blob: 50899d6c7cb6a2b7e522d5f976a71a3ccf49c214 [file] [log] [blame] [edit]
<!DOCTYPE html><html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width, initial-scale=1.0"><meta name="generator" content="rustdoc"><meta name="description" content="Source of the Rust file `lib/ungrammar/src/lexer.rs`."><title>lexer.rs - source</title><script>if(window.location.protocol!=="file:")document.head.insertAdjacentHTML("beforeend","SourceSerif4-Regular-6b053e98.ttf.woff2,FiraSans-Italic-81dc35de.woff2,FiraSans-Regular-0fe48ade.woff2,FiraSans-MediumItalic-ccf7e434.woff2,FiraSans-Medium-e1aa3f0a.woff2,SourceCodePro-Regular-8badfe75.ttf.woff2,SourceCodePro-Semibold-aa29a496.ttf.woff2".split(",").map(f=>`<link rel="preload" as="font" type="font/woff2"href="../../static.files/${f}">`).join(""))</script><link rel="stylesheet" href="../../static.files/normalize-9960930a.css"><link rel="stylesheet" href="../../static.files/rustdoc-77263533.css"><meta name="rustdoc-vars" data-root-path="../../" data-static-root-path="../../static.files/" data-current-crate="ungrammar" data-themes="" data-resource-suffix="" data-rustdoc-version="1.94.0 (4a4ef493e 2026-03-02)" data-channel="1.94.0" data-search-js="search-9e2438ea.js" data-stringdex-js="stringdex-b897f86f.js" data-settings-js="settings-c38705f0.js" ><script src="../../static.files/storage-e2aeef58.js"></script><script defer src="../../static.files/src-script-813739b1.js"></script><script defer src="../../src-files.js"></script><script defer src="../../static.files/main-7bab91a1.js"></script><noscript><link rel="stylesheet" href="../../static.files/noscript-ffcac47a.css"></noscript><link rel="alternate icon" type="image/png" href="../../static.files/favicon-32x32-eab170b8.png"><link rel="icon" type="image/svg+xml" href="../../static.files/favicon-044be391.svg"></head><body class="rustdoc src"><!--[if lte IE 11]><div class="warning">This old browser is unsupported and will most likely display funky things.</div><![endif]--><nav class="sidebar"><div class="src-sidebar-title"><h2>Files</h2></div></nav><div class="sidebar-resizer" title="Drag to resize sidebar"></div><main><section id="main-content" class="content"><div class="main-heading"><h1><div class="sub-heading">ungrammar/</div>lexer.rs</h1><rustdoc-toolbar></rustdoc-toolbar></div><div class="example-wrap digits-3"><pre class="rust"><code><a href=#1 id=1 data-nosnippet>1</a><span class="doccomment">//! Simple hand-written ungrammar lexer
<a href=#2 id=2 data-nosnippet>2</a></span><span class="kw">use </span><span class="kw">crate</span>::error::{<span class="prelude-ty">Result</span>, bail};
<a href=#3 id=3 data-nosnippet>3</a>
<a href=#4 id=4 data-nosnippet>4</a><span class="attr">#[derive(Debug, Eq, PartialEq)]
<a href=#5 id=5 data-nosnippet>5</a></span><span class="kw">pub</span>(<span class="kw">crate</span>) <span class="kw">enum </span>TokenKind {
<a href=#6 id=6 data-nosnippet>6</a> Node(String),
<a href=#7 id=7 data-nosnippet>7</a> Token(String),
<a href=#8 id=8 data-nosnippet>8</a> Eq,
<a href=#9 id=9 data-nosnippet>9</a> Star,
<a href=#10 id=10 data-nosnippet>10</a> Pipe,
<a href=#11 id=11 data-nosnippet>11</a> QMark,
<a href=#12 id=12 data-nosnippet>12</a> Colon,
<a href=#13 id=13 data-nosnippet>13</a> LParen,
<a href=#14 id=14 data-nosnippet>14</a> RParen,
<a href=#15 id=15 data-nosnippet>15</a>}
<a href=#16 id=16 data-nosnippet>16</a>
<a href=#17 id=17 data-nosnippet>17</a><span class="attr">#[derive(Debug)]
