diff options
author | bors[bot] <bors[bot]@users.noreply.github.com> | 2019-05-27 08:28:13 +0100 |
---|---|---|
committer | bors[bot] <bors[bot]@users.noreply.github.com> | 2019-05-27 08:28:13 +0100 |
commit | ce694ae11854a806031db98c51c068253f927519 (patch) | |
tree | 2996ecd85ff9aa57b6f208e83d42dd03a7370d1e | |
parent | 4f4e50db908ba44f113faeb356ae2b3d0788d308 (diff) | |
parent | 90764fc54b2be1e0fc5d6ac9c9e960d7bb059b14 (diff) |
Merge #1328
1328: Change TokenSource to iteration based r=matklad a=edwin0cheng
This PR change the `TokenSource` trait from random access to be an iteration based trait:
```rust
/// `TokenSource` abstracts the source of the tokens parser operates one.
///
/// Hopefully this will allow us to treat text and token trees in the same way!
pub trait TokenSource {
fn current(&self) -> Token;
/// Lookahead n token
fn lookahead_nth(&self, n: usize) -> Token;
/// bump cursor to next token
fn bump(&mut self);
/// Is the current token a specified keyword?
fn is_keyword(&self, kw: &str) -> bool;
}
/// `TokenCursor` abstracts the cursor of `TokenSource` operates one.
#[derive(Debug, Copy, Clone, Eq, PartialEq)]
pub struct Token {
/// What is the current token?
pub kind: SyntaxKind,
/// Is the current token joined to the next one (`> >` vs `>>`).
pub is_jointed_to_next: bool,
}
```
Note that the refactoring based on this new trait will be separated to incoming PRs
Co-authored-by: Edwin Cheng <[email protected]>
-rw-r--r-- | crates/ra_mbe/src/subtree_parser.rs | 4 | ||||
-rw-r--r-- | crates/ra_mbe/src/subtree_source.rs | 59 | ||||
-rw-r--r-- | crates/ra_mbe/src/syntax_bridge.rs | 63 | ||||
-rw-r--r-- | crates/ra_parser/src/lib.rs | 54 | ||||
-rw-r--r-- | crates/ra_parser/src/parser.rs | 57 | ||||
-rw-r--r-- | crates/ra_syntax/src/parsing.rs | 4 | ||||
-rw-r--r-- | crates/ra_syntax/src/parsing/reparsing.rs | 4 | ||||
-rw-r--r-- | crates/ra_syntax/src/parsing/text_token_source.rs | 47 |
8 files changed, 166 insertions, 126 deletions
diff --git a/crates/ra_mbe/src/subtree_parser.rs b/crates/ra_mbe/src/subtree_parser.rs index 709b87a38..9cc989b23 100644 --- a/crates/ra_mbe/src/subtree_parser.rs +++ b/crates/ra_mbe/src/subtree_parser.rs | |||
@@ -68,13 +68,13 @@ impl<'a> Parser<'a> { | |||
68 | 68 | ||
69 | fn parse<F>(self, f: F) -> Option<tt::TokenTree> | 69 | fn parse<F>(self, f: F) -> Option<tt::TokenTree> |
70 | where | 70 | where |
71 | F: FnOnce(&dyn TokenSource, &mut dyn TreeSink), | 71 | F: FnOnce(&mut dyn TokenSource, &mut dyn TreeSink), |
72 | { | 72 | { |
73 | let buffer = TokenBuffer::new(&self.subtree.token_trees[*self.cur_pos..]); | 73 | let buffer = TokenBuffer::new(&self.subtree.token_trees[*self.cur_pos..]); |
74 | let mut src = SubtreeTokenSource::new(&buffer); | 74 | let mut src = SubtreeTokenSource::new(&buffer); |
75 | let mut sink = OffsetTokenSink { token_pos: 0, error: false }; | 75 | let mut sink = OffsetTokenSink { token_pos: 0, error: false }; |
76 | 76 | ||
77 | f(&src, &mut sink); | 77 | f(&mut src, &mut sink); |
78 | 78 | ||
79 | let r = self.finish(sink.token_pos, &mut src); | 79 | let r = self.finish(sink.token_pos, &mut src); |
80 | if sink.error { | 80 | if sink.error { |
diff --git a/crates/ra_mbe/src/subtree_source.rs b/crates/ra_mbe/src/subtree_source.rs index 972af4a7c..c4f79f38a 100644 --- a/crates/ra_mbe/src/subtree_source.rs +++ b/crates/ra_mbe/src/subtree_source.rs | |||
@@ -1,6 +1,7 @@ | |||
1 | use ra_parser::{TokenSource}; | 1 | use ra_parser::{TokenSource, Token}; |
2 | use ra_syntax::{classify_literal, SmolStr, SyntaxKind, SyntaxKind::*, T}; | 2 | use ra_syntax::{classify_literal, SmolStr, SyntaxKind, SyntaxKind::*, T}; |
3 | use std::cell::{RefCell, Cell}; | 3 | use std::cell::{RefCell, Cell}; |
4 | use std::sync::Arc; | ||
4 | use tt::buffer::{TokenBuffer, Cursor}; | 5 | use tt::buffer::{TokenBuffer, Cursor}; |
5 | 6 | ||
6 | pub(crate) trait Querier { | 7 | pub(crate) trait Querier { |
@@ -65,7 +66,7 @@ impl<'a> SubtreeWalk<'a> { | |||
65 | return cached[pos].clone(); | 66 | return cached[pos].clone(); |
66 | } | 67 | } |
67 | 68 | ||
68 | fn collect_token_trees(&mut self, n: usize) -> Vec<tt::TokenTree> { | 69 | fn collect_token_trees(&self, n: usize) -> Vec<tt::TokenTree> { |
69 | let mut res = vec![]; | 70 | let mut res = vec![]; |
70 | 71 | ||
71 | let mut pos = 0; | 72 | let mut pos = 0; |
@@ -117,43 +118,59 @@ impl<'a> Querier for SubtreeWalk<'a> { | |||
117 | } | 118 | } |
118 | 119 | ||
119 | pub(crate) struct SubtreeTokenSource<'a> { | 120 | pub(crate) struct SubtreeTokenSource<'a> { |
120 | walker: SubtreeWalk<'a>, | 121 | walker: Arc<SubtreeWalk<'a>>, |
122 | curr: (Token, usize), | ||
121 | } | 123 | } |
122 | 124 | ||
123 | impl<'a> SubtreeTokenSource<'a> { | 125 | impl<'a> SubtreeTokenSource<'a> { |
124 | pub fn new(buffer: &'a TokenBuffer) -> SubtreeTokenSource<'a> { | 126 | pub fn new(buffer: &'a TokenBuffer) -> SubtreeTokenSource<'a> { |
