Skip to content
Navigation Menu
{{ message }}
-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathtokenizer.h
More file actions
426 lines (382 loc) · 16.2 KB
/
Copy pathtokenizer.h
File metadata and controls
426 lines (382 loc) · 16.2 KB
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
#ifndef SQL_PARSER_TOKENIZER_H
#define SQL_PARSER_TOKENIZER_H
#include "sql_parser/token.h"
#include "sql_parser/keywords_mysql.h"
#include "sql_parser/keywords_pgsql.h"
namespace sql_parser {
template <Dialect D>
class Tokenizer {
public:
void reset(const char* input, size_t len) {
start_ = input;
cursor_ = input;
end_ = input + len;
has_peeked_ = false;
has_error_ = false;
}
// True iff the tokenizer has emitted at least one TK_ERROR token since
// the last reset(). Lets callers distinguish "couldn't build an AST"
// from "the input was syntactically invalid" and surface the latter
// as ParseResult::ERROR rather than PARTIAL.
bool has_error() const { return has_error_; }
// Hook for parser-level (non-tokenizer) errors -- when the parser
// detects clearly invalid input (e.g. `SET = X`, `SET x = ;`,
// `SET x = ,foo`) it can flag the error so the eventual ParseResult
// is ERROR rather than PARTIAL with a null AST.
void flag_error() { has_error_ = true; }
Token next_token() {
if (has_peeked_) {
has_peeked_ = false;
return peeked_;
}
return scan_token();
}
Token peek() {
if (!has_peeked_) {
peeked_ = scan_token();
has_peeked_ = true;
}
return peeked_;
}
void skip() {
if (has_peeked_) {
has_peeked_ = false;
} else {
scan_token();
}
}
// Expose start/end of input for remaining-input calculation and for
// safe lookback (e.g. detecting whether an identifier token was source-
// delimited via a preceding backtick / double-quote byte).
const char* input_begin() const { return start_; }
const char* input_end() const { return end_; }
private:
const char* start_ = nullptr;
const char* cursor_ = nullptr;
const char* end_ = nullptr;
Token peeked_;
bool has_peeked_ = false;
bool has_error_ = false;
uint32_t offset() const {
return static_cast<uint32_t>(cursor_ - start_);
}
char current() const { return (cursor_ < end_) ? *cursor_ : '\0'; }
char advance() {
char c = current();
if (cursor_ < end_) ++cursor_;
return c;
}
char peek_char(size_t ahead = 0) const {
const char* p = cursor_ + ahead;
return (p < end_) ? *p : '\0';
}
void skip_whitespace_and_comments() {
while (cursor_ < end_) {
char c = *cursor_;
// Whitespace
if (c == ' ' || c == '\t' || c == '\n' || c == '\r') {
++cursor_;
continue;
}
// -- line comment (MySQL requires space after --, PgSQL doesn't but we handle both)
if (c == '-' && peek_char(1) == '-') {
cursor_ += 2;
while (cursor_ < end_ && *cursor_ != '\n') ++cursor_;
continue;
}
// # line comment (MySQL only)
if constexpr (D == Dialect::MySQL) {
if (c == '#') {
++cursor_;
while (cursor_ < end_ && *cursor_ != '\n') ++cursor_;
continue;
}
}
// /* block comment */
if (c == '/' && peek_char(1) == '*') {
cursor_ += 2;
if constexpr (D == Dialect::PostgreSQL) {
// PostgreSQL supports nested block comments
int depth = 1;
while (cursor_ < end_ && depth > 0) {
if (*cursor_ == '/' && peek_char(1) == '*') {
++depth;
cursor_ += 2;
} else if (*cursor_ == '*' && peek_char(1) == '/') {
--depth;
cursor_ += 2;
} else {
++cursor_;
}
}
} else {
// MySQL: no nesting
while (cursor_ < end_) {
if (*cursor_ == '*' && peek_char(1) == '/') {
cursor_ += 2;
break;
}
++cursor_;
}
}
continue;
}
break; // not whitespace or comment
}
}
Token make_token(TokenType type, const char* start, uint32_t len) {
if (type == TokenType::TK_ERROR) has_error_ = true;
return Token{type, StringRef{start, len},
static_cast<uint32_t>(start - start_)};
}
Token scan_identifier_or_keyword() {
const char* start = cursor_;
while (cursor_ < end_) {
char c = *cursor_;
bool is_cont = (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') ||
(c >= '0' && c <= '9') || c == '_';
// PostgreSQL allows `$` as an identifier continuation char (but
// not as the first char, which is enforced because $ at start
// is handled by the $$ / $N branches in next_token_impl()).
