domq 0.5.0

Tiny, zero-dependency, forgiving HTML parser: turn messy real-world HTML into a clean DOM tree (and JSON). WASM-first, no_std-friendly.
Documentation 
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

//! Numeric character references: WHATWG Windows-1252 remapping and the
//! U+FFFD replacement rules.

use domq::parse;

fn decode(html: &str) -> String {
    parse(html)
        .root()
        .map(|n| n.text_content())
        .unwrap_or_default()
}

#[test]
fn c1_range_is_remapped_through_windows_1252() {
    assert_eq!(decode("<p>&#150;</p>"), "\u{2013}"); // en dash
    assert_eq!(decode("<p>&#151;</p>"), "\u{2014}"); // em dash
    assert_eq!(decode("<p>&#128;</p>"), "\u{20ac}"); // euro
    assert_eq!(decode("<p>&#x99;</p>"), "\u{2122}"); // trade mark
    assert_eq!(decode("<p>&#X92;s</p>"), "\u{2019}s"); // right single quote
    assert_eq!(
        decode("<p>&#145;&#146; &#147;&#148;</p>"),
        "\u{2018}\u{2019} \u{201c}\u{201d}"
    );
    assert_eq!(
        decode("<p>&#133;&#149;&#159;</p>"),
        "\u{2026}\u{2022}\u{178}"
    );
}

#[test]
fn unmapped_c1_slots_decode_to_themselves() {
    for code in [0x81u32, 0x8D, 0x8F, 0x90, 0x9D] {
        let html = format!("<p>&#{code};</p>");
        assert_eq!(decode(&html), char::from_u32(code).unwrap().to_string());
    }
}

#[test]
fn neighbours_of_the_c1_range_are_untouched() {
    assert_eq!(decode("<p>&#127;</p>"), "\u{7f}");
    assert_eq!(decode("<p>&#160;</p>"), "\u{a0}");
    assert_eq!(decode("<p>&#xA9;</p>"), "\u{a9}");
}

#[test]
fn invalid_code_points_become_replacement_characters() {
    assert_eq!(decode("<p>&#0;</p>"), "\u{fffd}");
    assert_eq!(decode("<p>&#x0;</p>"), "\u{fffd}");
    assert_eq!(decode("<p>&#xD800;</p>"), "\u{fffd}"); // surrogate
    assert_eq!(decode("<p>&#xDFFF;</p>"), "\u{fffd}");
    assert_eq!(decode("<p>&#x110000;</p>"), "\u{fffd}"); // above U+10FFFF
    assert_eq!(decode("<p>&#1114112;</p>"), "\u{fffd}");
    // Overflowing u32 still decodes (to U+FFFD) rather than staying verbatim.
    assert_eq!(decode("<p>&#99999999999999;</p>"), "\u{fffd}");
    assert_eq!(decode("<p>&#xFFFFFFFFFFFF;</p>"), "\u{fffd}");
    // The largest valid scalar value is fine.
    assert_eq!(decode("<p>&#x10FFFF;</p>"), "\u{10ffff}");
}

#[test]
fn malformed_numeric_references_stay_verbatim() {
    assert_eq!(decode("<p>&#;</p>"), "&#;");
    assert_eq!(decode("<p>&#x;</p>"), "&#x;");
    assert_eq!(decode("<p>&#xZZ;</p>"), "&#xZZ;");
    assert_eq!(decode("<p>&#+65;</p>"), "&#+65;");
    assert_eq!(decode("<p>&#65;</p>"), "A");
}

#[test]
fn remapping_applies_in_attributes_too() {
    let dom = parse("<a title='a&#150;b'>x</a>");
    assert_eq!(dom.root().unwrap().attr("title"), Some("a\u{2013}b"));
}

#[test]
fn entity_lookahead_is_bounded_with_multibyte_text() {
    // Byte 32 after each `&` falls inside a multi-byte character. The window
    // used to fall back to the whole rest of the input, so every `&` scanned
    // to EOF looking for a `;` (quadratic).
    let unit = format!("&{}é", "a".repeat(31));
    let html = format!("<p>{}</p>", unit.repeat(80_000));
    let start = std::time::Instant::now();
    let dom = parse(&html);
    assert!(
        start.elapsed().as_millis() < 1000,
        "took {:?}",
        start.elapsed()
    );
    assert_eq!(dom.root().unwrap().text_content(), unit.repeat(80_000));
}

