xml5ever/
driver.rs

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
10use crate::tokenizer::{XmlTokenizer, XmlTokenizerOpts};
11use crate::tree_builder::{TreeSink, XmlTreeBuilder, XmlTreeBuilderOpts};
12
13use std::borrow::Cow;
14
15use crate::tendril;
16use crate::tendril::stream::{TendrilSink, Utf8LossyDecoder};
17use crate::tendril::StrTendril;
18use markup5ever::buffer_queue::BufferQueue;
19
20/// All-encompasing parser setting structure.
21#[derive(Clone, Default)]
22pub struct XmlParseOpts {
23    /// Xml tokenizer options.
24    pub tokenizer: XmlTokenizerOpts,
25    /// Xml tree builder .
26    pub tree_builder: XmlTreeBuilderOpts,
27}
28
29/// Parse and send results to a `TreeSink`.
30///
31/// ## Example
32///
33/// ```ignore
34/// let mut sink = MySink;
35/// parse_document(&mut sink, iter::once(my_str), Default::default());
36/// ```
37pub fn parse_document<Sink>(sink: Sink, opts: XmlParseOpts) -> XmlParser<Sink>
38where
39    Sink: TreeSink,
40{
41    let tb = XmlTreeBuilder::new(sink, opts.tree_builder);
42    let tok = XmlTokenizer::new(tb, opts.tokenizer);
43    XmlParser {
44        tokenizer: tok,
45        input_buffer: BufferQueue::default(),
46    }
47}
48
49/// An XML parser,
50/// ready to receive Unicode input through the `tendril::TendrilSink` trait’s methods.
51pub struct XmlParser<Sink>
52where
53    Sink: TreeSink,
54{
55    /// Tokenizer used by XmlParser.
56    pub tokenizer: XmlTokenizer<XmlTreeBuilder<Sink::Handle, Sink>>,
57    /// Input used by XmlParser.
58    pub input_buffer: BufferQueue,
59}
60
61impl<Sink: TreeSink> TendrilSink<tendril::fmt::UTF8> for XmlParser<Sink> {
62    type Output = Sink::Output;
63
64    fn process(&mut self, t: StrTendril) {
65        self.input_buffer.push_back(t);
66        // FIXME: Properly support </script> somehow.
67        let _ = self.tokenizer.feed(&self.input_buffer);
68    }
69
70    // FIXME: Is it too noisy to report every character decoding error?
71    fn error(&mut self, desc: Cow<'static, str>) {
72        self.tokenizer.sink.sink.parse_error(desc)
73    }
74
75    fn finish(self) -> Self::Output {
76        self.tokenizer.end();
77        self.tokenizer.sink.sink.finish()
78    }
79}
80
81impl<Sink: TreeSink> XmlParser<Sink> {
82    /// Wrap this parser into a `TendrilSink` that accepts UTF-8 bytes.
83    ///
84    /// Use this when your input is bytes that are known to be in the UTF-8 encoding.
85    /// Decoding is lossy, like `String::from_utf8_lossy`.
86    #[allow(clippy::wrong_self_convention)]
87    pub fn from_utf8(self) -> Utf8LossyDecoder<Self> {
88        Utf8LossyDecoder::new(self)
89    }
90}