/dom.v
@@ -96,6 +96,12 @@ fn (mut dom DocumentObjectModel) add_tag_by_attribute(tag &Tag) {
 
 fn (mut dom DocumentObjectModel) construct(tag_list []&Tag) {
 	dom.constructed = true
+
+	// If there are no tags, accessing `tag_list` below does panic.
+	if tag_list.len == 0 {
+		return
+	}
+
 	mut temp_map := map[string]int{}
 	mut temp_int := null_element
 	mut temp_string := ''
@@ -106,6 +112,7 @@ fn (mut dom DocumentObjectModel) construct(tag_list []&Tag) {
 	temp_map['0'] = dom.btree.add_children(tag_list[0])
 	stack.push(0)
 	root_index := 0
+
 	for index := 1; index < tag_list.len; index++ {
 		mut tag := tag_list[index]
 		dom.print_debug(tag.str())
diff --git a/vlib/net/html/parser.v b/vlib/net/html/parser.v
@@ -13,6 +13,7 @@ mut:
 	is_attribute     bool
 	opened_code_type string
 	line_count       int
+	outside_tag      bool
 	lexeme_builder   strings.Builder = strings.new_builder(100)
 	code_tags        map[string]bool = {
 		'script': true
@@ -90,6 +91,7 @@ fn (mut parser Parser) init() {
 	parser.tags = []&Tag{}
 	parser.dom.close_tags['/!document'] = true
 	parser.lexical_attributes.current_tag = &Tag{}
+	parser.lexical_attributes.outside_tag = true
 	parser.initialized = true
 }
 
@@ -231,19 +233,40 @@ pub fn (mut parser Parser) split_parse(data string) {
 			parser.lexical_attributes.lexeme_builder.go_back_to(0)
 			parser.generate_tag()
 			parser.lexical_attributes.open_tag = true
+			parser.lexical_attributes.outside_tag = false
 		} else {
 			parser.lexical_attributes.lexeme_builder.write_u8(chr)
 		}
 	}
+
+	// If `data` has not tags but has only text.
+	if parser.lexical_attributes.outside_tag {
+		temp_string := parser.lexical_attributes.lexeme_builder.str()
+
+		if parser.tags.len == 0 {
+			parser.tags << &Tag{
+				name: 'text'
+				content: temp_string
+			}
+		} else if parser.tags.len == 1 {
+			mut tag := parser.tags.first()
+
+			if tag.name == 'text' {
+				tag.content += temp_string
+			}
+		}
+	}
 }
 
 // parse_html parses the given HTML string
 pub fn (mut parser Parser) parse_html(data string) {
 	parser.init()
 	mut lines := data.split_into_lines()
-	for line in lines {
+	for index, line in lines {
 		parser.lexical_attributes.line_count++
-		parser.split_parse(line)
+		// Parser shouldn't replace `\n`, because it may break JS code or text which sticks together.
+		// After `split_into_lines()` we need to add `\n` again.
+		parser.split_parse(if index < lines.len - 1 { '${line}\n' } else { line })
 	}
 	parser.generate_tag()
 	parser.dom.debug_file = parser.debug_file
diff --git a/vlib/net/html/parser_test.v b/vlib/net/html/parser_test.v
@@ -2,6 +2,34 @@ module html
 
 import strings
 
+fn test_parse_empty_string() {
+	mut parser := Parser{}
+
+	parser.parse_html('')
+
+	assert parser.tags.len == 0
+}
+
+fn test_parse_text() {
+	mut parser := Parser{}
+	text_content := 'test\nparse\ntext'
+
+	parser.parse_html(text_content)
+
+	assert parser.tags.len == 1
+	assert parser.tags.first().text() == text_content
+}
+
+fn test_parse_one_tag_with_text() {
+	mut parser := Parser{}
+	text_content := 'tag\nwith\ntext'
+	p_tag := '<p>${text_content}</p>'
+
+	parser.parse_html(p_tag)
+
+	assert parser.tags.first().text() == text_content
+}
+
 fn test_split_parse() {
 	mut parser := Parser{}
 	parser.init()
@@ -37,5 +65,5 @@ fn test_script_tag() {
 	script_content := "\nvar googletag = googletag || {};\ngoogletag.cmd = googletag.cmd || [];if(3 > 5) {console.log('Birl');}\n"
 	temp_html := '<html><body><script>${script_content}</script></body></html>'
 	parser.parse_html(temp_html)
-	assert parser.tags[2].content.len == script_content.replace('\n', '').len
+	assert parser.tags[2].content.len == script_content.len
 }
diff --git a/vlib/net/html/tag.v b/vlib/net/html/tag.v
@@ -40,7 +40,7 @@ pub fn (tag Tag) text() string {
 		return '\n'
 	}
 	mut text_str := strings.new_builder(200)
-	text_str.write_string(tag.content.replace('\n', ''))
+	text_str.write_string(tag.content)
 	for child in tag.children {
 		text_str.write_string(child.text())
 	}

Original file line number	Diff line number	Diff line change
`@@ -40,7 +40,7 @@ pub fn (tag Tag) text() string {`
`40`	`40`	`return '\n'`
`41`	`41`	`}`
`42`	`42`	`mut text_str := strings.new_builder(200)`
`43`		`- text_str.write_string(tag.content.replace('\n', ''))`
	`43`	`+ text_str.write_string(tag.content)`
`44`	`44`	`for child in tag.children {`
`45`	`45`	`text_str.write_string(child.text())`
`46`	`46`	`}`