125 | SubtreeTokenSource { walker: SubtreeWalk::new(buffer.begin()) } | 127 | let mut res = SubtreeTokenSource { |
128 | walker: Arc::new(SubtreeWalk::new(buffer.begin())), | ||
129 | curr: (Token { kind: EOF, is_jointed_to_next: false }, 0), | ||
130 | }; | ||
131 | res.curr = (res.mk_token(0), 0); | ||
132 | res | ||
126 | } | 133 | } |
127 | 134 | ||
128 | pub fn querier<'b>(&'a self) -> &'b SubtreeWalk<'a> | 135 | pub fn querier(&self) -> Arc<SubtreeWalk<'a>> { |
129 | where | 136 | self.walker.clone() |
130 | 'a: 'b, | ||
131 | { | ||
132 | &self.walker | ||
133 | } | 137 | } |
134 | 138 | ||
135 | pub(crate) fn bump_n(&mut self, parsed_tokens: usize) -> Vec<tt::TokenTree> { | 139 | pub(crate) fn bump_n(&mut self, parsed_tokens: usize) -> Vec<tt::TokenTree> { |
136 | let res = self.walker.collect_token_trees(parsed_tokens); | 140 | let res = self.walker.collect_token_trees(parsed_tokens); |
137 | res | 141 | res |
138 | } | 142 | } |
143 | |||
144 | fn mk_token(&self, pos: usize) -> Token { | ||
145 | match self.walker.get(pos) { | ||
146 | Some(tt) => Token { kind: tt.kind, is_jointed_to_next: tt.is_joint_to_next }, | ||
147 | None => Token { kind: EOF, is_jointed_to_next: false }, | ||
148 | } | ||
149 | } | ||
139 | } | 150 | } |
140 | 151 | ||
141 | impl<'a> TokenSource for SubtreeTokenSource<'a> { | 152 | impl<'a> TokenSource for SubtreeTokenSource<'a> { |
142 | fn token_kind(&self, pos: usize) -> SyntaxKind { | 153 | fn current(&self) -> Token { |
143 | if let Some(tok) = self.walker.get(pos) { | 154 | self.curr.0 |
144 | tok.kind | ||
145 | } else { | ||
146 | SyntaxKind::EOF | ||
147 | } | ||
148 | } | 155 | } |
149 | fn is_token_joint_to_next(&self, pos: usize) -> bool { | 156 | |
150 | match self.walker.get(pos) { | 157 | /// Lookahead n token |
151 | Some(t) => t.is_joint_to_next, | 158 | fn lookahead_nth(&self, n: usize) -> Token { |
152 | _ => false, | 159 | self.mk_token(self.curr.1 + n) |
160 | } | ||
161 | |||
162 | /// bump cursor to next token | ||
163 | fn bump(&mut self) { | ||
164 | if self.current().kind == EOF { | ||
165 | return; | ||
153 | } | 166 | } |
167 | |||
168 | self.curr = (self.mk_token(self.curr.1 + 1), self.curr.1 + 1) | ||
154 | } | 169 | } |
155 | fn is_keyword(&self, pos: usize, kw: &str) -> bool { | 170 | |
156 | match self.walker.get(pos) { | 171 | /// Is the current token a specified keyword? |
172 | fn is_keyword(&self, kw: &str) -> bool { | ||
173 | match self.walker.get(self.curr.1) { | ||
157 | Some(t) => t.text == *kw, | 174 | Some(t) => t.text == *kw, |
158 | _ => false, | 175 | _ => false, |
159 | } | 176 | } |
diff --git a/crates/ra_mbe/src/syntax_bridge.rs b/crates/ra_mbe/src/syntax_bridge.rs index 0a75305b4..0a7e50c4e 100644 --- a/crates/ra_mbe/src/syntax_bridge.rs +++ b/crates/ra_mbe/src/syntax_bridge.rs | |||
@@ -45,17 +45,25 @@ pub fn syntax_node_to_token_tree(node: &SyntaxNode) -> Option<(tt::Subtree, Toke | |||
45 | // | 45 | // |
46 | // | 46 | // |
47 | 47 | ||
48 | /// Parses the token tree (result of macro expansion) to an expression | 48 | fn token_tree_to_syntax_node<F>(tt: &tt::Subtree, f: F) -> Result<TreeArc<SyntaxNode>, ExpandError> |
49 | pub fn token_tree_to_expr(tt: &tt::Subtree) -> Result<TreeArc<ast::Expr>, ExpandError> { | 49 | where |
50 | F: Fn(&mut ra_parser::TokenSource, &mut ra_parser::TreeSink), | ||
51 | { | ||
50 | let buffer = tt::buffer::TokenBuffer::new(&[tt.clone().into()]); | 52 | let buffer = tt::buffer::TokenBuffer::new(&[tt.clone().into()]); |
51 | let token_source = SubtreeTokenSource::new(&buffer); | 53 | let mut token_source = SubtreeTokenSource::new(&buffer); |
52 | let mut tree_sink = TtTreeSink::new(token_source.querier()); | 54 | let querier = token_source.querier(); |
53 | ra_parser::parse_expr(&token_source, &mut tree_sink); | 55 | let mut tree_sink = TtTreeSink::new(querier.as_ref()); |
56 | f(&mut token_source, &mut tree_sink); | ||
54 | if tree_sink.roots.len() != 1 { | 57 | if tree_sink.roots.len() != 1 { |
55 | return Err(ExpandError::ConversionError); | 58 | return Err(ExpandError::ConversionError); |
56 | } | 59 | } |
57 | 60 | ||
58 | let syntax = tree_sink.inner.finish(); | 61 | Ok(tree_sink.inner.finish()) |
62 | } | ||
63 | |||
64 | /// Parses the token tree (result of macro expansion) to an expression | ||
65 | pub fn token_tree_to_expr(tt: &tt::Subtree) -> Result<TreeArc<ast::Expr>, ExpandError> { | ||
66 | let syntax = token_tree_to_syntax_node(tt, ra_parser::parse_expr)?; | ||
59 | ast::Expr::cast(&syntax) | 67 | ast::Expr::cast(&syntax) |
60 | .map(|m| m.to_owned()) | 68 | .map(|m| m.to_owned()) |
61 | .ok_or_else(|| crate::ExpandError::ConversionError) | 69 | .ok_or_else(|| crate::ExpandError::ConversionError) |
@@ -63,28 +71,13 @@ pub fn token_tree_to_expr(tt: &tt::Subtree) -> Result<TreeArc<ast::Expr>, Expand | |||
