]>
Commit | Line | Data |
---|---|---|
1 | /* | |
2 | * JSON lexer | |
3 | * | |
4 | * Copyright IBM, Corp. 2009 | |
5 | * | |
6 | * Authors: | |
7 | * Anthony Liguori <aliguori@us.ibm.com> | |
8 | * | |
9 | * This work is licensed under the terms of the GNU LGPL, version 2.1 or later. | |
10 | * See the COPYING.LIB file in the top-level directory. | |
11 | * | |
12 | */ | |
13 | ||
14 | #include "qstring.h" | |
15 | #include "qlist.h" | |
16 | #include "qdict.h" | |
17 | #include "qint.h" | |
18 | #include "qemu-common.h" | |
19 | #include "json-lexer.h" | |
20 | ||
21 | /* | |
22 | * \"([^\\\"]|(\\\"\\'\\\\\\/\\b\\f\\n\\r\\t\\u[0-9a-fA-F][0-9a-fA-F][0-9a-fA-F][0-9a-fA-F]))*\" | |
23 | * '([^\\']|(\\\"\\'\\\\\\/\\b\\f\\n\\r\\t\\u[0-9a-fA-F][0-9a-fA-F][0-9a-fA-F][0-9a-fA-F]))*' | |
24 | * 0|([1-9][0-9]*(.[0-9]+)?([eE]([-+])?[0-9]+)) | |
25 | * [{}\[\],:] | |
26 | * [a-z]+ | |
27 | * | |
28 | */ | |
29 | ||
30 | enum json_lexer_state { | |
31 | ERROR = 0, | |
32 | IN_DONE_STRING, | |
33 | IN_DQ_UCODE3, | |
34 | IN_DQ_UCODE2, | |
35 | IN_DQ_UCODE1, | |
36 | IN_DQ_UCODE0, | |
37 | IN_DQ_STRING_ESCAPE, | |
38 | IN_DQ_STRING, | |
39 | IN_SQ_UCODE3, | |
40 | IN_SQ_UCODE2, | |
41 | IN_SQ_UCODE1, | |
42 | IN_SQ_UCODE0, | |
43 | IN_SQ_STRING_ESCAPE, | |
44 | IN_SQ_STRING, | |
45 | IN_ZERO, | |
46 | IN_DIGITS, | |
47 | IN_DIGIT, | |
48 | IN_EXP_E, | |
49 | IN_MANTISSA, | |
50 | IN_MANTISSA_DIGITS, | |
51 | IN_NONZERO_NUMBER, | |
52 | IN_NEG_NONZERO_NUMBER, | |
53 | IN_KEYWORD, | |
54 | IN_ESCAPE, | |
55 | IN_ESCAPE_L, | |
56 | IN_ESCAPE_LL, | |
57 | IN_ESCAPE_I, | |
58 | IN_ESCAPE_I6, | |
59 | IN_ESCAPE_I64, | |
60 | IN_ESCAPE_DONE, | |
61 | IN_WHITESPACE, | |
62 | IN_OPERATOR_DONE, | |
63 | IN_START, | |
64 | }; | |
65 | ||
66 | #define TERMINAL(state) [0 ... 0x7F] = (state) | |
67 | ||
68 | static const uint8_t json_lexer[][256] = { | |
69 | [IN_DONE_STRING] = { | |
70 | TERMINAL(JSON_STRING), | |
71 | }, | |
72 | ||
73 | /* double quote string */ | |
74 | [IN_DQ_UCODE3] = { | |
75 | ['0' ... '9'] = IN_DQ_STRING, | |
76 | ['a' ... 'f'] = IN_DQ_STRING, | |
77 | ['A' ... 'F'] = IN_DQ_STRING, | |
78 | }, | |
79 | [IN_DQ_UCODE2] = { | |
80 | ['0' ... '9'] = IN_DQ_UCODE3, | |
81 | ['a' ... 'f'] = IN_DQ_UCODE3, | |
82 | ['A' ... 'F'] = IN_DQ_UCODE3, | |
83 | }, | |
84 | [IN_DQ_UCODE1] = { | |
85 | ['0' ... '9'] = IN_DQ_UCODE2, | |
86 | ['a' ... 'f'] = IN_DQ_UCODE2, | |
87 | ['A' ... 'F'] = IN_DQ_UCODE2, | |
88 | }, | |
89 | [IN_DQ_UCODE0] = { | |
90 | ['0' ... '9'] = IN_DQ_UCODE1, | |
91 | ['a' ... 'f'] = IN_DQ_UCODE1, | |
92 | ['A' ... 'F'] = IN_DQ_UCODE1, | |
93 | }, | |
94 | [IN_DQ_STRING_ESCAPE] = { | |
95 | ['b'] = IN_DQ_STRING, | |
96 | ['f'] = IN_DQ_STRING, | |
97 | ['n'] = IN_DQ_STRING, | |
98 | ['r'] = IN_DQ_STRING, | |
99 | ['t'] = IN_DQ_STRING, | |
100 | ['/'] = IN_DQ_STRING, | |