<a href=#18 id=18 data-nosnippet>18</a></span><span class="kw">pub</span>(<span class="kw">crate</span>) <span class="kw">struct </span>Token {
<a href=#19 id=19 data-nosnippet>19</a> <span class="kw">pub</span>(<span class="kw">crate</span>) kind: TokenKind,
<a href=#20 id=20 data-nosnippet>20</a> <span class="kw">pub</span>(<span class="kw">crate</span>) loc: Location,
<a href=#21 id=21 data-nosnippet>21</a>}
<a href=#22 id=22 data-nosnippet>22</a>
<a href=#23 id=23 data-nosnippet>23</a><span class="attr">#[derive(Copy, Clone, Default, Debug)]
<a href=#24 id=24 data-nosnippet>24</a></span><span class="kw">pub</span>(<span class="kw">crate</span>) <span class="kw">struct </span>Location {
<a href=#25 id=25 data-nosnippet>25</a> <span class="kw">pub</span>(<span class="kw">crate</span>) line: usize,
<a href=#26 id=26 data-nosnippet>26</a> <span class="kw">pub</span>(<span class="kw">crate</span>) column: usize,
<a href=#27 id=27 data-nosnippet>27</a>}
<a href=#28 id=28 data-nosnippet>28</a>
<a href=#29 id=29 data-nosnippet>29</a><span class="kw">impl </span>Location {
<a href=#30 id=30 data-nosnippet>30</a> <span class="kw">fn </span>advance(<span class="kw-2">&amp;mut </span><span class="self">self</span>, text: <span class="kw-2">&amp;</span>str) {
<a href=#31 id=31 data-nosnippet>31</a> <span class="kw">match </span>text.rfind(<span class="string">'\n'</span>) {
<a href=#32 id=32 data-nosnippet>32</a> <span class="prelude-val">Some</span>(idx) =&gt; {
<a href=#33 id=33 data-nosnippet>33</a> <span class="self">self</span>.line += text.chars().filter(|<span class="kw-2">&amp;</span>it| it == <span class="string">'\n'</span>).count();
<a href=#34 id=34 data-nosnippet>34</a> <span class="self">self</span>.column = text[idx + <span class="number">1</span>..].chars().count();
<a href=#35 id=35 data-nosnippet>35</a> }
<a href=#36 id=36 data-nosnippet>36</a> <span class="prelude-val">None </span>=&gt; <span class="self">self</span>.column += text.chars().count(),
<a href=#37 id=37 data-nosnippet>37</a> }
<a href=#38 id=38 data-nosnippet>38</a> }
<a href=#39 id=39 data-nosnippet>39</a>}
<a href=#40 id=40 data-nosnippet>40</a>
<a href=#41 id=41 data-nosnippet>41</a><span class="kw">pub</span>(<span class="kw">crate</span>) <span class="kw">fn </span>tokenize(<span class="kw-2">mut </span>input: <span class="kw-2">&amp;</span>str) -&gt; <span class="prelude-ty">Result</span>&lt;Vec&lt;Token&gt;&gt; {
<a href=#42 id=42 data-nosnippet>42</a> <span class="kw">let </span><span class="kw-2">mut </span>res = Vec::new();
<a href=#43 id=43 data-nosnippet>43</a> <span class="kw">let </span><span class="kw-2">mut </span>loc = Location::default();
<a href=#44 id=44 data-nosnippet>44</a> <span class="kw">while </span>!input.is_empty() {
<a href=#45 id=45 data-nosnippet>45</a> <span class="kw">let </span>old_input = input;
<a href=#46 id=46 data-nosnippet>46</a> skip_ws(<span class="kw-2">&amp;mut </span>input);
<a href=#47 id=47 data-nosnippet>47</a> skip_comment(<span class="kw-2">&amp;mut </span>input);
<a href=#48 id=48 data-nosnippet>48</a> <span class="kw">if </span>old_input.len() == input.len() {
<a href=#49 id=49 data-nosnippet>49</a> <span class="kw">match </span>advance(<span class="kw-2">&amp;mut </span>input) {
<a href=#50 id=50 data-nosnippet>50</a> <span class="prelude-val">Ok</span>(kind) =&gt; {
<a href=#51 id=51 data-nosnippet>51</a> res.push(Token { kind, loc });
<a href=#52 id=52 data-nosnippet>52</a> }
<a href=#53 id=53 data-nosnippet>53</a> <span class="prelude-val">Err</span>(err) =&gt; <span class="kw">return </span><span class="prelude-val">Err</span>(err.with_location(loc)),
