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
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
/// Output of PlaceholderIter
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub enum PlaceholderPart<'a> {
Sql(&'a str),
Placeholder(&'a str), // name WITHOUT ':'
}
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
enum Mode {
Normal,
SingleQuote,
DoubleQuote,
Backtick,
BracketIdent,
LineComment,
BlockComment,
DollarQuote { tag_start: usize, tag_end: usize }, // tag is s[tag_start..tag_end], may be empty
}
/// Non-allocating, panic-free iterator over SQL + :name placeholders.
/// - Only detects ASCII placeholders: :[A-Za-z_][A-Za-z0-9_]*
/// - Skips: strings, quoted identifiers, line/block comments, PG dollar quotes ($$ or $tag$)
pub struct PlaceholderIter<'a> {
s: &'a str,
b: &'a [u8],
i: usize, // scan cursor (byte index)
chunk_start: usize,
mode: Mode,
done: bool,
}
impl<'a> PlaceholderIter<'a> {
pub fn new(s: &'a str) -> Self {
Self {
s,
b: s.as_bytes(),
i: 0,
chunk_start: 0,
mode: Mode::Normal,
done: false,
}
}
#[inline]
fn is_name_start(x: u8) -> bool {
x.is_ascii_lowercase() || x.is_ascii_uppercase() || x == b'_'
}
#[inline]
fn is_name_char(x: u8) -> bool {
Self::is_name_start(x) || x.is_ascii_digit()
}
#[inline]
fn is_word_char(x: u8) -> bool {
Self::is_name_char(x)
}
#[inline]
fn slice(&self, a: usize, b: usize) -> &'a str {
// Safe because we only advance on UTF-8 character boundaries.
// Indices are always in-range by construction (checked before advancing).
&self.s[a..b]
}
/// Advance by one UTF-8 character. Returns the byte length of the character.
/// This ensures we never land in the middle of a multi-byte character.
/// Returns 0 if at end of string. Steps 1 byte on invalid UTF-8.
#[inline]
fn advance_one_char(&mut self) -> usize {
let Some(&b0) = self.b.get(self.i) else {
return 0;
};
// Fast path for ASCII (vast majority of SQL)
if b0 < 0x80 {
self.i += 1;
return 1;
}
// Determine UTF-8 character length from leading byte
let len = if (b0 & 0b1110_0000) == 0b1100_0000 {
2
} else if (b0 & 0b1111_0000) == 0b1110_0000 {
3
} else if (b0 & 0b1111_1000) == 0b1111_0000 {
4
} else {
1 // Invalid lead byte, step 1
};
let end = self.i.saturating_add(len).min(self.b.len());
// Validate continuation bytes (must match 0b10xx_xxxx pattern)
if end - self.i == len
&& self.b[self.i + 1..end]
.iter()
.all(|&c| (c & 0b1100_0000) == 0b1000_0000)
{
self.i = end;
len
} else {
// Invalid UTF-8 sequence, step 1 byte
self.i += 1;
1
}
}
#[inline]
fn peek(&self) -> Option<u8> {
self.b.get(self.i).copied()
}
#[inline]
fn peek_n(&self, n: usize) -> Option<u8> {
self.b.get(self.i + n).copied()
}
#[inline]
fn starts_with_at(&self, pos: usize, pat: &[u8]) -> bool {
self.b.get(pos..pos + pat.len()) == Some(pat)
}
// Attempts to parse a dollar-quote opener at current `i` (which must be on '$').
// On success, sets mode and advances `i` past opener, returning true.
fn try_enter_dollar_quote(&mut self) -> bool {
// at '$'
// $$ ...
if self.starts_with_at(self.i, b"$$") {
self.mode = Mode::DollarQuote {
tag_start: self.i + 1,
tag_end: self.i + 1, // empty tag
};
self.advance_one_char(); // first $
self.advance_one_char(); // second $
return true;
}
// $tag$
let tag_start = self.i + 1;
let first = self.b.get(tag_start).copied();
if first.is_none_or(|b| !Self::is_name_start(b)) {
return false;
}
let mut j = tag_start + 1;
while let Some(&ch) = self.b.get(j) {
if Self::is_name_char(ch) {
j += 1;
} else {
break;
}
}
// must end with '$'
if self.b.get(j).copied() == Some(b'$') {
self.mode = Mode::DollarQuote {
tag_start,
tag_end: j, // excludes '$'
};
// Advance past $ + tag + $
// Tag is ASCII (alphanumeric/underscore) so byte-wise is safe
self.i = j + 1;
return true;
}
false
}
// Attempts to detect dollar-quote closer at current `i` (which must be on '$').
// On success, consumes closer and returns true.