63 | 71 | ||
64 | /// Parses the token tree (result of macro expansion) to a Pattern | 72 | /// Parses the token tree (result of macro expansion) to a Pattern |
65 | pub fn token_tree_to_pat(tt: &tt::Subtree) -> Result<TreeArc<ast::Pat>, ExpandError> { | 73 | pub fn token_tree_to_pat(tt: &tt::Subtree) -> Result<TreeArc<ast::Pat>, ExpandError> { |
66 | let buffer = tt::buffer::TokenBuffer::new(&[tt.clone().into()]); | 74 | let syntax = token_tree_to_syntax_node(tt, ra_parser::parse_pat)?; |
67 | let token_source = SubtreeTokenSource::new(&buffer); | ||
68 | let mut tree_sink = TtTreeSink::new(token_source.querier()); | ||
69 | ra_parser::parse_pat(&token_source, &mut tree_sink); | ||
70 | if tree_sink.roots.len() != 1 { | ||
71 | return Err(ExpandError::ConversionError); | ||
72 | } | ||
73 | |||
74 | let syntax = tree_sink.inner.finish(); | ||
75 | ast::Pat::cast(&syntax).map(|m| m.to_owned()).ok_or_else(|| ExpandError::ConversionError) | 75 | ast::Pat::cast(&syntax).map(|m| m.to_owned()).ok_or_else(|| ExpandError::ConversionError) |
76 | } | 76 | } |
77 | 77 | ||
78 | /// Parses the token tree (result of macro expansion) to a Type | 78 | /// Parses the token tree (result of macro expansion) to a Type |
79 | pub fn token_tree_to_ty(tt: &tt::Subtree) -> Result<TreeArc<ast::TypeRef>, ExpandError> { | 79 | pub fn token_tree_to_ty(tt: &tt::Subtree) -> Result<TreeArc<ast::TypeRef>, ExpandError> { |
80 | let buffer = tt::buffer::TokenBuffer::new(&[tt.clone().into()]); | 80 | let syntax = token_tree_to_syntax_node(tt, ra_parser::parse_ty)?; |
81 | let token_source = SubtreeTokenSource::new(&buffer); | ||
82 | let mut tree_sink = TtTreeSink::new(token_source.querier()); | ||
83 | ra_parser::parse_ty(&token_source, &mut tree_sink); | ||
84 | if tree_sink.roots.len() != 1 { | ||
85 | return Err(ExpandError::ConversionError); | ||
86 | } | ||
87 | let syntax = tree_sink.inner.finish(); | ||
88 | ast::TypeRef::cast(&syntax).map(|m| m.to_owned()).ok_or_else(|| ExpandError::ConversionError) | 81 | ast::TypeRef::cast(&syntax).map(|m| m.to_owned()).ok_or_else(|| ExpandError::ConversionError) |
89 | } | 82 | } |
90 | 83 | ||
@@ -92,14 +85,7 @@ pub fn token_tree_to_ty(tt: &tt::Subtree) -> Result<TreeArc<ast::TypeRef>, Expan | |||
92 | pub fn token_tree_to_macro_stmts( | 85 | pub fn token_tree_to_macro_stmts( |
93 | tt: &tt::Subtree, | 86 | tt: &tt::Subtree, |
94 | ) -> Result<TreeArc<ast::MacroStmts>, ExpandError> { | 87 | ) -> Result<TreeArc<ast::MacroStmts>, ExpandError> { |
95 | let buffer = tt::buffer::TokenBuffer::new(&[tt.clone().into()]); | 88 | let syntax = token_tree_to_syntax_node(tt, ra_parser::parse_macro_stmts)?; |
96 | let token_source = SubtreeTokenSource::new(&buffer); | ||
97 | let mut tree_sink = TtTreeSink::new(token_source.querier()); | ||
98 | ra_parser::parse_macro_stmts(&token_source, &mut tree_sink); | ||
99 | if tree_sink.roots.len() != 1 { | ||
100 | return Err(ExpandError::ConversionError); | ||
101 | } | ||
102 | let syntax = tree_sink.inner.finish(); | ||
103 | ast::MacroStmts::cast(&syntax).map(|m| m.to_owned()).ok_or_else(|| ExpandError::ConversionError) | 89 | ast::MacroStmts::cast(&syntax).map(|m| m.to_owned()).ok_or_else(|| ExpandError::ConversionError) |
104 | } | 90 | } |
105 | 91 | ||
@@ -107,24 +93,13 @@ pub fn token_tree_to_macro_stmts( | |||
107 | pub fn token_tree_to_macro_items( | 93 | pub fn token_tree_to_macro_items( |
108 | tt: &tt::Subtree, | 94 | tt: &tt::Subtree, |
109 | ) -> Result<TreeArc<ast::MacroItems>, ExpandError> { | 95 | ) -> Result<TreeArc<ast::MacroItems>, ExpandError> { |
110 | let buffer = tt::buffer::TokenBuffer::new(&[tt.clone().into()]); | 96 | let syntax = token_tree_to_syntax_node(tt, ra_parser::parse_macro_items)?; |
111 | let token_source = SubtreeTokenSource::new(&buffer); | ||
112 | let mut tree_sink = TtTreeSink::new(token_source.querier()); | ||
113 | ra_parser::parse_macro_items(&token_source, &mut tree_sink); | ||
114 | if tree_sink.roots.len() != 1 { | ||
115 | return Err(ExpandError::ConversionError); | ||
116 | } | ||
117 | let syntax = tree_sink.inner.finish(); | ||
118 | ast::MacroItems::cast(&syntax).map(|m| m.to_owned()).ok_or_else(|| ExpandError::ConversionError) | 97 | ast::MacroItems::cast(&syntax).map(|m| m.to_owned()).ok_or_else(|| ExpandError::ConversionError) |
119 | } | 98 | } |
120 | 99 | ||
121 | /// Parses the token tree (result of macro expansion) as a sequence of items | 100 | /// Parses the token tree (result of macro expansion) as a sequence of items |
122 | pub fn token_tree_to_ast_item_list(tt: &tt::Subtree) -> TreeArc<ast::SourceFile> { | 101 | pub fn token_tree_to_ast_item_list(tt: &tt::Subtree) -> TreeArc<ast::SourceFile> { |
123 | let buffer = tt::buffer::TokenBuffer::new(&[tt.clone().into()]); | 102 | let syntax = token_tree_to_syntax_node(tt, ra_parser::parse).unwrap(); |
