aboutsummaryrefslogtreecommitdiff
path: root/src/parser.rs
blob: 79e9ec8276fde7bba8e28f46850d0783249121d0 (plain) (blame)
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
use crate::ast::*;

// (λx:T.y): T z
pub fn parse(input: &str) -> Expression {
    match parse_lambda(input) {
        Ok(expr) => return expr,
        Err(e) => println!("invalid expression! {:?}", e)
    }
    return Expression::Constant { term: Term { val: 0, kind: Type::Empty } };
}

/// Parses a Nim-like language into an AST.
pub fn parse_file(path: &str) -> Vec<Expression> {
    match std::fs::read_to_string(path) {
        Ok(file) => match lex(&file) {
            Ok(input) => match parse_lang(&input) {
                Ok(expr) => return expr,
                Err(e) => println!("failed to parse file! {:?}", e),
            },
            Err(e) => println!("failed to lex file! {:?}", e),
        },
        Err(e) => println!("failed to read file! {:?}", e),
    }
    return Vec::new();
}

/// Parses a lambda-calculus-like language into an AST.
pub fn parse_lambda(input: &str) -> Result<Expression, peg::error::ParseError<peg::str::LineCol>> {
    // this is kinda awful, i miss my simple nim pegs
    peg::parser!{
        grammar lambda() for str {
            rule identifier() -> String
            = i:['a'..='z' | 'A'..='Z' | '0'..='9']+ {
                i.iter().collect::<String>()
            }
            rule constant() -> Expression
            = p:"-"? c:['0'..='9']+ {
                let value = c.iter().collect::<String>().parse::<Value>().unwrap();
                Expression::Constant {
                    term: Term {
                        val: if let Some(_) = p {
                            value.wrapping_neg()
                        } else {
                            value
                        },
                        kind: Type::Empty
                    }
                }
            }
            // fucking awful but i don't know another way
            // k:("empty" / "unit" / etc) returns ()
            // and i can't seem to match and raise a parse error
            // so ¯\_(ツ)_/¯
            rule empty() -> Type = k:"empty" {Type::Empty}
            rule unit() -> Type = k:"unit" {Type::Unit}
            rule boolean() -> Type = k:"bool" {Type::Boolean}
            rule natural() -> Type = k:"nat" {Type::Natural}
            rule integer() -> Type = k:"int" {Type::Integer}
            rule kind() -> Type
             = k:(empty() / unit() / boolean() / natural() / integer()) {
                k
            }
            rule annotation() -> Expression
            = e:(conditional() / abstraction() / application() / constant() / variable()) " "* ":" " "* k:kind() {
                Expression::Annotation {
                    expr: Box::new(e),
                    kind: k
                }
            }
            rule variable() -> Expression
            = v:identifier() {
                Expression::Variable {
                    id: v
                }
            }
            rule abstraction() -> Expression
            = ("λ" / "lambda ") " "* p:identifier() " "* "." " "* f:expression() {
                Expression::Abstraction {
                    param: p,
                    func: Box::new(f)
                }
            }
            // fixme: more cases should parse, but how?
            rule application() -> Expression
            = "(" f:(annotation() / abstraction()) ")" " "* a:expression() {
                Expression::Application {
                    func: Box::new(f),
                    arg: Box::new(a)
                }
            }
            rule conditional() -> Expression
            = "if" " "+ c:expression() " "+ "then" " "+ t:expression() " "+ "else" " "+ e:expression() {
                Expression::Conditional {
                    if_cond: Box::new(c),
                    if_then: Box::new(t),
                    if_else: Box::new(e)
                }
            }
            pub rule expression() -> Expression
            = e:(conditional() / annotation() / abstraction() / application() / constant() / variable()) {
                e
            }
            pub rule ast() -> Vec<Expression>
            = expression() ** ("\n"+)
        }
    }
    return lambda::expression(input.trim());
}

/// Converts a whitespace-indented language into a regular bracketed language for matching with PEGs
/// Then, tokens are known to be separated by [\n ]+ (except strings. problem for later.)
pub fn lex(input: &str) -> Result<String, &'static str> {
    #[derive(Eq, PartialEq)]
    enum Previous {
        Start,
        Block,
        Line,
    }
    struct State {
        blank: bool,    // is the line entirely whitespace so far?
        level: usize,   // current indentation level
        count: usize,   // current whitespace count
        previous: Previous,
    }
    let indent_size: usize = 2;

    let mut state = State { blank: true, level: 0, count: 0, previous: Previous::Start };
    let mut buffer = String::new();
    let mut result = String::new();

    // wow lexers are hard
    for c in input.chars() {
        match c {
            '\n' => {
                if !buffer.is_empty() {
                    if state.count == state.level {
                        if state.previous != Previous::Start {
                            result.push(';');
                            result.push('\n');
                        }
                        state.previous = Previous::Line;
                    } else if state.level + indent_size == state.count {
                        result.push(' ');
                        result.push('{');
                        result.push('\n');
                        state.level = state.count;
                        state.previous = Previous::Line;
                    } else if state.count > state.level + indent_size {
                        return Err("invalid jump in indentation");
                    } else if state.count % indent_size != 0 {
                        return Err("incorrect indentation offset, must be a multiple of indent_size");
                    } else if state.level > state.count {
                        while state.level > state.count {
                            if state.previous == Previous::Line {
                                result.push(';');
                            }
                            state.level -= indent_size;
                            result.push('\n');
                            result.push_str(&" ".repeat(state.level));
                            result.push('}');
                            state.previous = Previous::Block;
                        }
                        result.push('\n');
                    } else {
                        return Err("unknown indentation error");
                    }

                    result.push_str(&" ".repeat(state.count));
                    result.push_str(&buffer);

                    state.count = 0;
                    buffer.clear();
                }
                state.blank = true;
            },
            ' ' if state.blank => {
                state.count += 1;
            },
            _ => {
                if state.blank {
                    state.blank = false;
                }
                buffer.push(c);
            },
        }
    }
    if state.previous == Previous::Line {
        result.push(';');
    }
    while state.level != 0 {
        state.level -= 2;
        result.push('\n');
        result.push_str(&" ".repeat(state.level));
        result.push('}');
    }
    return Ok(result);
}

/// Parses a simple language with bracket-based indentation and end-of-term semicolons.
/// The lex() function can turn an indentation-based language into a language recognizable by this.
#[allow(unused_variables)]
pub fn parse_lang(input: &str) -> Result<Vec<Expression>, peg::error::ParseError<peg::str::LineCol>> {
    todo!();
}