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
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
// SPDX-License-Identifier: MIT
// Copyright (c) 2026 Moderately AI Inc.
//! Lexical token types: the M1 token categories and their operator/punctuation
//! sub-kinds.
use crateSpan;
use crateKeyword;
/// A single lexical token: a category plus its byte range in the source.
///
/// `Token` is `Copy` and deliberately carries no borrow. The token *text* is
/// recovered later as `&source[span]` (zero-copy tokens): keeping the
/// token free of a `&str` makes the token stream cache-dense and decouples it
/// from the source buffer's lifetime, so a `Vec<Token>` can outlive the borrow
/// the cursor held while scanning.
/// The M1 lexical categories.
///
/// Literal *unescaping* is deferred to lazy materialization: a
/// [`String`] or [`QuotedIdent`] span still includes its delimiters and raw
/// escapes.
///
/// [`String`]: TokenKind::String
/// [`QuotedIdent`]: TokenKind::QuotedIdent
/// Operator spellings recognized in M1.
///
/// The set is *closed over the operator bytes* the shared lexer-class table
/// marks (`+ - * / % = < > ! | & ^ ~`): every such byte maps to a variant here,
/// either alone or as the lead of a fixed two-byte operator (`<=`, `>=`, `<>`,
/// `!=`, `||`, and the always-munched shifts `<<`/`>>`). That keeps the enum small
/// and total without rejecting a byte the dialect data calls an operator. Operator
/// *precedence and semantics* live in the AST precedence data and the parser, not here.
///
/// A handful of spellings are recognised only when their dialect feature is on, so
/// they are scanned with the feature data rather than the byte-class table:
/// - the named-argument arrows [`Arrow`](Self::Arrow) (`=>`) and
/// [`ColonEquals`](Self::ColonEquals) (`:=`) under named arguments (the `:=` lead
/// byte is punctuation-class, so it is munched in the punctuation scanner);
/// - the PostgreSQL containment and JSON operators [`AtGt`](Self::AtGt) (`@>`),
/// [`LtAt`](Self::LtAt) (`<@`), [`MinusGt`](Self::MinusGt) (`->`), and
/// [`MinusGtGt`](Self::MinusGtGt) (`->>`), gated by the containment / JSON-arrow
/// expression-syntax flags, and PostgreSQL's XOR [`Hash`](Self::Hash) (`#`), gated by
/// [`FeatureSet::hash_bitwise_xor`](crate::ast::dialect::FeatureSet). The `@` and `#` lead bytes are
/// not in the lexer-class table, so `@>` / `#` reach the operator scanner through a
/// feature-gated dispatch arm. The PostgreSQL prefix `@` absolute-value operator is a
/// scoped follow-up (its `@` lexeme contends with the T-SQL/MySQL `@name` sigils,
/// needing a tracked conflict).
/// - the PostgreSQL `jsonb` operators [`Question`](Self::Question) (`?`),
/// [`QuestionPipe`](Self::QuestionPipe) (`?|`), [`QuestionAmp`](Self::QuestionAmp) (`?&`),
/// [`AtQuestion`](Self::AtQuestion) (`@?`), [`AtAt`](Self::AtAt) (`@@`),
/// [`HashGt`](Self::HashGt) (`#>`), [`HashGtGt`](Self::HashGtGt) (`#>>`), and
/// [`HashMinus`](Self::HashMinus) (`#-`), gated by
/// [`OperatorSyntax::jsonb_operators`](crate::ast::dialect::OperatorSyntax::jsonb_operators).
/// The `?` and `@?`/`@@` forms reach the scanner through feature-gated dispatch arms (like
/// `@>`); the `#`-led forms are munched inside the scanner ahead of the bare `#` XOR.
/// Structural punctuation recognized in M1.
///
/// Closed over the punctuation bytes the shared lexer-class table marks
/// (`( ) , ; . [ ] { } :`), so the tokenizer never rejects a byte the dialect
/// data calls punctuation. The `:` byte is the one with two forms: `::` (the
/// PostgreSQL typecast operator) takes maximal-munch priority over a lone `:`
/// (the array-slice separator).