omnicoach 0.3.0

Lightweight, high-signal static analysis and command linter for Python and Rust
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

//! Multiline string literal analysis for Rust (`ra_ap_syntax`).

use super::attributes::attribute_terminal_name;
use super::names::resolve_preceding_macro_path;
use super::trivia::previous_non_trivia_token;
use crate::code_lint::ast::multiline_strings::{
    MultilineKind, MultilineString, escaped_multiline_newlines,
};
use crate::code_lint::ast::tree::{AstNode, ParsedFile, span_from_rowan_range};
use ra_ap_syntax::ast;
use ra_ap_syntax::{AstNode as _, AstToken as _, SyntaxElement, SyntaxKind, SyntaxToken};

/// Collects all multiline string literals in a Rust `file` (see
/// [`crate::code_lint::ast::collect_multiline_strings`]).
#[must_use]
pub fn collect_multiline_strings(file: &ParsedFile) -> Vec<MultilineString<'_>> {
    let Some(parsed) = file.rs_parsed() else {
        return Vec::new();
    };
    parsed
        .tree()
        .syntax()
        .descendants_with_tokens()
        .filter_map(SyntaxElement::into_token)
        .filter_map(|token| {
            let kind = token_multiline_kind(file, &token)?;
            Some(MultilineString {
                node: AstNode::from_span(file, span_from_rowan_range(token.text_range())),
                kind,
            })
        })
        .collect()
}

/// Classifies `token` as a source or escaped multiline string literal, or `None` when it is not a
/// string literal, is a doc attribute or `insta` inline snapshot, or is a single-line literal.
fn token_multiline_kind(file: &ParsedFile, token: &SyntaxToken) -> Option<MultilineKind> {
    if !matches!(
        token.kind(),
        SyntaxKind::STRING | SyntaxKind::BYTE_STRING | SyntaxKind::C_STRING
    ) || is_insta_inline_snapshot(token)
        || is_enclosed_in_doc_attribute(token)
    {
        return None;
    }
    if is_source_multiline_token(token) {
        return Some(MultilineKind::Source {
            enclosing_calls: enclosing_macro_paths(file, token),
        });
    }
    token_escaped_newlines(token).map(|newlines| MultilineKind::Escaped { newlines })
}

/// Returns true if `token` spans multiple source lines and contains runtime newlines (raw strings
/// spanning lines, or standard strings with at least one non-final line not ending in `\`).
#[must_use]
fn is_source_multiline_token(token: &SyntaxToken) -> bool {
    let text = token.text();
    if !text.contains('\n') {
        return false;
    }
    if text.starts_with('r') || text.starts_with("br") || text.starts_with("cr") {
        return true;
    }
    let mut lines = text.lines();
    lines.next_back();
    lines.any(|line| !line.trim_end().ends_with('\\'))
}

/// Decodes a single-line `STRING`, `BYTE_STRING`, or `C_STRING` token and returns its escaped
/// `\n` count when it spans at least two non-blank lines.
fn token_escaped_newlines(token: &SyntaxToken) -> Option<usize> {
    ast::String::cast(token.clone())
        .and_then(|string| {
            string
                .value()
                .ok()
                .and_then(|val| escaped_multiline_newlines(val.as_bytes()))
        })
        .or_else(|| {
            ast::ByteString::cast(token.clone()).and_then(|byte_string| {
                byte_string
                    .value()
                    .ok()
                    .and_then(|val| escaped_multiline_newlines(&val))
            })
        })
        .or_else(|| {
            ast::CString::cast(token.clone()).and_then(|c_string| {
                c_string
                    .value()
                    .ok()
                    .and_then(|val| escaped_multiline_newlines(&val))
            })
        })
}

/// Returns true if `token` is preceded by `@` (an `insta` inline snapshot literal `@"..."`).
#[must_use]
fn is_insta_inline_snapshot(token: &SyntaxToken) -> bool {
    previous_non_trivia_token(token).is_some_and(|previous| previous.kind() == SyntaxKind::AT)
}

/// Returns true if `token` is enclosed inside a `#[doc = "..."]` attribute.
#[must_use]
fn is_enclosed_in_doc_attribute(token: &SyntaxToken) -> bool {
    token
        .parent_ancestors()
        .filter_map(ast::Attr::cast)
        .any(|attribute| is_doc_attribute(&attribute))
}

/// Returns true if `attribute` represents a `#[doc = "..."]` attribute.
#[must_use]
fn is_doc_attribute(attribute: &ast::Attr) -> bool {
    attribute_terminal_name(attribute).is_some_and(|terminal| terminal.text() == "doc")
}

/// Collects the canonical import-resolved paths of non-local macro invocations enclosing `token`
/// within its innermost function or closure scope.
fn enclosing_macro_paths(file: &ParsedFile, token: &SyntaxToken) -> Vec<String> {
    token
        .parent_ancestors()
        .take_while(|ancestor| {
            !matches!(ancestor.kind(), SyntaxKind::FN | SyntaxKind::CLOSURE_EXPR)
        })
        .filter(|ancestor| ancestor.kind() == SyntaxKind::TOKEN_TREE)
        .filter_map(|ancestor| {
            resolve_preceding_macro_path(&ancestor)
                .and_then(|name| name.resolve(file))
                .map(|name| name.path)
        })
        .collect()
}