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
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
//! The Davinci span type: two byte offsets into the authored file.
//!
//! `Span` is the L0 source-model coordinate (`docs/davinci/architecture.md`):
//! every node records where it came from as a half-open byte range
//! `[start, end)` into the one authored source string, and nothing else.
//! Text is recovered on demand with [`Span::slice`]; line/column exist only
//! at diagnostic-rendering time, derived from offsets (see
//! [`crate::line_index::LineIndex`]). Since Davinci P1-3 the relief
//! `SourceLocation` carries one of these instead of an owned `source: String`
//! copy of the covered text; the migration record lives in
//! `docs/davinci/plan/sourcelocation-inventory.md`.
//!
//! # Offset contract
//!
//! Offsets are **UTF-8 byte offsets** measured from the start of the file the
//! span was produced against. Producers (parsers) only ever cut on character
//! boundaries, so well-formed spans always land on `char` boundaries of that
//! source. [`Span::slice`] is nevertheless total: it never panics on
//! malformed input, it clamps and snaps as documented there.
/// A half-open byte range `[start, end)` into an authored source string.
#[derive(Copy, Clone, PartialEq, Eq, Hash, Debug, serde::Serialize, serde::Deserialize)]
pub struct Span {
/// Start byte offset (inclusive).
pub start: u32,
/// End byte offset (exclusive).
pub end: u32,
}
/// `Span` is passed and stored by value everywhere; keep it two words at most.
const _: () = assert!(core::mem::size_of::<Span>() == 8);
impl Span {
/// Create a new span. No ordering is enforced; consumers treat an
/// inverted span (`start > end`) as empty.
#[inline]
pub const fn new(start: u32, end: u32) -> Self {
Self { start, end }
}
/// Byte length of the span. Inverted spans (`start > end`) saturate to 0.
#[inline]
pub const fn len(&self) -> u32 {
self.end.saturating_sub(self.start)
}
/// Whether the span covers no bytes (`start >= end`).
#[inline]
pub const fn is_empty(&self) -> bool {
self.start >= self.end
}
/// Slice `source` to the text this span covers.
///
/// Never panics. The exact semantics, in order:
///
/// 1. `end` is clamped to `source.len()`, then `start` is clamped to the
/// clamped `end` (an inverted or out-of-range span yields `""`).
/// 2. Each clamped offset is snapped **down** to the nearest `char`
/// boundary of `source`. Snapping is monotone, so the result is always
/// a valid, possibly empty, subslice.
///
/// For the documented contract — byte offsets produced on character
/// boundaries of `source` — both steps are identity and this is exactly
/// `&source[start..end]`. Callers slicing a *different* string than the
/// one the span was measured against get deterministic garbage, not a
/// panic; that mismatch is a caller bug the type cannot detect.
#[inline]
pub fn slice<'a>(&self, source: &'a str) -> &'a str {
let end = (self.end as usize).min(source.len());
let start = (self.start as usize).min(end);
source
.get(floor_char_boundary(source, start)..floor_char_boundary(source, end))
.unwrap_or_default()
}
/// Rebase this span from file-absolute offsets to block-relative offsets
/// by subtracting `block_start` from both ends.
///
/// Saturating: offsets before `block_start` clamp to 0, so a span that
/// starts before the block degenerates toward an empty span at the block
/// start instead of wrapping.
#[inline]
pub const fn to_block_relative(self, block_start: u32) -> Span {
Span {
start: self.start.saturating_sub(block_start),
end: self.end.saturating_sub(block_start),
}
}
/// Hash the **block-relative** form of this span, per the rustc
/// relative-span import: equal content at different absolute positions
/// (e.g. the same template block shifted by an edit above it) hashes
/// equally when each side passes its own block start.
///
/// Exactly equivalent to `self.to_block_relative(block_start)` hashed
/// with the derived [`core::hash::Hash`] impl.
#[inline]
pub fn hash_relative<H: core::hash::Hasher>(&self, hasher: &mut H, block_start: u32) {
core::hash::Hash::hash(&self.to_block_relative(block_start), hasher);
}
}
/// Largest index `<= index` that is a `char` boundary of `source`.
///
/// `index` must be `<= source.len()` (callers clamp first). A UTF-8 sequence
/// is at most 4 bytes, so at most 3 probes run.