// e.g. `SET search_path = schema$1` — `schema$1` is a single
// identifier, not `schema` followed by the placeholder `$1`.
if (!is_cont && D == Dialect::PostgreSQL && c == '$' && cursor_ > start) {
is_cont = true;
}
if (is_cont) {
++cursor_;
} else {
break;
}
}
uint32_t len = static_cast<uint32_t>(cursor_ - start);
// Keyword lookup
TokenType kw;
if constexpr (D == Dialect::MySQL) {
kw = mysql_keywords::lookup(start, len);
} else {
kw = pgsql_keywords::lookup(start, len);
}
return make_token(kw, start, len);
}
Token scan_number() {
const char* start = cursor_;
bool has_dot = false;
while (cursor_ < end_) {
char c = *cursor_;
if (c >= '0' && c <= '9') {
++cursor_;
} else if (c == '.' && !has_dot) {
has_dot = true;
++cursor_;
} else {
break;
}
}
uint32_t len = static_cast<uint32_t>(cursor_ - start);
return make_token(has_dot ? TokenType::TK_FLOAT : TokenType::TK_INTEGER,
start, len);
}
Token scan_single_quoted_string() {
++cursor_; // skip opening quote
const char* content_start = cursor_;
while (cursor_ < end_) {
if (*cursor_ == '\'') {
// Check for doubled single-quote escape ('')
if (cursor_ + 1 < end_ && *(cursor_ + 1) == '\'') {
cursor_ += 2; // skip both quotes
continue;
}
break; // end of string
}
if (*cursor_ == '\\') {
++cursor_; // skip escaped char
if (cursor_ < end_) ++cursor_;
} else {
++cursor_;
}
}
uint32_t len = static_cast<uint32_t>(cursor_ - content_start);
if (cursor_ < end_) ++cursor_; // skip closing quote
return make_token(TokenType::TK_STRING, content_start, len);
}
// MySQL: backtick-quoted identifier
// Unclosed backticks emit TK_ERROR so the parser can fail cleanly with
// ParseResult::ERROR -- otherwise the tokenizer would silently swallow
// the rest of the input as one giant identifier (e.g. `SET x = `foo`
// would become identifier `foo` followed by an unclosed-backtick scan
// that consumes everything to EOF as another identifier).
Token scan_backtick_identifier() {
const char* open_pos = cursor_;
++cursor_; // skip opening backtick
const char* content_start = cursor_;
while (cursor_ < end_ && *cursor_ != '`') ++cursor_;
if (cursor_ >= end_) {
return make_token(TokenType::TK_ERROR, open_pos, 1);
}
uint32_t len = static_cast<uint32_t>(cursor_ - content_start);
++cursor_; // skip closing backtick
return make_token(TokenType::TK_IDENTIFIER, content_start, len);
}
// PostgreSQL: double-quoted identifier
// Unclosed `"` emits TK_ERROR -- same rationale as scan_backtick_identifier
// above. `SET search_path = "unclosed_quote, public` would otherwise be
// treated as identifier `unclosed_quote, public` (commas, spaces and all),
// pass validation, and corrupt search_path with garbage.
Token scan_double_quoted_identifier() {
const char* open_pos = cursor_;
++cursor_; // skip opening quote
const char* content_start = cursor_;
while (cursor_ < end_ && *cursor_ != '"') ++cursor_;
if (cursor_ >= end_) {
return make_token(TokenType::TK_ERROR, open_pos, 1);
}
uint32_t len = static_cast<uint32_t>(cursor_ - content_start);
++cursor_; // skip closing quote
return make_token(TokenType::TK_IDENTIFIER, content_start, len);
}
// PostgreSQL: $$...$$ dollar-quoted string
Token scan_dollar_string() {
// We're at the first $. Simple form: $$content$$
cursor_ += 2; // skip opening $$
const char* content_start = cursor_;
while (cursor_ < end_) {
if (*cursor_ == '$' && peek_char(1) == '$') {
uint32_t len = static_cast<uint32_t>(cursor_ - content_start);
cursor_ += 2; // skip closing $$
return make_token(TokenType::TK_STRING, content_start, len);
}
++cursor_;
}
// Unterminated — return what we have
uint32_t len = static_cast<uint32_t>(cursor_ - content_start);
return make_token(TokenType::TK_STRING, content_start, len);
}
Token scan_token() {
skip_whitespace_and_comments();
if (cursor_ >= end_) {
return make_token(TokenType::TK_EOF, cursor_, 0);
}
char c = *cursor_;
// Identifiers and keywords
if ((c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || c == '_') {
return scan_identifier_or_keyword();
}
// Numbers
if (c >= '0' && c <= '9') {
return scan_number();
}
// Dot — could be start of .123 float or just dot
if (c == '.' && cursor_ + 1 < end_ &&
peek_char(1) >= '0' && peek_char(1) <= '9') {
return scan_number();
}
// String literals
if (c == '\'') return scan_single_quoted_string();
// MySQL: double-quoted strings; PostgreSQL: double-quoted identifiers
if (c == '"') {
if constexpr (D == Dialect::MySQL) {
// In MySQL, double quotes are strings (unless ANSI_QUOTES mode)
++cursor_;
const char* content_start = cursor_;
while (cursor_ < end_ && *cursor_ != '"') {
if (*cursor_ == '\\') { ++cursor_; if (cursor_ < end_) ++cursor_; }
else ++cursor_;
}
uint32_t len = static_cast<uint32_t>(cursor_ - content_start);
if (cursor_ < end_) ++cursor_;
return make_token(TokenType::TK_STRING, content_start, len);
} else {
return scan_double_quoted_identifier();
}
}
// Backtick identifier (MySQL only)
if constexpr (D == Dialect::MySQL) {
if (c == '`') return scan_backtick_identifier();
}
// @ and @@
if (c == '@') {
if (peek_char(1) == '@') {
const char* s = cursor_;
cursor_ += 2;
return make_token(TokenType::TK_DOUBLE_AT, s, 2);
}
const char* s = cursor_;
++cursor_;
return make_token(TokenType::TK_AT, s, 1);
}
// $ — PostgreSQL: $N placeholder or $$string$$
if constexpr (D == Dialect::PostgreSQL) {
if (c == '$') {
if (peek_char(1) == '$') {
return scan_dollar_string();
}
if (peek_char(1) >= '0' && peek_char(1) <= '9') {
const char* start = cursor_;
++cursor_; // skip $
while (cursor_ < end_ && *cursor_ >= '0' && *cursor_ <= '9')
++cursor_;
uint32_t len = static_cast<uint32_t>(cursor_ - start);
return make_token(TokenType::TK_DOLLAR_NUM, start, len);
}
// $<letter|underscore>... is NOT a valid PG token at this
// position -- it looks like a parameter placeholder but is
// syntactically invalid (placeholders must be numeric, e.g.
