quick_xml/reader/
state.rs

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
#[cfg(feature = "encoding")]
use encoding_rs::UTF_8;

use crate::encoding::Decoder;
use crate::errors::{Error, Result};
use crate::events::{BytesCData, BytesDecl, BytesEnd, BytesStart, BytesText, Event};
#[cfg(feature = "encoding")]
use crate::reader::EncodingRef;
use crate::reader::{is_whitespace, BangType, ParseState};

use memchr;

/// A struct that holds a current reader state and a parser configuration.
/// It is independent on a way of reading data: the reader feed data into it and
/// get back produced [`Event`]s.
#[derive(Clone)]
pub(super) struct ReaderState {
    /// Number of bytes read from the source of data since the reader was created
    pub offset: usize,
    /// Defines how to process next byte
    pub state: ParseState,
    /// Expand empty element into an opening and closing element
    pub expand_empty_elements: bool,
    /// Trims leading whitespace in Text events, skip the element if text is empty
    pub trim_text_start: bool,
    /// Trims trailing whitespace in Text events.
    pub trim_text_end: bool,
    /// Trims trailing whitespaces from markup names in closing tags `</a >`
    pub trim_markup_names_in_closing_tags: bool,
    /// Check if [`Event::End`] nodes match last [`Event::Start`] node
    pub check_end_names: bool,
    /// Check if comments contains `--` (false per default)
    pub check_comments: bool,
    /// All currently Started elements which didn't have a matching
    /// End element yet.
    ///
    /// For an XML
    ///
    /// ```xml
    /// <root><one/><inner attr="value">|<tag></inner></root>
    /// ```
    /// when cursor at the `|` position buffer contains:
    ///
    /// ```text
    /// rootinner
    /// ^   ^
    /// ```
    ///
    /// The `^` symbols shows which positions stored in the [`Self::opened_starts`]
    /// (0 and 4 in that case).
    opened_buffer: Vec<u8>,
    /// Opened name start indexes into [`Self::opened_buffer`]. See documentation
    /// for that field for details
    opened_starts: Vec<usize>,

    #[cfg(feature = "encoding")]
    /// Reference to the encoding used to read an XML
    pub encoding: EncodingRef,
}

impl ReaderState {
    /// Trims whitespaces from `bytes`, if required, and returns a [`Text`] event.
    ///
    /// # Parameters
    /// - `bytes`: data from the start of stream to the first `<` or from `>` to `<`
    ///
    /// [`Text`]: Event::Text
    pub fn emit_text<'b>(&mut self, bytes: &'b [u8]) -> Result<Event<'b>> {
        let mut content = bytes;

        if self.trim_text_end {
            // Skip the ending '<'
            let len = bytes
                .iter()
                .rposition(|&b| !is_whitespace(b))
                .map_or_else(|| bytes.len(), |p| p + 1);
            content = &bytes[..len];
        }

        Ok(Event::Text(BytesText::wrap(content, self.decoder())))
    }

    /// reads `BytesElement` starting with a `!`,
    /// return `Comment`, `CData` or `DocType` event
    pub fn emit_bang<'b>(&mut self, bang_type: BangType, buf: &'b [u8]) -> Result<Event<'b>> {
        let uncased_starts_with = |string: &[u8], prefix: &[u8]| {
            string.len() >= prefix.len() && string[..prefix.len()].eq_ignore_ascii_case(prefix)
        };

        let len = buf.len();
        match bang_type {
            BangType::Comment if buf.starts_with(b"!--") => {
                debug_assert!(buf.ends_with(b"--"));
                if self.check_comments {
                    // search if '--' not in comments
                    if let Some(p) = memchr::memchr_iter(b'-', &buf[3..len - 2])
                        .position(|p| buf[3 + p + 1] == b'-')
                    {
                        self.offset += len - p;
                        return Err(Error::UnexpectedToken("--".to_string()));
                    }
                }
                Ok(Event::Comment(BytesText::wrap(
                    &buf[3..len - 2],
                    self.decoder(),
                )))
            }
            BangType::CData if uncased_starts_with(buf, b"![CDATA[") => {
                debug_assert!(buf.ends_with(b"]]"));
                Ok(Event::CData(BytesCData::wrap(
                    &buf[8..len - 2],
                    self.decoder(),
                )))
            }
            BangType::DocType if uncased_starts_with(buf, b"!DOCTYPE") => {
                let start = buf[8..]
                    .iter()
                    .position(|b| !is_whitespace(*b))
                    .unwrap_or(len - 8);
                if start + 8 >= len {
                    return Err(Error::EmptyDocType);
                }
                Ok(Event::DocType(BytesText::wrap(
                    &buf[8 + start..],
                    self.decoder(),
                )))
            }
            _ => Err(bang_type.to_err()),
        }
    }

