1 // Copyright 2014-2017 The html5ever Project Developers. See the
2 // COPYRIGHT file at the top-level directory of this distribution.
3 //
4 // Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
5 // http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
6 // <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
7 // option. This file may not be copied, modified, or distributed
8 // except according to those terms.
9
10 extern crate html5ever;
11
12 use std::default::Default;
13 use std::io;
14
15 use html5ever::tendril::*;
16 use html5ever::tokenizer::BufferQueue;
17 use html5ever::tokenizer::{CharacterTokens, EndTag, NullCharacterToken, StartTag, TagToken};
18 use html5ever::tokenizer::{
19 ParseError, Token, TokenSink, TokenSinkResult, Tokenizer, TokenizerOpts,
20 };
21
22 #[derive(Copy, Clone)]
23 struct TokenPrinter {
24 in_char_run: bool,
25 }
26
27 impl TokenPrinter {
is_char(&mut self, is_char: bool)28 fn is_char(&mut self, is_char: bool) {
29 match (self.in_char_run, is_char) {
30 (false, true) => print!("CHAR : \""),
31 (true, false) => println!("\""),
32 _ => (),
33 }
34 self.in_char_run = is_char;
35 }
36
do_char(&mut self, c: char)37 fn do_char(&mut self, c: char) {
38 self.is_char(true);
39 print!("{}", c.escape_default().collect::<String>());
40 }
41 }
42
43 impl TokenSink for TokenPrinter {
44 type Handle = ();
45
process_token(&mut self, token: Token, _line_number: u64) -> TokenSinkResult<()>46 fn process_token(&mut self, token: Token, _line_number: u64) -> TokenSinkResult<()> {
47 match token {
48 CharacterTokens(b) => {
49 for c in b.chars() {
50 self.do_char(c);
51 }
52 },
53 NullCharacterToken => self.do_char('\0'),
54 TagToken(tag) => {
55 self.is_char(false);
56 // This is not proper HTML serialization, of course.
57 match tag.kind {
58 StartTag => print!("TAG : <\x1b[32m{}\x1b[0m", tag.name),
59 EndTag => print!("TAG : <\x1b[31m/{}\x1b[0m", tag.name),
60 }
61 for attr in tag.attrs.iter() {
62 print!(
63 " \x1b[36m{}\x1b[0m='\x1b[34m{}\x1b[0m'",
64 attr.name.local, attr.value
65 );
66 }
67 if tag.self_closing {
68 print!(" \x1b[31m/\x1b[0m");
69 }
70 println!(">");
71 },
72 ParseError(err) => {
73 self.is_char(false);
74 println!("ERROR: {}", err);
75 },
76 _ => {
77 self.is_char(false);
78 println!("OTHER: {:?}", token);
79 },
80 }
81 TokenSinkResult::Continue
82 }
83 }
84
main()85 fn main() {
86 let mut sink = TokenPrinter { in_char_run: false };
87 let mut chunk = ByteTendril::new();
88 io::stdin().read_to_tendril(&mut chunk).unwrap();
89 let mut input = BufferQueue::new();
90 input.push_back(chunk.try_reinterpret().unwrap());
91
92 let mut tok = Tokenizer::new(
93 sink,
94 TokenizerOpts {
95 profile: true,
96 ..Default::default()
97 },
98 );
99 let _ = tok.feed(&mut input);
100 assert!(input.is_empty());
101 tok.end();
102 sink.is_char(false);
103 }
104