// $1). Emit TK_ERROR so the caller can fail cleanly with
// ParseResult::ERROR instead of returning PARTIAL with a
// null AST and confusing downstream consumers.
{
const char* start = cursor_;
++cursor_; // consume $ so error offset is precise
return make_token(TokenType::TK_ERROR, start, 1);
}
}
}
// Two-character operators
if (cursor_ + 1 < end_) {
char c2 = peek_char(1);
if (c == '<' && c2 == '=') { auto s = cursor_; cursor_ += 2; return make_token(TokenType::TK_LESS_EQUAL, s, 2); }
if (c == '>' && c2 == '=') { auto s = cursor_; cursor_ += 2; return make_token(TokenType::TK_GREATER_EQUAL, s, 2); }
if (c == '!' && c2 == '=') { auto s = cursor_; cursor_ += 2; return make_token(TokenType::TK_NOT_EQUAL, s, 2); }
if (c == '<' && c2 == '>') { auto s = cursor_; cursor_ += 2; return make_token(TokenType::TK_NOT_EQUAL, s, 2); }
if (c == '|' && c2 == '|') { auto s = cursor_; cursor_ += 2; return make_token(TokenType::TK_DOUBLE_PIPE, s, 2); }
if constexpr (D == Dialect::MySQL) {
if (c == ':' && c2 == '=') { auto s = cursor_; cursor_ += 2; return make_token(TokenType::TK_COLON_EQUAL, s, 2); }
}
if constexpr (D == Dialect::PostgreSQL) {
if (c == ':' && c2 == ':') { auto s = cursor_; cursor_ += 2; return make_token(TokenType::TK_DOUBLE_COLON, s, 2); }
}
}
// Single-character operators/punctuation
const char* s = cursor_;
++cursor_;
switch (c) {
case '(': return make_token(TokenType::TK_LPAREN, s, 1);
case ')': return make_token(TokenType::TK_RPAREN, s, 1);
case '[': return make_token(TokenType::TK_LBRACKET, s, 1);
case ']': return make_token(TokenType::TK_RBRACKET, s, 1);
case ',': return make_token(TokenType::TK_COMMA, s, 1);
case ';': return make_token(TokenType::TK_SEMICOLON, s, 1);
case '.': return make_token(TokenType::TK_DOT, s, 1);
case '*': return make_token(TokenType::TK_ASTERISK, s, 1);
case '+': return make_token(TokenType::TK_PLUS, s, 1);
case '-': return make_token(TokenType::TK_MINUS, s, 1);
case '/': return make_token(TokenType::TK_SLASH, s, 1);
case '%': return make_token(TokenType::TK_PERCENT, s, 1);
case '=': return make_token(TokenType::TK_EQUAL, s, 1);
case '<': return make_token(TokenType::TK_LESS, s, 1);
case '>': return make_token(TokenType::TK_GREATER, s, 1);
case '&': return make_token(TokenType::TK_AMPERSAND, s, 1);
case '|': return make_token(TokenType::TK_PIPE, s, 1);
case '^': return make_token(TokenType::TK_CARET, s, 1);
case '~': return make_token(TokenType::TK_TILDE, s, 1);
case '!': return make_token(TokenType::TK_EXCLAIM, s, 1);
case ':': return make_token(TokenType::TK_COLON, s, 1);
case '?': return make_token(TokenType::TK_QUESTION, s, 1);
case '#': return make_token(TokenType::TK_HASH, s, 1);
default: return make_token(TokenType::TK_ERROR, s, 1);
}
}
};
} // namespace sql_parser
#endif // SQL_PARSER_TOKENIZER_H
You can’t perform that action at this time.