#[test]
fn far_semicolon_is_not_part_of_a_reference() {
    // A `;` beyond the 32-byte lookahead never turns the run into a
    // reference: `&amp` decodes as a legacy name without its `;`.
    assert_eq!(
        decode("<p>&ampéééééééééééééééééé;</p>"),
        "&éééééééééééééééééé;"
    );
    assert_eq!(
        decode("<p>&bogusééééééééééééééé;</p>"),
        "&bogusééééééééééééééé;"
    );
    assert_eq!(decode("<p>&amp;é</p>"), "&é");
}

#[test]
fn legacy_names_decode_without_their_semicolon() {
    // As in browsers, with or without `entities-full`; reported as a missing
    // `;`. Only the 106 legacy names: `&hellip` stays.
    assert_eq!(
        decode("<p>&amp &copy 2026 &lt &hellip</p>"),
        "& \u{a9} 2026 < &hellip"
    );
    let dom = parse("<p>AT&ampT</p>");
    assert_eq!(dom.root().unwrap().text_content(), "AT&T");
    assert!(
        dom.errors()
            .iter()
            .any(|e| e.kind == domq::ErrorKind::MissingSemicolon)
    );
    assert_eq!(decode("<p>&amp;&copy;&lt;</p>"), "&©<");
    // In attribute values a legacy name before `=` or an alphanumeric stays.
    let dom = parse("<a href='?a=1&copy=2&lang=en' title='&copy'>x</a>");
    let a = dom.root().unwrap();
    assert_eq!(a.attr("href"), Some("?a=1&copy=2&lang=en"));
    assert_eq!(a.attr("title"), Some("\u{a9}"));
}

#[test]
fn numeric_references_decode_without_their_semicolon() {
    assert_eq!(decode("<p>&#169 2026 &#x41ZOO</p>"), "\u{a9} 2026 AZOO");
}

#[cfg(not(feature = "entities-full"))]
#[test]
fn curated_table_leaves_rare_names_verbatim() {
    assert_eq!(
        decode("<p>&fjlig; &NotNestedGreaterGreater;</p>"),
        "&fjlig; &NotNestedGreaterGreater;"
    );
}

#[cfg(feature = "entities-full")]
mod full_table {
    use super::*;

    #[test]
    fn multi_code_point_and_long_names() {
        assert_eq!(
            decode("<p>&NotNestedGreaterGreater;</p>"),
            "\u{2aa2}\u{338}"
        );
        assert_eq!(decode("<p>&fjlig;</p>"), "fj");
        assert_eq!(decode("<p>&ThickSpace;</p>"), "\u{205f}\u{200a}");
        // The longest name in the table (31 bytes) still fits the lookahead.
        assert_eq!(
            decode("<p>&CounterClockwiseContourIntegral;</p>"),
            "\u{2233}"
        );
        assert_eq!(decode("<p>&bsol;&quot;&Tab;&NewLine;</p>"), "\\\"\t\n");
    }

    #[test]
    fn names_are_case_sensitive() {
        assert_eq!(decode("<p>&AMP; &amp; &Amp;</p>"), "& & &Amp;");
        assert_eq!(decode("<p>&Aacute;&aacute;</p>"), "\u{c1}\u{e1}");
    }

    #[test]
    fn unknown_names_stay_verbatim_and_are_reported() {
        let dom = parse("<p>&nosuchentity; &Nosuchentity;</p>");
        assert_eq!(
            dom.root().unwrap().text_content(),
            "&nosuchentity; &Nosuchentity;"
        );
        // A known legacy name at the start still decodes, as in browsers.
        assert_eq!(decode("<p>&notanentity;</p>"), "\u{ac}anentity;");
        assert!(
            dom.errors()
                .iter()
                .any(|e| e.kind == domq::ErrorKind::UnknownEntity)
        );
    }

    #[test]
    fn full_names_decode_in_attributes() {
        let dom = parse("<a title='&fjlig;&ShortRightArrow;'>x</a>");
        assert_eq!(dom.root().unwrap().attr("title"), Some("fj\u{2192}"));
    }
}