124 | let token_source = SubtreeTokenSource::new(&buffer); | ||
125 | let mut tree_sink = TtTreeSink::new(token_source.querier()); | ||
126 | ra_parser::parse(&token_source, &mut tree_sink); | ||
127 | let syntax = tree_sink.inner.finish(); | ||
128 | ast::SourceFile::cast(&syntax).unwrap().to_owned() | 103 | ast::SourceFile::cast(&syntax).unwrap().to_owned() |
129 | } | 104 | } |
130 | 105 | ||
diff --git a/crates/ra_parser/src/lib.rs b/crates/ra_parser/src/lib.rs index 697d1b794..3d88be642 100644 --- a/crates/ra_parser/src/lib.rs +++ b/crates/ra_parser/src/lib.rs | |||
@@ -31,12 +31,26 @@ pub struct ParseError(pub String); | |||
31 | /// | 31 | /// |
32 | /// Hopefully this will allow us to treat text and token trees in the same way! | 32 | /// Hopefully this will allow us to treat text and token trees in the same way! |
33 | pub trait TokenSource { | 33 | pub trait TokenSource { |
34 | fn current(&self) -> Token; | ||
35 | |||
36 | /// Lookahead n token | ||
37 | fn lookahead_nth(&self, n: usize) -> Token; | ||
38 | |||
39 | /// bump cursor to next token | ||
40 | fn bump(&mut self); | ||
41 | |||
42 | /// Is the current token a specified keyword? | ||
43 | fn is_keyword(&self, kw: &str) -> bool; | ||
44 | } | ||
45 | |||
46 | /// `TokenCursor` abstracts the cursor of `TokenSource` operates one. | ||
47 | #[derive(Debug, Copy, Clone, Eq, PartialEq)] | ||
48 | pub struct Token { | ||
34 | /// What is the current token? | 49 | /// What is the current token? |
35 | fn token_kind(&self, pos: usize) -> SyntaxKind; | 50 | pub kind: SyntaxKind, |
51 | |||
36 | /// Is the current token joined to the next one (`> >` vs `>>`). | 52 | /// Is the current token joined to the next one (`> >` vs `>>`). |
37 | fn is_token_joint_to_next(&self, pos: usize) -> bool; | 53 | pub is_jointed_to_next: bool, |
38 | /// Is the current token a specified keyword? | ||
39 | fn is_keyword(&self, pos: usize, kw: &str) -> bool; | ||
40 | } | 54 | } |
41 | 55 | ||
42 | /// `TreeSink` abstracts details of a particular syntax tree implementation. | 56 | /// `TreeSink` abstracts details of a particular syntax tree implementation. |
@@ -54,7 +68,7 @@ pub trait TreeSink { | |||
54 | fn error(&mut self, error: ParseError); | 68 | fn error(&mut self, error: ParseError); |
55 | } | 69 | } |
56 | 70 | ||
57 | fn parse_from_tokens<F>(token_source: &dyn TokenSource, tree_sink: &mut dyn TreeSink, f: F) | 71 | fn parse_from_tokens<F>(token_source: &mut dyn TokenSource, tree_sink: &mut dyn TreeSink, f: F) |
58 | where | 72 | where |
59 | F: FnOnce(&mut parser::Parser), | 73 | F: FnOnce(&mut parser::Parser), |
60 | { | 74 | { |
@@ -65,61 +79,65 @@ where | |||
65 | } | 79 | } |
66 | 80 | ||
67 | /// Parse given tokens into the given sink as a rust file. | 81 | /// Parse given tokens into the given sink as a rust file. |
68 | pub fn parse(token_source: &dyn TokenSource, tree_sink: &mut dyn TreeSink) { | 82 | pub fn parse(token_source: &mut dyn TokenSource, tree_sink: &mut dyn TreeSink) { |
69 | parse_from_tokens(token_source, tree_sink, grammar::root); | 83 | parse_from_tokens(token_source, tree_sink, grammar::root); |
70 | } | 84 | } |
71 | 85 | ||
72 | /// Parse given tokens into the given sink as a path | 86 | /// Parse given tokens into the given sink as a path |
73 | pub fn parse_path(token_source: &dyn TokenSource, tree_sink: &mut dyn TreeSink) { | 87 | pub fn parse_path(token_source: &mut dyn TokenSource, tree_sink: &mut dyn TreeSink) { |
74 | parse_from_tokens(token_source, tree_sink, grammar::path); | 88 | parse_from_tokens(token_source, tree_sink, grammar::path); |
75 | } | 89 | } |
76 | 90 | ||
77 | /// Parse given tokens into the given sink as a expression | 91 | /// Parse given tokens into the given sink as a expression |
78 | pub fn parse_expr(token_source: &dyn TokenSource, tree_sink: &mut dyn TreeSink) { | 92 | pub fn parse_expr(token_source: &mut dyn TokenSource, tree_sink: &mut dyn TreeSink) { |
79 | parse_from_tokens(token_source, tree_sink, grammar::expr); | 93 | parse_from_tokens(token_source, tree_sink, grammar::expr); |
80 | } | 94 | } |
81 | 95 | ||
82 | /// Parse given tokens into the given sink as a ty | 96 | /// Parse given tokens into the given sink as a ty |
83 | pub fn parse_ty(token_source: &dyn TokenSource, tree_sink: &mut dyn TreeSink) { | 97 | pub fn parse_ty(token_source: &mut dyn TokenSource, tree_sink: &mut dyn TreeSink) { |
84 | parse_from_tokens(token_source, tree_sink, grammar::type_); | 98 | parse_from_tokens(token_source, tree_sink, grammar::type_); |
85 | } | 99 | } |
86 | 100 | ||
87 | /// Parse given tokens into the given sink as a pattern | 101 | /// Parse given tokens into the given sink as a pattern |
88 | pub fn parse_pat(token_source: &dyn TokenSource, tree_sink: &mut dyn TreeSink) { | 102 | pub fn parse_pat(token_source: &mut dyn TokenSource, tree_sink: &mut dyn TreeSink) { |