    /// Wraps content of `buf` into the [`Event::End`] event. Does the check that
    /// end name matches the last opened start name if `self.check_end_names` is set.
    pub fn emit_end<'b>(&mut self, buf: &'b [u8]) -> Result<Event<'b>> {
        // Strip the `/` character. `content` contains data between `</` and `>`
        let content = &buf[1..];
        // XML standard permits whitespaces after the markup name in closing tags.
        // Let's strip them from the buffer before comparing tag names.
        let name = if self.trim_markup_names_in_closing_tags {
            if let Some(pos_end_name) = content.iter().rposition(|&b| !is_whitespace(b)) {
                &content[..pos_end_name + 1]
            } else {
                content
            }
        } else {
            content
        };

        let decoder = self.decoder();
        let mismatch_err = |expected: String, found: &[u8], offset: &mut usize| {
            *offset -= buf.len();
            Err(Error::EndEventMismatch {
                expected,
                found: decoder.decode(found).unwrap_or_default().into_owned(),
            })
        };

        // Get the index in self.opened_buffer of the name of the last opened tag
        match self.opened_starts.pop() {
            Some(start) => {
                if self.check_end_names {
                    let expected = &self.opened_buffer[start..];
                    if name != expected {
                        let expected = decoder.decode(expected).unwrap_or_default().into_owned();
                        // #513: In order to allow error recovery we should drop content of the buffer
                        self.opened_buffer.truncate(start);

                        return mismatch_err(expected, name, &mut self.offset);
                    }
                }

                self.opened_buffer.truncate(start);
            }
            None => {
                if self.check_end_names {
                    return mismatch_err("".to_string(), &buf[1..], &mut self.offset);
                }
            }
        }

        Ok(Event::End(BytesEnd::wrap(name.into())))
    }

    /// reads `BytesElement` starting with a `?`,
    /// return `Decl` or `PI` event
    pub fn emit_question_mark<'b>(&mut self, buf: &'b [u8]) -> Result<Event<'b>> {
        let len = buf.len();
        if len > 2 && buf[len - 1] == b'?' {
            if len > 5 && &buf[1..4] == b"xml" && is_whitespace(buf[4]) {
                let event = BytesDecl::from_start(BytesStart::wrap(&buf[1..len - 1], 3));

                // Try getting encoding from the declaration event
                #[cfg(feature = "encoding")]
                if self.encoding.can_be_refined() {
                    if let Some(encoding) = event.encoder() {
                        self.encoding = EncodingRef::XmlDetected(encoding);
                    }
                }

                Ok(Event::Decl(event))
            } else {
                Ok(Event::PI(BytesText::wrap(&buf[1..len - 1], self.decoder())))
            }
        } else {
            self.offset -= len;
            Err(Error::UnexpectedEof("XmlDecl".to_string()))
        }
    }

    /// Converts content of a tag to a `Start` or an `Empty` event
    ///
    /// # Parameters
    /// - `content`: Content of a tag between `<` and `>`
    pub fn emit_start<'b>(&mut self, content: &'b [u8]) -> Result<Event<'b>> {
        let len = content.len();
        let name_end = content
            .iter()
            .position(|&b| is_whitespace(b))
            .unwrap_or(len);
        if let Some(&b'/') = content.last() {
            // This is self-closed tag `<something/>`
            let name_len = if name_end < len { name_end } else { len - 1 };
            let event = BytesStart::wrap(&content[..len - 1], name_len);

            if self.expand_empty_elements {
                self.state = ParseState::Empty;
                self.opened_starts.push(self.opened_buffer.len());
                self.opened_buffer.extend(&content[..name_len]);
                Ok(Event::Start(event))
            } else {
                Ok(Event::Empty(event))
            }
        } else {
            // #514: Always store names event when .check_end_names == false,
            // because checks can be temporary disabled and when they would be
            // enabled, we should have that information
            self.opened_starts.push(self.opened_buffer.len());
            self.opened_buffer.extend(&content[..name_end]);
            Ok(Event::Start(BytesStart::wrap(content, name_end)))
        }
    }

    #[inline]
    pub fn close_expanded_empty(&mut self) -> Result<Event<'static>> {
        self.state = ParseState::ClosedTag;
        let name = self
            .opened_buffer
            .split_off(self.opened_starts.pop().unwrap());
        Ok(Event::End(BytesEnd::wrap(name.into())))
    }

    /// Get the decoder, used to decode bytes, read by this reader, to the strings.
    ///
    /// If [`encoding`] feature is enabled, the used encoding may change after
    /// parsing the XML declaration, otherwise encoding is fixed to UTF-8.
    ///
    /// If [`encoding`] feature is enabled and no encoding is specified in declaration,
    /// defaults to UTF-8.
    ///
    /// [`encoding`]: ../../index.html#encoding
    pub fn decoder(&self) -> Decoder {
        Decoder {
            #[cfg(feature = "encoding")]
            encoding: self.encoding.encoding(),
        }
    }
}

impl Default for ReaderState {
    fn default() -> Self {
        Self {
            offset: 0,
            state: ParseState::Init,
            expand_empty_elements: false,
            trim_text_start: false,
            trim_text_end: false,
            trim_markup_names_in_closing_tags: true,
            check_end_names: true,
            check_comments: false,
            opened_buffer: Vec::new(),
            opened_starts: Vec::new(),

            #[cfg(feature = "encoding")]
            encoding: EncodingRef::Implicit(UTF_8),
        }
    }
}