<a href=#54 id=54 data-nosnippet>54</a> }
<a href=#55 id=55 data-nosnippet>55</a> }
<a href=#56 id=56 data-nosnippet>56</a> <span class="kw">let </span>consumed = old_input.len() - input.len();
<a href=#57 id=57 data-nosnippet>57</a> loc.advance(<span class="kw-2">&amp;</span>old_input[..consumed]);
<a href=#58 id=58 data-nosnippet>58</a> }
<a href=#59 id=59 data-nosnippet>59</a>
<a href=#60 id=60 data-nosnippet>60</a> <span class="prelude-val">Ok</span>(res)
<a href=#61 id=61 data-nosnippet>61</a>}
<a href=#62 id=62 data-nosnippet>62</a>
<a href=#63 id=63 data-nosnippet>63</a><span class="kw">fn </span>skip_ws(input: <span class="kw-2">&amp;mut &amp;</span>str) {
<a href=#64 id=64 data-nosnippet>64</a> <span class="kw-2">*</span>input = input.trim_start_matches(is_whitespace)
<a href=#65 id=65 data-nosnippet>65</a>}
<a href=#66 id=66 data-nosnippet>66</a><span class="kw">fn </span>skip_comment(input: <span class="kw-2">&amp;mut &amp;</span>str) {
<a href=#67 id=67 data-nosnippet>67</a> <span class="kw">if </span>input.starts_with(<span class="string">"//"</span>) {
<a href=#68 id=68 data-nosnippet>68</a> <span class="kw">let </span>idx = input.find(<span class="string">'\n'</span>).map_or(input.len(), |it| it + <span class="number">1</span>);
<a href=#69 id=69 data-nosnippet>69</a> <span class="kw-2">*</span>input = <span class="kw-2">&amp;</span>input[idx..]
<a href=#70 id=70 data-nosnippet>70</a> }
<a href=#71 id=71 data-nosnippet>71</a>}
<a href=#72 id=72 data-nosnippet>72</a>
<a href=#73 id=73 data-nosnippet>73</a><span class="kw">fn </span>advance(input: <span class="kw-2">&amp;mut &amp;</span>str) -&gt; <span class="prelude-ty">Result</span>&lt;TokenKind&gt; {
<a href=#74 id=74 data-nosnippet>74</a> <span class="kw">let </span><span class="kw-2">mut </span>chars = input.chars();
<a href=#75 id=75 data-nosnippet>75</a> <span class="kw">let </span>c = chars.next().unwrap();
<a href=#76 id=76 data-nosnippet>76</a> <span class="kw">let </span>res = <span class="kw">match </span>c {
<a href=#77 id=77 data-nosnippet>77</a> <span class="string">'=' </span>=&gt; TokenKind::Eq,
<a href=#78 id=78 data-nosnippet>78</a> <span class="string">'*' </span>=&gt; TokenKind::Star,
<a href=#79 id=79 data-nosnippet>79</a> <span class="string">'?' </span>=&gt; TokenKind::QMark,
<a href=#80 id=80 data-nosnippet>80</a> <span class="string">'(' </span>=&gt; TokenKind::LParen,
<a href=#81 id=81 data-nosnippet>81</a> <span class="string">')' </span>=&gt; TokenKind::RParen,
<a href=#82 id=82 data-nosnippet>82</a> <span class="string">'|' </span>=&gt; TokenKind::Pipe,
<a href=#83 id=83 data-nosnippet>83</a> <span class="string">':' </span>=&gt; TokenKind::Colon,
<a href=#84 id=84 data-nosnippet>84</a> <span class="string">'\'' </span>=&gt; {
<a href=#85 id=85 data-nosnippet>85</a> <span class="kw">let </span><span class="kw-2">mut </span>buf = String::new();
<a href=#86 id=86 data-nosnippet>86</a> <span class="kw">loop </span>{
<a href=#87 id=87 data-nosnippet>87</a> <span class="kw">match </span>chars.next() {
<a href=#88 id=88 data-nosnippet>88</a> <span class="prelude-val">None </span>=&gt; <span class="macro">bail!</span>(<span class="string">"unclosed token literal"</span>),
<a href=#89 id=89 data-nosnippet>89</a> <span class="prelude-val">Some</span>(<span class="string">'\\'</span>) =&gt; <span class="kw">match </span>chars.next() {
<a href=#90 id=90 data-nosnippet>90</a> <span class="prelude-val">Some</span>(c) <span class="kw">if </span>is_escapable(c) =&gt; buf.push(c),