89 | parse_from_tokens(token_source, tree_sink, grammar::pattern); | 103 | parse_from_tokens(token_source, tree_sink, grammar::pattern); |
90 | } | 104 | } |
91 | 105 | ||
92 | /// Parse given tokens into the given sink as a statement | 106 | /// Parse given tokens into the given sink as a statement |
93 | pub fn parse_stmt(token_source: &dyn TokenSource, tree_sink: &mut dyn TreeSink, with_semi: bool) { | 107 | pub fn parse_stmt( |
108 | token_source: &mut dyn TokenSource, | ||
109 | tree_sink: &mut dyn TreeSink, | ||
110 | with_semi: bool, | ||
111 | ) { | ||
94 | parse_from_tokens(token_source, tree_sink, |p| grammar::stmt(p, with_semi)); | 112 | parse_from_tokens(token_source, tree_sink, |p| grammar::stmt(p, with_semi)); |
95 | } | 113 | } |
96 | 114 | ||
97 | /// Parse given tokens into the given sink as a block | 115 | /// Parse given tokens into the given sink as a block |
98 | pub fn parse_block(token_source: &dyn TokenSource, tree_sink: &mut dyn TreeSink) { | 116 | pub fn parse_block(token_source: &mut dyn TokenSource, tree_sink: &mut dyn TreeSink) { |
99 | parse_from_tokens(token_source, tree_sink, grammar::block); | 117 | parse_from_tokens(token_source, tree_sink, grammar::block); |
100 | } | 118 | } |
101 | 119 | ||
102 | pub fn parse_meta(token_source: &dyn TokenSource, tree_sink: &mut dyn TreeSink) { | 120 | pub fn parse_meta(token_source: &mut dyn TokenSource, tree_sink: &mut dyn TreeSink) { |
103 | parse_from_tokens(token_source, tree_sink, grammar::meta_item); | 121 | parse_from_tokens(token_source, tree_sink, grammar::meta_item); |
104 | } | 122 | } |
105 | 123 | ||
106 | /// Parse given tokens into the given sink as an item | 124 | /// Parse given tokens into the given sink as an item |
107 | pub fn parse_item(token_source: &dyn TokenSource, tree_sink: &mut dyn TreeSink) { | 125 | pub fn parse_item(token_source: &mut dyn TokenSource, tree_sink: &mut dyn TreeSink) { |
108 | parse_from_tokens(token_source, tree_sink, grammar::item); | 126 | parse_from_tokens(token_source, tree_sink, grammar::item); |
109 | } | 127 | } |
110 | 128 | ||
111 | /// Parse given tokens into the given sink as an visibility qualifier | 129 | /// Parse given tokens into the given sink as an visibility qualifier |
112 | pub fn parse_vis(token_source: &dyn TokenSource, tree_sink: &mut dyn TreeSink) { | 130 | pub fn parse_vis(token_source: &mut dyn TokenSource, tree_sink: &mut dyn TreeSink) { |
113 | parse_from_tokens(token_source, tree_sink, |p| { | 131 | parse_from_tokens(token_source, tree_sink, |p| { |
114 | grammar::opt_visibility(p); | 132 | grammar::opt_visibility(p); |
115 | }); | 133 | }); |
116 | } | 134 | } |
117 | 135 | ||
118 | pub fn parse_macro_items(token_source: &dyn TokenSource, tree_sink: &mut dyn TreeSink) { | 136 | pub fn parse_macro_items(token_source: &mut dyn TokenSource, tree_sink: &mut dyn TreeSink) { |
119 | parse_from_tokens(token_source, tree_sink, grammar::macro_items); | 137 | parse_from_tokens(token_source, tree_sink, grammar::macro_items); |
120 | } | 138 | } |
121 | 139 | ||
122 | pub fn parse_macro_stmts(token_source: &dyn TokenSource, tree_sink: &mut dyn TreeSink) { | 140 | pub fn parse_macro_stmts(token_source: &mut dyn TokenSource, tree_sink: &mut dyn TreeSink) { |
123 | parse_from_tokens(token_source, tree_sink, grammar::macro_stmts); | 141 | parse_from_tokens(token_source, tree_sink, grammar::macro_stmts); |
124 | } | 142 | } |
125 | 143 | ||
@@ -140,7 +158,7 @@ impl Reparser { | |||
140 | /// | 158 | /// |
141 | /// Tokens must start with `{`, end with `}` and form a valid brace | 159 | /// Tokens must start with `{`, end with `}` and form a valid brace |
142 | /// sequence. | 160 | /// sequence. |
143 | pub fn parse(self, token_source: &dyn TokenSource, tree_sink: &mut dyn TreeSink) { | 161 | pub fn parse(self, token_source: &mut dyn TokenSource, tree_sink: &mut dyn TreeSink) { |
144 | let Reparser(r) = self; | 162 | let Reparser(r) = self; |
145 | let mut p = parser::Parser::new(token_source); | 163 | let mut p = parser::Parser::new(token_source); |
146 | r(&mut p); | 164 | r(&mut p); |
diff --git a/crates/ra_parser/src/parser.rs b/crates/ra_parser/src/parser.rs index 4434dfb09..8f654f04c 100644 --- a/crates/ra_parser/src/parser.rs +++ b/crates/ra_parser/src/parser.rs | |||
@@ -19,15 +19,14 @@ use crate::{ | |||
19 | /// "start expression, consume number literal, | 19 | /// "start expression, consume number literal, |
20 | /// finish expression". See `Event` docs for more. | 20 | /// finish expression". See `Event` docs for more. |
21 | pub(crate) struct Parser<'t> { | 21 | pub(crate) struct Parser<'t> { |
22 | token_source: &'t dyn TokenSource, | 22 | token_source: &'t mut dyn TokenSource, |
23 | token_pos: usize, | ||
24 | events: Vec<Event>, | 23 | events: Vec<Event>, |
25 | steps: Cell<u32>, | 24 | steps: Cell<u32>, |