101 | ['\\'] = IN_DQ_STRING, | |
102 | ['\''] = IN_DQ_STRING, | |
103 | ['\"'] = IN_DQ_STRING, | |
104 | ['u'] = IN_DQ_UCODE0, | |
105 | }, | |
106 | [IN_DQ_STRING] = { | |
107 | [1 ... 0xFF] = IN_DQ_STRING, | |
108 | ['\\'] = IN_DQ_STRING_ESCAPE, | |
109 | ['"'] = IN_DONE_STRING, | |
110 | }, | |
111 | ||
112 | /* single quote string */ | |
113 | [IN_SQ_UCODE3] = { | |
114 | ['0' ... '9'] = IN_SQ_STRING, | |
115 | ['a' ... 'f'] = IN_SQ_STRING, | |
116 | ['A' ... 'F'] = IN_SQ_STRING, | |
117 | }, | |
118 | [IN_SQ_UCODE2] = { | |
119 | ['0' ... '9'] = IN_SQ_UCODE3, | |
120 | ['a' ... 'f'] = IN_SQ_UCODE3, | |
121 | ['A' ... 'F'] = IN_SQ_UCODE3, | |
122 | }, | |
123 | [IN_SQ_UCODE1] = { | |
124 | ['0' ... '9'] = IN_SQ_UCODE2, | |
125 | ['a' ... 'f'] = IN_SQ_UCODE2, | |
126 | ['A' ... 'F'] = IN_SQ_UCODE2, | |
127 | }, | |
128 | [IN_SQ_UCODE0] = { | |
129 | ['0' ... '9'] = IN_SQ_UCODE1, | |
130 | ['a' ... 'f'] = IN_SQ_UCODE1, | |
131 | ['A' ... 'F'] = IN_SQ_UCODE1, | |
132 | }, | |
133 | [IN_SQ_STRING_ESCAPE] = { | |
134 | ['b'] = IN_SQ_STRING, | |
135 | ['f'] = IN_SQ_STRING, | |
136 | ['n'] = IN_SQ_STRING, | |
137 | ['r'] = IN_SQ_STRING, | |
138 | ['t'] = IN_SQ_STRING, | |
139 | ['/'] = IN_DQ_STRING, | |
140 | ['\\'] = IN_DQ_STRING, | |
141 | ['\''] = IN_SQ_STRING, | |
142 | ['\"'] = IN_SQ_STRING, | |
143 | ['u'] = IN_SQ_UCODE0, | |
144 | }, | |
145 | [IN_SQ_STRING] = { | |
146 | [1 ... 0xFF] = IN_SQ_STRING, | |
147 | ['\\'] = IN_SQ_STRING_ESCAPE, | |
148 | ['\''] = IN_DONE_STRING, | |
149 | }, | |
150 | ||
151 | /* Zero */ | |
152 | [IN_ZERO] = { | |
153 | TERMINAL(JSON_INTEGER), | |
154 | ['0' ... '9'] = ERROR, | |
155 | ['.'] = IN_MANTISSA, | |
156 | }, | |
157 | ||
158 | /* Float */ | |
159 | [IN_DIGITS] = { | |
160 | TERMINAL(JSON_FLOAT), | |
161 | ['0' ... '9'] = IN_DIGITS, | |
162 | }, | |
163 | ||
164 | [IN_DIGIT] = { | |
165 | ['0' ... '9'] = IN_DIGITS, | |
166 | }, | |
167 | ||
168 | [IN_EXP_E] = { | |
169 | ['-'] = IN_DIGIT, | |
170 | ['+'] = IN_DIGIT, | |
171 | ['0' ... '9'] = IN_DIGITS, | |
172 | }, | |
173 | ||
174 | [IN_MANTISSA_DIGITS] = { | |
175 | TERMINAL(JSON_FLOAT), | |
176 | ['0' ... '9'] = IN_MANTISSA_DIGITS, | |
177 | ['e'] = IN_EXP_E, | |
178 | ['E'] = IN_EXP_E, | |
179 | }, | |
180 | ||
181 | [IN_MANTISSA] = { | |
182 | ['0' ... '9'] = IN_MANTISSA_DIGITS, | |
183 | }, | |
184 | ||
185 | /* Number */ | |
186 | [IN_NONZERO_NUMBER] = { | |
187 | TERMINAL(JSON_INTEGER), | |
188 | ['0' ... '9'] = IN_NONZERO_NUMBER, | |
189 | ['e'] = IN_EXP_E, | |
190 | ['E'] = IN_EXP_E, | |
191 | ['.'] = IN_MANTISSA, | |
192 | }, | |
193 | ||
194 | [IN_NEG_NONZERO_NUMBER] = { | |
195 | ['0'] = IN_ZERO, | |
196 | ['1' ... '9'] = IN_NONZERO_NUMBER, | |
197 | }, | |
198 | ||
199 | /* keywords */ | |
200 | [IN_KEYWORD] = { | |
201 | TERMINAL(JSON_KEYWORD), | |
