perl_parser_core/syntax/
text_line.rs

1//! Text-line cursor helpers.
2//!
3//! This crate has a single responsibility: map cursor offsets to line
4//! boundaries and provide conservative token-boundary primitives for
5//! single-line scanning.
6
7#![deny(unsafe_code)]
8#![warn(rust_2018_idioms)]
9#![warn(missing_docs)]
10
11/// Return the byte span of the line containing `cursor_pos`.
12///
13/// The returned range is inclusive of the first line byte and exclusive of
14/// one past the last byte, matching half-open Rust range conventions.
15#[must_use]
16pub fn line_bounds_at(text: &str, cursor_pos: usize) -> (usize, usize) {
17    let cursor = cursor_pos.min(text.len());
18    let start = text[..cursor].rfind('\n').map_or(0, |idx| idx + 1);
19    let end = text[cursor..].find('\n').map_or(text.len(), |idx| cursor + idx);
20    (start, end)
21}
22
23/// Return `true` when `byte` is an identifier character (`[A-Za-z0-9_]`).
24#[must_use]
25pub fn is_identifier_byte(byte: u8) -> bool {
26    byte.is_ascii_alphanumeric() || byte == b'_'
27}
28
29/// Return `true` when token `keyword` bytes in `[start, start + len)` are
30/// bounded on both sides by non-identifier bytes.
31#[must_use]
32pub fn is_keyword_boundary(bytes: &[u8], start: usize, len: usize) -> bool {
33    if start > bytes.len() {
34        return false;
35    }
36
37    let end = start.saturating_add(len);
38    if end > bytes.len() {
39        return false;
40    }
41
42    if start > 0 && is_identifier_byte(bytes[start - 1]) {
43        return false;
44    }
45
46    if end < bytes.len() && is_identifier_byte(bytes[end]) {
47        return false;
48    }
49
50    true
51}
52
53/// Advance `idx` while bytes at the cursor are ASCII whitespace.
54#[must_use]
55pub fn skip_ascii_whitespace(bytes: &[u8], mut idx: usize) -> usize {
56    while idx < bytes.len() && bytes[idx].is_ascii_whitespace() {
57        idx += 1;
58    }
59    idx
60}
perl_parser_core/syntax/text_line.rs

perl_parser_core/syntax/
text_line.rs