26 | } | 25 | } |
27 | 26 | ||
28 | impl<'t> Parser<'t> { | 27 | impl<'t> Parser<'t> { |
29 | pub(super) fn new(token_source: &'t dyn TokenSource) -> Parser<'t> { | 28 | pub(super) fn new(token_source: &'t mut dyn TokenSource) -> Parser<'t> { |
30 | Parser { token_source, token_pos: 0, events: Vec::new(), steps: Cell::new(0) } | 29 | Parser { token_source, events: Vec::new(), steps: Cell::new(0) } |
31 | } | 30 | } |
32 | 31 | ||
33 | pub(crate) fn finish(self) -> Vec<Event> { | 32 | pub(crate) fn finish(self) -> Vec<Event> { |
@@ -49,7 +48,7 @@ impl<'t> Parser<'t> { | |||
49 | let c1 = self.nth(0); | 48 | let c1 = self.nth(0); |
50 | let c2 = self.nth(1); | 49 | let c2 = self.nth(1); |
51 | 50 | ||
52 | if self.token_source.is_token_joint_to_next(self.token_pos) { | 51 | if self.token_source.current().is_jointed_to_next { |
53 | Some((c1, c2)) | 52 | Some((c1, c2)) |
54 | } else { | 53 | } else { |
55 | None | 54 | None |
@@ -64,8 +63,8 @@ impl<'t> Parser<'t> { | |||
64 | let c1 = self.nth(0); | 63 | let c1 = self.nth(0); |
65 | let c2 = self.nth(1); | 64 | let c2 = self.nth(1); |
66 | let c3 = self.nth(2); | 65 | let c3 = self.nth(2); |
67 | if self.token_source.is_token_joint_to_next(self.token_pos) | 66 | if self.token_source.current().is_jointed_to_next |
68 | && self.token_source.is_token_joint_to_next(self.token_pos + 1) | 67 | && self.token_source.lookahead_nth(1).is_jointed_to_next |
69 | { | 68 | { |
70 | Some((c1, c2, c3)) | 69 | Some((c1, c2, c3)) |
71 | } else { | 70 | } else { |
@@ -76,6 +75,8 @@ impl<'t> Parser<'t> { | |||
76 | /// Lookahead operation: returns the kind of the next nth | 75 | /// Lookahead operation: returns the kind of the next nth |
77 | /// token. | 76 | /// token. |
78 | pub(crate) fn nth(&self, n: usize) -> SyntaxKind { | 77 | pub(crate) fn nth(&self, n: usize) -> SyntaxKind { |
78 | assert!(n <= 3); | ||
79 | |||
79 | let steps = self.steps.get(); | 80 | let steps = self.steps.get(); |
80 | assert!(steps <= 10_000_000, "the parser seems stuck"); | 81 | assert!(steps <= 10_000_000, "the parser seems stuck"); |
81 | self.steps.set(steps + 1); | 82 | self.steps.set(steps + 1); |
@@ -86,7 +87,7 @@ impl<'t> Parser<'t> { | |||
86 | let mut i = 0; | 87 | let mut i = 0; |
87 | 88 | ||
88 | loop { | 89 | loop { |
89 | let mut kind = self.token_source.token_kind(self.token_pos + i); | 90 | let mut kind = self.token_source.lookahead_nth(i).kind; |
90 | if let Some((composited, step)) = self.is_composite(kind, i) { | 91 | if let Some((composited, step)) = self.is_composite(kind, i) { |
91 | kind = composited; | 92 | kind = composited; |
92 | i += step; | 93 | i += step; |
@@ -115,7 +116,7 @@ impl<'t> Parser<'t> { | |||
115 | 116 | ||
116 | /// Checks if the current token is contextual keyword with text `t`. | 117 | /// Checks if the current token is contextual keyword with text `t`. |
117 | pub(crate) fn at_contextual_kw(&self, kw: &str) -> bool { | 118 | pub(crate) fn at_contextual_kw(&self, kw: &str) -> bool { |
118 | self.token_source.is_keyword(self.token_pos, kw) | 119 | self.token_source.is_keyword(kw) |
119 | } | 120 | } |
120 | 121 | ||
121 | /// Starts a new node in the syntax tree. All nodes and tokens | 122 | /// Starts a new node in the syntax tree. All nodes and tokens |
@@ -130,12 +131,12 @@ impl<'t> Parser<'t> { | |||
130 | /// Advances the parser by one token unconditionally | 131 | /// Advances the parser by one token unconditionally |
131 | /// Mainly use in `token_tree` parsing | 132 | /// Mainly use in `token_tree` parsing |
132 | pub(crate) fn bump_raw(&mut self) { | 133 | pub(crate) fn bump_raw(&mut self) { |
133 | let mut kind = self.token_source.token_kind(self.token_pos); | 134 | let mut kind = self.token_source.current().kind; |
134 | 135 | ||
135 | // Skip dollars, do_bump will eat these later | 136 | // Skip dollars, do_bump will eat these later |
136 | let mut i = 0; | 137 | let mut i = 0; |
137 | while kind == SyntaxKind::L_DOLLAR || kind == SyntaxKind::R_DOLLAR { | 138 | while kind == SyntaxKind::L_DOLLAR || kind == SyntaxKind::R_DOLLAR { |
138 | kind = self.token_source.token_kind(self.token_pos + i); | 139 | kind = self.token_source.lookahead_nth(i).kind; |
139 | i += 1; | 140 | i += 1; |
140 | } | 141 | } |
141 | 142 | ||
@@ -236,7 +237,11 @@ impl<'t> Parser<'t> { | |||
236 | 237 | ||
237 | fn do_bump(&mut self, kind: SyntaxKind, n_raw_tokens: u8) { | 238 | fn do_bump(&mut self, kind: SyntaxKind, n_raw_tokens: u8) { |
238 | self.eat_dollars(); | 239 | self.eat_dollars(); |
239 | self.token_pos += usize::from(n_raw_tokens); | 240 | |
241 | for _ in 0..n_raw_tokens { | ||
242 | self.token_source.bump(); | ||
243 | } | ||
244 | |||
240 | self.push_event(Event::Token { kind, n_raw_tokens }); | 245 | self.push_event(Event::Token { kind, n_raw_tokens }); |
241 | } | 246 | } |
242 | 247 | ||