#[inline]
fn floor_char_boundary(source: &str, mut index: usize) -> usize {
while !source.is_char_boundary(index) {
index -= 1;
}
index
}
#[cfg(test)]
mod tests {
use core::hash::{Hash, Hasher};
use std::hash::DefaultHasher;
use super::Span;
fn hash_of(value: impl Hash) -> u64 {
let mut hasher = DefaultHasher::new();
value.hash(&mut hasher);
hasher.finish()
}
fn relative_hash_of(span: Span, block_start: u32) -> u64 {
let mut hasher = DefaultHasher::new();
span.hash_relative(&mut hasher, block_start);
hasher.finish()
}
#[test]
fn size_is_eight_bytes() {
assert_eq!(core::mem::size_of::<Span>(), 8);
assert_eq!(core::mem::align_of::<Span>(), 4);
}
#[test]
fn new_stores_offsets_verbatim() {
let span = Span::new(3, 7);
assert_eq!(span.start, 3);
assert_eq!(span.end, 7);
let inverted = Span::new(7, 3);
assert_eq!(inverted.start, 7);
assert_eq!(inverted.end, 3);
}
#[test]
fn len_counts_bytes() {
assert_eq!(Span::new(3, 7).len(), 4);
assert_eq!(Span::new(0, 0).len(), 0);
assert_eq!(Span::new(5, 5).len(), 0);
assert_eq!(Span::new(0, u32::MAX).len(), u32::MAX);
}
#[test]
fn len_saturates_on_inverted_spans() {
assert_eq!(Span::new(7, 3).len(), 0);
assert_eq!(Span::new(u32::MAX, 0).len(), 0);
}
#[test]
fn is_empty_is_start_at_or_past_end() {
assert!(Span::new(0, 0).is_empty());
assert!(Span::new(5, 5).is_empty());
assert!(Span::new(7, 3).is_empty());
assert!(!Span::new(3, 7).is_empty());
assert!(!Span::new(0, 1).is_empty());
}
#[test]
fn slice_returns_covered_text() {
let source = "let x = 1;";
assert_eq!(Span::new(0, 3).slice(source), "let");
assert_eq!(Span::new(4, 5).slice(source), "x");
assert_eq!(Span::new(0, 10).slice(source), "let x = 1;");
}
#[test]
fn slice_of_empty_span_is_empty() {
assert_eq!(Span::new(4, 4).slice("let x = 1;"), "");
assert_eq!(Span::new(0, 0).slice(""), "");
}
#[test]
fn slice_at_end_of_source() {
let source = "abc";
assert_eq!(Span::new(3, 3).slice(source), "");
assert_eq!(Span::new(2, 3).slice(source), "c");
}
#[test]
fn slice_clamps_end_past_source() {
let source = "abc";
assert_eq!(Span::new(1, 100).slice(source), "bc");
assert_eq!(Span::new(0, u32::MAX).slice(source), "abc");
}
#[test]
fn slice_clamps_start_past_source_to_empty() {
let source = "abc";
assert_eq!(Span::new(50, 100).slice(source), "");
assert_eq!(Span::new(3, 100).slice(source), "");
}
#[test]
fn slice_of_inverted_span_is_empty() {
assert_eq!(Span::new(7, 3).slice("let x = 1;"), "");
assert_eq!(Span::new(u32::MAX, 0).slice("abc"), "");
}
#[test]
fn slice_on_char_boundaries_of_multibyte_source() {
// "aéb": a=0..1, é=1..3, b=3..4
let source = "a\u{e9}b";
assert_eq!(Span::new(1, 3).slice(source), "\u{e9}");
assert_eq!(Span::new(0, 4).slice(source), "a\u{e9}b");
assert_eq!(Span::new(3, 4).slice(source), "b");
}
#[test]
fn slice_snaps_mid_char_offsets_down() {
// "aéb": offset 2 is inside the two-byte é (1..3).
let source = "a\u{e9}b";
// Mid-char start snaps down to 1: the é is included.
assert_eq!(Span::new(2, 4).slice(source), "\u{e9}b");
// Mid-char end snaps down to 1: nothing of the é remains.
assert_eq!(Span::new(1, 2).slice(source), "");
// Both ends inside the same char collapse to empty.
assert_eq!(Span::new(2, 2).slice(source), "");
// Four-byte scalar: "🦀" is 0..4; every interior offset snaps to 0.
let crab = "\u{1f980}";
assert_eq!(Span::new(1, 3).slice(crab), "");
assert_eq!(Span::new(0, 3).slice(crab), "");
assert_eq!(Span::new(0, 4).slice(crab), "\u{1f980}");
}
#[test]
fn to_block_relative_subtracts_block_start() {
assert_eq!(Span::new(10, 14).to_block_relative(10), Span::new(0, 4));
assert_eq!(Span::new(10, 14).to_block_relative(3), Span::new(7, 11));
assert_eq!(Span::new(10, 14).to_block_relative(0), Span::new(10, 14));
}
#[test]
fn to_block_relative_saturates_before_block_start() {
// Block start inside the span: start clamps to 0, end still rebases.
assert_eq!(Span::new(10, 14).to_block_relative(12), Span::new(0, 2));
// Block start past the span: both clamp to 0.
assert_eq!(Span::new(10, 14).to_block_relative(20), Span::new(0, 0));
assert_eq!(Span::new(0, 0).to_block_relative(u32::MAX), Span::new(0, 0));
}
#[test]
fn hash_relative_equal_content_at_different_absolute_positions() {
// The same 4-byte content in blocks starting at 10 and 50.
assert_eq!(
relative_hash_of(Span::new(10, 14), 10),
relative_hash_of(Span::new(50, 54), 50),
);
// Offset within the block is preserved: 12..14 in block 10 matches
// 52..54 in block 50.
assert_eq!(
relative_hash_of(Span::new(12, 14), 10),
relative_hash_of(Span::new(52, 54), 50),
);
}
#[test]
fn hash_relative_distinguishes_block_offset_and_length() {
// Same absolute span, different block starts: different relative form.
assert_ne!(
relative_hash_of(Span::new(12, 14), 10),
relative_hash_of(Span::new(12, 14), 12),
);
// Same start, different lengths.
assert_ne!(
relative_hash_of(Span::new(10, 14), 10),
relative_hash_of(Span::new(10, 15), 10),
);
}
#[test]
fn hash_relative_matches_derived_hash_of_rebased_span() {
let span = Span::new(37, 41);
assert_eq!(
relative_hash_of(span, 30),
hash_of(span.to_block_relative(30)),
);
// With block_start 0 it is exactly the derived hash of the span.
assert_eq!(relative_hash_of(span, 0), hash_of(span));
}
}