<a href=#91 id=91 data-nosnippet>91</a> <span class="kw">_ </span>=&gt; <span class="macro">bail!</span>(<span class="string">"invalid escape in token literal"</span>),
<a href=#92 id=92 data-nosnippet>92</a> },
<a href=#93 id=93 data-nosnippet>93</a> <span class="prelude-val">Some</span>(<span class="string">'\''</span>) =&gt; <span class="kw">break</span>,
<a href=#94 id=94 data-nosnippet>94</a> <span class="prelude-val">Some</span>(c) =&gt; buf.push(c),
<a href=#95 id=95 data-nosnippet>95</a> }
<a href=#96 id=96 data-nosnippet>96</a> }
<a href=#97 id=97 data-nosnippet>97</a> TokenKind::Token(buf)
<a href=#98 id=98 data-nosnippet>98</a> }
<a href=#99 id=99 data-nosnippet>99</a> c <span class="kw">if </span>is_ident_char(c) =&gt; {
<a href=#100 id=100 data-nosnippet>100</a> <span class="kw">let </span><span class="kw-2">mut </span>buf = String::new();
<a href=#101 id=101 data-nosnippet>101</a> buf.push(c);
<a href=#102 id=102 data-nosnippet>102</a> <span class="kw">loop </span>{
<a href=#103 id=103 data-nosnippet>103</a> <span class="kw">match </span>chars.clone().next() {
<a href=#104 id=104 data-nosnippet>104</a> <span class="prelude-val">Some</span>(c) <span class="kw">if </span>is_ident_char(c) =&gt; {
<a href=#105 id=105 data-nosnippet>105</a> chars.next();
<a href=#106 id=106 data-nosnippet>106</a> buf.push(c);
<a href=#107 id=107 data-nosnippet>107</a> }
<a href=#108 id=108 data-nosnippet>108</a> <span class="kw">_ </span>=&gt; <span class="kw">break</span>,
<a href=#109 id=109 data-nosnippet>109</a> }
<a href=#110 id=110 data-nosnippet>110</a> }
<a href=#111 id=111 data-nosnippet>111</a> TokenKind::Node(buf)
<a href=#112 id=112 data-nosnippet>112</a> }
<a href=#113 id=113 data-nosnippet>113</a> <span class="string">'\r' </span>=&gt; <span class="macro">bail!</span>(<span class="string">"unexpected `\\r`, only Unix-style line endings allowed"</span>),
<a href=#114 id=114 data-nosnippet>114</a> c =&gt; <span class="macro">bail!</span>(<span class="string">"unexpected character: `{}`"</span>, c),
<a href=#115 id=115 data-nosnippet>115</a> };
<a href=#116 id=116 data-nosnippet>116</a>
<a href=#117 id=117 data-nosnippet>117</a> <span class="kw-2">*</span>input = chars.as_str();
<a href=#118 id=118 data-nosnippet>118</a> <span class="prelude-val">Ok</span>(res)
<a href=#119 id=119 data-nosnippet>119</a>}
<a href=#120 id=120 data-nosnippet>120</a>
<a href=#121 id=121 data-nosnippet>121</a><span class="kw">fn </span>is_escapable(c: char) -&gt; bool {
<a href=#122 id=122 data-nosnippet>122</a> <span class="macro">matches!</span>(c, <span class="string">'\\' </span>| <span class="string">'\''</span>)
<a href=#123 id=123 data-nosnippet>123</a>}
<a href=#124 id=124 data-nosnippet>124</a><span class="kw">fn </span>is_whitespace(c: char) -&gt; bool {
<a href=#125 id=125 data-nosnippet>125</a> <span class="macro">matches!</span>(c, <span class="string">' ' </span>| <span class="string">'\t' </span>| <span class="string">'\n'</span>)
<a href=#126 id=126 data-nosnippet>126</a>}
<a href=#127 id=127 data-nosnippet>127</a><span class="kw">fn </span>is_ident_char(c: char) -&gt; bool {
<a href=#128 id=128 data-nosnippet>128</a> <span class="macro">matches!</span>(c, <span class="string">'a'</span>..=<span class="string">'z' </span>| <span class="string">'A'</span>..=<span class="string">'Z' </span>| <span class="string">'_'</span>)
<a href=#129 id=129 data-nosnippet>129</a>}
</code></pre></div></section></main></body></html>