@@ -249,10 +254,14 @@ impl<'t> Parser<'t> { | |||
249 | // We assume the dollars will not occuried between | 254 | // We assume the dollars will not occuried between |
250 | // mult-byte tokens | 255 | // mult-byte tokens |
251 | 256 | ||
252 | let jn1 = self.token_source.is_token_joint_to_next(self.token_pos + n); | 257 | let first = self.token_source.lookahead_nth(n); |
253 | let la2 = self.token_source.token_kind(self.token_pos + n + 1); | 258 | let second = self.token_source.lookahead_nth(n + 1); |
254 | let jn2 = self.token_source.is_token_joint_to_next(self.token_pos + n + 1); | 259 | let third = self.token_source.lookahead_nth(n + 2); |
255 | let la3 = self.token_source.token_kind(self.token_pos + n + 2); | 260 | |
261 | let jn1 = first.is_jointed_to_next; | ||
262 | let la2 = second.kind; | ||
263 | let jn2 = second.is_jointed_to_next; | ||
264 | let la3 = third.kind; | ||
256 | 265 | ||
257 | match kind { | 266 | match kind { |
258 | T![.] if jn1 && la2 == T![.] && jn2 && la3 == T![.] => Some((T![...], 3)), | 267 | T![.] if jn1 && la2 == T![.] && jn2 && la3 == T![.] => Some((T![...], 3)), |
@@ -271,9 +280,9 @@ impl<'t> Parser<'t> { | |||
271 | 280 | ||
272 | fn eat_dollars(&mut self) { | 281 | fn eat_dollars(&mut self) { |
273 | loop { | 282 | loop { |
274 | match self.token_source.token_kind(self.token_pos) { | 283 | match self.token_source.current().kind { |
275 | k @ SyntaxKind::L_DOLLAR | k @ SyntaxKind::R_DOLLAR => { | 284 | k @ SyntaxKind::L_DOLLAR | k @ SyntaxKind::R_DOLLAR => { |
276 | self.token_pos += 1; | 285 | self.token_source.bump(); |
277 | self.push_event(Event::Token { kind: k, n_raw_tokens: 1 }); | 286 | self.push_event(Event::Token { kind: k, n_raw_tokens: 1 }); |
278 | } | 287 | } |
279 | _ => { | 288 | _ => { |
@@ -286,9 +295,9 @@ impl<'t> Parser<'t> { | |||
286 | pub(crate) fn eat_l_dollars(&mut self) -> usize { | 295 | pub(crate) fn eat_l_dollars(&mut self) -> usize { |
287 | let mut ate_count = 0; | 296 | let mut ate_count = 0; |
288 | loop { | 297 | loop { |
289 | match self.token_source.token_kind(self.token_pos) { | 298 | match self.token_source.current().kind { |
290 | k @ SyntaxKind::L_DOLLAR => { | 299 | k @ SyntaxKind::L_DOLLAR => { |
291 | self.token_pos += 1; | 300 | self.token_source.bump(); |
292 | self.push_event(Event::Token { kind: k, n_raw_tokens: 1 }); | 301 | self.push_event(Event::Token { kind: k, n_raw_tokens: 1 }); |
293 | ate_count += 1; | 302 | ate_count += 1; |
294 | } | 303 | } |
@@ -302,9 +311,9 @@ impl<'t> Parser<'t> { | |||
302 | pub(crate) fn eat_r_dollars(&mut self, max_count: usize) -> usize { | 311 | pub(crate) fn eat_r_dollars(&mut self, max_count: usize) -> usize { |
303 | let mut ate_count = 0; | 312 | let mut ate_count = 0; |
304 | loop { | 313 | loop { |
305 | match self.token_source.token_kind(self.token_pos) { | 314 | match self.token_source.current().kind { |
306 | k @ SyntaxKind::R_DOLLAR => { | 315 | k @ SyntaxKind::R_DOLLAR => { |
307 | self.token_pos += 1; | 316 | self.token_source.bump(); |
308 | self.push_event(Event::Token { kind: k, n_raw_tokens: 1 }); | 317 | self.push_event(Event::Token { kind: k, n_raw_tokens: 1 }); |
309 | ate_count += 1; | 318 | ate_count += 1; |
310 | 319 | ||
@@ -320,12 +329,12 @@ impl<'t> Parser<'t> { | |||
320 | } | 329 | } |
321 | 330 | ||
322 | pub(crate) fn at_l_dollar(&self) -> bool { | 331 | pub(crate) fn at_l_dollar(&self) -> bool { |
323 | let kind = self.token_source.token_kind(self.token_pos); | 332 | let kind = self.token_source.current().kind; |
324 | (kind == SyntaxKind::L_DOLLAR) | 333 | (kind == SyntaxKind::L_DOLLAR) |
325 | } | 334 | } |
326 | 335 | ||
327 | pub(crate) fn at_r_dollar(&self) -> bool { | 336 | pub(crate) fn at_r_dollar(&self) -> bool { |
328 | let kind = self.token_source.token_kind(self.token_pos); | 337 | let kind = self.token_source.current().kind; |
329 | (kind == SyntaxKind::R_DOLLAR) | 338 | (kind == SyntaxKind::R_DOLLAR) |
330 | } | 339 | } |
331 | } | 340 | } |
diff --git a/crates/ra_syntax/src/parsing.rs b/crates/ra_syntax/src/parsing.rs index 15d69c5ab..4c1fa6c4f 100644 --- a/crates/ra_syntax/src/parsing.rs +++ b/crates/ra_syntax/src/parsing.rs | |||
@@ -17,8 +17,8 @@ pub(crate) use self::reparsing::incremental_reparse; | |||
17 | 17 | ||
18 | pub(crate) fn parse_text(text: &str) -> (GreenNode, Vec<SyntaxError>) { | 18 | pub(crate) fn parse_text(text: &str) -> (GreenNode, Vec<SyntaxError>) { |
19 | let tokens = tokenize(&text); | 19 | let tokens = tokenize(&text); |
20 | let token_source = text_token_source::TextTokenSource::new(text, &tokens); | 20 | let mut token_source = text_token_source::TextTokenSource::new(text, &tokens); |
21 | let mut tree_sink = text_tree_sink::TextTreeSink::new(text, &tokens); | 21 | let mut tree_sink = text_tree_sink::TextTreeSink::new(text, &tokens); |