fn try_exit_dollar_quote(&mut self) -> bool {
let (tag_start, tag_end) = match self.mode {
Mode::DollarQuote { tag_start, tag_end } => (tag_start, tag_end),
_ => return false,
};
// Empty tag: "$$"
if tag_start == tag_end {
if self.starts_with_at(self.i, b"$$") {
self.advance_one_char(); // first $
self.advance_one_char(); // second $
self.mode = Mode::Normal;
return true;
}
return false;
}
// Non-empty: "$tag$"
// Need: '$' + tag + '$'
let tag = self.b.get(tag_start..tag_end).unwrap_or(&[]);
if self.peek() != Some(b'$') {
return false;
}
let after_dollar = self.i + 1;
if self.b.get(after_dollar..after_dollar + tag.len()) != Some(tag) {
return false;
}
if self.b.get(after_dollar + tag.len()).copied() != Some(b'$') {
return false;
}
// Advance past the closing delimiter: $ + tag + $
self.advance_one_char(); // opening $
// Tag is ASCII (alphanumeric/underscore), so byte-wise advancement is safe
self.i += tag.len();
self.advance_one_char(); // closing $
self.mode = Mode::Normal;
true
}
}
impl<'a> Iterator for PlaceholderIter<'a> {
type Item = PlaceholderPart<'a>;
fn next(&mut self) -> Option<Self::Item> {
if self.done {
return None;
}
while self.i < self.b.len() {
match self.mode {
Mode::Normal => {
let c = self.b[self.i];
// -- line comment
if c == b'-' && self.peek_n(1) == Some(b'-') {
self.advance_one_char(); // first -
self.advance_one_char(); // second -
self.mode = Mode::LineComment;
continue;
}
// # line comment
if c == b'#' {
self.advance_one_char();
self.mode = Mode::LineComment;
continue;
}
// /* block comment */
if c == b'/' && self.peek_n(1) == Some(b'*') {
self.advance_one_char(); // /
self.advance_one_char(); // *
self.mode = Mode::BlockComment;
continue;
}
// quotes / identifiers
if c == b'\'' {
self.advance_one_char();
self.mode = Mode::SingleQuote;
continue;
}
if c == b'"' {
self.advance_one_char();
self.mode = Mode::DoubleQuote;
continue;
}
if c == b'`' {
self.advance_one_char();
self.mode = Mode::Backtick;
continue;
}
if c == b'[' {
self.advance_one_char();
self.mode = Mode::BracketIdent;
continue;
}
// dollar quote
if c == b'$' {
if self.try_enter_dollar_quote() {
continue;
}
self.advance_one_char();
continue;
}
// placeholder :name (NOT :: and NOT abc:name)
if c == b':' {
let prev = if self.i == 0 {
None
} else {
Some(self.b[self.i - 1])
};
// Only reject if prev is a word char AND we didn't just emit a placeholder
let prev_is_word = prev.map(Self::is_word_char).unwrap_or(false);
let prev_is_colon = prev == Some(b':');
let just_emitted_placeholder = self.chunk_start == self.i;
if !prev_is_colon
&& (!prev_is_word || just_emitted_placeholder)
&& let Some(n0) = self.peek_n(1)
&& Self::is_name_start(n0)
{
let name_start = self.i + 1;
let mut j = name_start + 1;
while let Some(&ch) = self.b.get(j) {
if Self::is_name_char(ch) {
j += 1;
} else {
break;
}
}
// Emit SQL chunk before placeholder, then placeholder in next call.
if self.chunk_start < self.i {
let sql = self.slice(self.chunk_start, self.i);
self.chunk_start = self.i; // placeholder begins here
return Some(PlaceholderPart::Sql(sql));
}
// Emit placeholder (name only), advance past it.
let name = self.slice(name_start, j);
self.i = j;
self.chunk_start = j;
return Some(PlaceholderPart::Placeholder(name));
}
self.advance_one_char();
continue;
}
self.advance_one_char();
}
Mode::SingleQuote => {
match self.peek() {
Some(b'\\') => {
// backslash: skip it and next char if present
self.advance_one_char();
if self.peek().is_some() {
self.advance_one_char();
}
}
Some(b'\'') => {
if self.peek_n(1) == Some(b'\'') {
self.advance_one_char(); // first '
self.advance_one_char(); // second '
} else {
self.advance_one_char();
self.mode = Mode::Normal;
}
}
_ => {
self.advance_one_char();
}
}
}
Mode::DoubleQuote => {
match self.peek() {
Some(b'\\') => {
// backslash: skip it and next char if present
self.advance_one_char();
if self.peek().is_some() {
self.advance_one_char();
}
}
Some(b'"') => {
if self.peek_n(1) == Some(b'"') {
self.advance_one_char(); // first "
self.advance_one_char(); // second "
} else {
self.advance_one_char();
self.mode = Mode::Normal;
}
}
_ => {
self.advance_one_char();
}
}
}
Mode::Backtick => {
match self.peek() {
Some(b'\\') => {
// backslash: skip it and next char if present
self.advance_one_char();
if self.peek().is_some() {
self.advance_one_char();
}
}
Some(b'`') => {
if self.peek_n(1) == Some(b'`') {
self.advance_one_char(); // first `
self.advance_one_char(); // second `
} else {
self.advance_one_char();
self.mode = Mode::Normal;
}
}
_ => {
self.advance_one_char();
}
}
}
Mode::BracketIdent => {
if self.peek() == Some(b']') {
self.advance_one_char();
self.mode = Mode::Normal;
} else {
self.advance_one_char();
}
}
Mode::LineComment => {
if self.peek() == Some(b'\n') {
self.advance_one_char();
self.mode = Mode::Normal;
} else {
self.advance_one_char();
}
}
Mode::BlockComment => {
if self.peek() == Some(b'*') && self.peek_n(1) == Some(b'/') {
self.advance_one_char(); // *
self.advance_one_char(); // /
self.mode = Mode::Normal;
} else {
self.advance_one_char();
}
}
Mode::DollarQuote { .. } => {
// scan until we see a '$' that matches closer
if self.peek() == Some(b'$') && self.try_exit_dollar_quote() {
continue;
}
self.advance_one_char();
}
}
}
// end: flush remaining SQL chunk
self.done = true;
if self.chunk_start < self.s.len() {
Some(PlaceholderPart::Sql(
self.slice(self.chunk_start, self.s.len()),
))
} else {
None
}
}
}