diff options
Diffstat (limited to 'crates/ra_parser/src/lib.rs')
-rw-r--r-- | crates/ra_parser/src/lib.rs | 88 |
1 files changed, 88 insertions, 0 deletions
diff --git a/crates/ra_parser/src/lib.rs b/crates/ra_parser/src/lib.rs new file mode 100644 index 000000000..ddc08e462 --- /dev/null +++ b/crates/ra_parser/src/lib.rs | |||
@@ -0,0 +1,88 @@ | |||
1 | //! The Rust parser. | ||
2 | //! | ||
3 | //! The parser doesn't know about concrete representation of tokens and syntax | ||
4 | //! trees. Abstract `TokenSource` and `TreeSink` traits are used instead. As a | ||
5 | //! consequence, this crates does not contain a lexer. | ||
6 | //! | ||
7 | //! The `Parser` struct from the `parser` module is a cursor into the sequence | ||
8 | //! of tokens. Parsing routines use `Parser` to inspect current state and | ||
9 | //! advance the parsing. | ||
10 | //! | ||
11 | //! The actual parsing happens in the `grammar` module. | ||
12 | //! | ||
13 | //! Tests for this crate live in `ra_syntax` crate. | ||
14 | |||
15 | #[macro_use] | ||
16 | mod token_set; | ||
17 | mod syntax_kind; | ||
18 | mod event; | ||
19 | mod parser; | ||
20 | mod grammar; | ||
21 | |||
22 | pub(crate) use token_set::TokenSet; | ||
23 | |||
24 | pub use syntax_kind::SyntaxKind; | ||
25 | |||
26 | #[derive(Debug, Clone, PartialEq, Eq, Hash)] | ||
27 | pub struct ParseError(pub String); | ||
28 | |||
29 | /// `TokenSource` abstracts the source of the tokens parser operates one. | ||
30 | /// | ||
31 | /// Hopefully this will allow us to treat text and token trees in the same way! | ||
32 | pub trait TokenSource { | ||
33 | /// What is the current token? | ||
34 | fn token_kind(&self, pos: usize) -> SyntaxKind; | ||
35 | /// Is the current token joined to the next one (`> >` vs `>>`). | ||
36 | fn is_token_joint_to_next(&self, pos: usize) -> bool; | ||
37 | /// Is the current token a specified keyword? | ||
38 | fn is_keyword(&self, pos: usize, kw: &str) -> bool; | ||
39 | } | ||
40 | |||
41 | /// `TreeSink` abstracts details of a particular syntax tree implementation. | ||
42 | pub trait TreeSink { | ||
43 | /// Adds new leaf to the current branch. | ||
44 | fn leaf(&mut self, kind: SyntaxKind, n_tokens: u8); | ||
45 | |||
46 | /// Start new branch and make it current. | ||
47 | fn start_branch(&mut self, kind: SyntaxKind); | ||
48 | |||
49 | /// Finish current branch and restore previous | ||
50 | /// branch as current. | ||
51 | fn finish_branch(&mut self); | ||
52 | |||
53 | fn error(&mut self, error: ParseError); | ||
54 | } | ||
55 | |||
56 | /// Parse given tokens into the given sink as a rust file. | ||
57 | pub fn parse(token_source: &dyn TokenSource, tree_sink: &mut dyn TreeSink) { | ||
58 | let mut p = parser::Parser::new(token_source); | ||
59 | grammar::root(&mut p); | ||
60 | let events = p.finish(); | ||
61 | event::process(tree_sink, events); | ||
62 | } | ||
63 | |||
64 | /// A parsing function for a specific braced-block. | ||
65 | pub struct Reparser(fn(&mut parser::Parser)); | ||
66 | |||
67 | impl Reparser { | ||
68 | /// If the node is a braced block, return the corresponding `Reparser`. | ||
69 | pub fn for_node( | ||
70 | node: SyntaxKind, | ||
71 | first_child: Option<SyntaxKind>, | ||
72 | parent: Option<SyntaxKind>, | ||
73 | ) -> Option<Reparser> { | ||
74 | grammar::reparser(node, first_child, parent).map(Reparser) | ||
75 | } | ||
76 | |||
77 | /// Re-parse given tokens using this `Reparser`. | ||
78 | /// | ||
79 | /// Tokens must start with `{`, end with `}` and form a valid brace | ||
80 | /// sequence. | ||
81 | pub fn parse(self, token_source: &dyn TokenSource, tree_sink: &mut dyn TreeSink) { | ||
82 | let Reparser(r) = self; | ||
83 | let mut p = parser::Parser::new(token_source); | ||
84 | r(&mut p); | ||
85 | let events = p.finish(); | ||
86 | event::process(tree_sink, events); | ||
87 | } | ||
88 | } | ||