22 | ra_parser::parse(&token_source, &mut tree_sink); | 22 | ra_parser::parse(&mut token_source, &mut tree_sink); |
23 | tree_sink.finish() | 23 | tree_sink.finish() |
24 | } | 24 | } |
diff --git a/crates/ra_syntax/src/parsing/reparsing.rs b/crates/ra_syntax/src/parsing/reparsing.rs index 6de02a15a..3b6687f61 100644 --- a/crates/ra_syntax/src/parsing/reparsing.rs +++ b/crates/ra_syntax/src/parsing/reparsing.rs | |||
@@ -85,9 +85,9 @@ fn reparse_block<'node>( | |||
85 | if !is_balanced(&tokens) { | 85 | if !is_balanced(&tokens) { |
86 | return None; | 86 | return None; |
87 | } | 87 | } |
88 | let token_source = TextTokenSource::new(&text, &tokens); | 88 | let mut token_source = TextTokenSource::new(&text, &tokens); |
89 | let mut tree_sink = TextTreeSink::new(&text, &tokens); | 89 | let mut tree_sink = TextTreeSink::new(&text, &tokens); |
90 | reparser.parse(&token_source, &mut tree_sink); | 90 | reparser.parse(&mut token_source, &mut tree_sink); |
91 | let (green, new_errors) = tree_sink.finish(); | 91 | let (green, new_errors) = tree_sink.finish(); |
92 | Some((node.replace_with(green), new_errors, node.range())) | 92 | Some((node.replace_with(green), new_errors, node.range())) |
93 | } | 93 | } |
diff --git a/crates/ra_syntax/src/parsing/text_token_source.rs b/crates/ra_syntax/src/parsing/text_token_source.rs index a6277f66f..71d2947f7 100644 --- a/crates/ra_syntax/src/parsing/text_token_source.rs +++ b/crates/ra_syntax/src/parsing/text_token_source.rs | |||
@@ -1,7 +1,8 @@ | |||
1 | use ra_parser::TokenSource; | 1 | use ra_parser::TokenSource; |
2 | use ra_parser::Token as PToken; | ||
2 | 3 | ||
3 | use crate::{ | 4 | use crate::{ |
4 | SyntaxKind, SyntaxKind::EOF, TextRange, TextUnit, | 5 | SyntaxKind::EOF, TextRange, TextUnit, |
5 | parsing::lexer::Token, | 6 | parsing::lexer::Token, |
6 | }; | 7 | }; |
7 | 8 | ||
@@ -23,31 +24,50 @@ pub(crate) struct TextTokenSource<'t> { | |||
23 | /// ``` | 24 | /// ``` |
24 | /// tokens: `[struct, Foo, {, }]` | 25 | /// tokens: `[struct, Foo, {, }]` |
25 | tokens: Vec<Token>, | 26 | tokens: Vec<Token>, |
27 | |||
28 | /// Current token and position | ||
29 | curr: (PToken, usize), | ||
26 | } | 30 | } |
27 | 31 | ||
28 | impl<'t> TokenSource for TextTokenSource<'t> { | 32 | impl<'t> TokenSource for TextTokenSource<'t> { |
29 | fn token_kind(&self, pos: usize) -> SyntaxKind { | 33 | fn current(&self) -> PToken { |
30 | if !(pos < self.tokens.len()) { | 34 | return self.curr.0; |
31 | return EOF; | ||
32 | } | ||
33 | self.tokens[pos].kind | ||
34 | } | 35 | } |
35 | fn is_token_joint_to_next(&self, pos: usize) -> bool { | 36 | |
36 | if !(pos + 1 < self.tokens.len()) { | 37 | fn lookahead_nth(&self, n: usize) -> PToken { |
37 | return true; | 38 | mk_token(self.curr.1 + n, &self.start_offsets, &self.tokens) |
39 | } | ||
40 | |||
41 | fn bump(&mut self) { | ||
42 | if self.curr.0.kind == EOF { | ||
43 | return; | ||
38 | } | 44 | } |
39 | self.start_offsets[pos] + self.tokens[pos].len == self.start_offsets[pos + 1] | 45 | |
46 | let pos = self.curr.1 + 1; | ||
47 | self.curr = (mk_token(pos, &self.start_offsets, &self.tokens), pos); | ||
40 | } | 48 | } |
41 | fn is_keyword(&self, pos: usize, kw: &str) -> bool { | 49 | |
50 | fn is_keyword(&self, kw: &str) -> bool { | ||
51 | let pos = self.curr.1; | ||
42 | if !(pos < self.tokens.len()) { | 52 | if !(pos < self.tokens.len()) { |
43 | return false; | 53 | return false; |
44 | } | 54 | } |
45 | let range = TextRange::offset_len(self.start_offsets[pos], self.tokens[pos].len); | 55 | let range = TextRange::offset_len(self.start_offsets[pos], self.tokens[pos].len); |
46 | |||
47 | self.text[range] == *kw | 56 | self.text[range] == *kw |
48 | } | 57 | } |
49 | } | 58 | } |
50 | 59 | ||
60 | fn mk_token(pos: usize, start_offsets: &[TextUnit], tokens: &[Token]) -> PToken { | ||
61 | let kind = tokens.get(pos).map(|t| t.kind).unwrap_or(EOF); | ||
62 | let is_jointed_to_next = if pos + 1 < start_offsets.len() { | ||
63 | start_offsets[pos] + tokens[pos].len == start_offsets[pos + 1] | ||
64 | } else { | ||
65 | false | ||
66 | }; | ||
67 | |||
68 | PToken { kind, is_jointed_to_next } | ||
69 | } | ||
70 | |||
51 | impl<'t> TextTokenSource<'t> { | 71 | impl<'t> TextTokenSource<'t> { |
52 | /// Generate input from tokens(expect comment and whitespace). | 72 | /// Generate input from tokens(expect comment and whitespace). |
53 | pub fn new(text: &'t str, raw_tokens: &'t [Token]) -> TextTokenSource<'t> { | 73 | pub fn new(text: &'t str, raw_tokens: &'t [Token]) -> TextTokenSource<'t> { |
@@ -62,6 +82,7 @@ impl<'t> TextTokenSource<'t> { | |||
62 | len += token.len; | 82 | len += token.len; |
63 | } | 83 | } |
64 | 84 | ||
65 | TextTokenSource { text, start_offsets, tokens } | 85 | let first = mk_token(0, &start_offsets, &tokens); |
86 | TextTokenSource { text, start_offsets, tokens, curr: (first, 0) } | ||
66 | } | 87 | } |
67 | } | 88 | } |