202 | ['a' ... 'z'] = IN_KEYWORD, | |
203 | }, | |
204 | ||
205 | /* whitespace */ | |
206 | [IN_WHITESPACE] = { | |
207 | TERMINAL(JSON_SKIP), | |
208 | [' '] = IN_WHITESPACE, | |
209 | ['\t'] = IN_WHITESPACE, | |
210 | ['\r'] = IN_WHITESPACE, | |
211 | ['\n'] = IN_WHITESPACE, | |
212 | }, | |
213 | ||
214 | /* operator */ | |
215 | [IN_OPERATOR_DONE] = { | |
216 | TERMINAL(JSON_OPERATOR), | |
217 | }, | |
218 | ||
219 | /* escape */ | |
220 | [IN_ESCAPE_DONE] = { | |
221 | TERMINAL(JSON_ESCAPE), | |
222 | }, | |
223 | ||
224 | [IN_ESCAPE_LL] = { | |
225 | ['d'] = IN_ESCAPE_DONE, | |
226 | }, | |
227 | ||
228 | [IN_ESCAPE_L] = { | |
229 | ['d'] = IN_ESCAPE_DONE, | |
230 | ['l'] = IN_ESCAPE_LL, | |
231 | }, | |
232 | ||
233 | [IN_ESCAPE_I64] = { | |
234 | ['d'] = IN_ESCAPE_DONE, | |
235 | }, | |
236 | ||
237 | [IN_ESCAPE_I6] = { | |
238 | ['4'] = IN_ESCAPE_I64, | |
239 | }, | |
240 | ||
241 | [IN_ESCAPE_I] = { | |
242 | ['6'] = IN_ESCAPE_I6, | |
243 | }, | |
244 | ||
245 | [IN_ESCAPE] = { | |
246 | ['d'] = IN_ESCAPE_DONE, | |
247 | ['i'] = IN_ESCAPE_DONE, | |
248 | ['p'] = IN_ESCAPE_DONE, | |
249 | ['s'] = IN_ESCAPE_DONE, | |
250 | ['f'] = IN_ESCAPE_DONE, | |
251 | ['l'] = IN_ESCAPE_L, | |
252 | ['I'] = IN_ESCAPE_I, | |
253 | }, | |
254 | ||
255 | /* top level rule */ | |
256 | [IN_START] = { | |
257 | ['"'] = IN_DQ_STRING, | |
258 | ['\''] = IN_SQ_STRING, | |
259 | ['0'] = IN_ZERO, | |
260 | ['1' ... '9'] = IN_NONZERO_NUMBER, | |
261 | ['-'] = IN_NEG_NONZERO_NUMBER, | |
262 | ['{'] = IN_OPERATOR_DONE, | |
263 | ['}'] = IN_OPERATOR_DONE, | |
264 | ['['] = IN_OPERATOR_DONE, | |
265 | [']'] = IN_OPERATOR_DONE, | |
266 | [','] = IN_OPERATOR_DONE, | |
267 | [':'] = IN_OPERATOR_DONE, | |
268 | ['a' ... 'z'] = IN_KEYWORD, | |
269 | ['%'] = IN_ESCAPE, | |
270 | [' '] = IN_WHITESPACE, | |
271 | ['\t'] = IN_WHITESPACE, | |
272 | ['\r'] = IN_WHITESPACE, | |
273 | ['\n'] = IN_WHITESPACE, | |
274 | }, | |
275 | }; | |
276 | ||
277 | void json_lexer_init(JSONLexer *lexer, JSONLexerEmitter func) | |
278 | { | |
279 | lexer->emit = func; | |
280 | lexer->state = IN_START; | |
281 | lexer->token = qstring_new(); | |
282 | lexer->x = lexer->y = 0; | |
283 | } | |
284 | ||
285 | static int json_lexer_feed_char(JSONLexer *lexer, char ch) | |
286 | { | |
287 | lexer->x++; | |
288 | if (ch == '\n') { | |
289 | lexer->x = 0; | |
290 | lexer->y++; | |
291 | } | |
292 | ||
293 | lexer->state = json_lexer[lexer->state][(uint8_t)ch]; | |
294 | ||
295 | switch (lexer->state) { | |
296 | case JSON_OPERATOR: | |
297 | case JSON_ESCAPE: | |
298 | case JSON_INTEGER: | |
299 | case JSON_FLOAT: | |
300 | case JSON_KEYWORD: | |
301 | case JSON_STRING: | |
302 | lexer->emit(lexer, lexer->token, lexer->state, lexer->x, lexer->y); | |
303 | case JSON_SKIP: | |
304 | lexer->state = json_lexer[IN_START][(uint8_t)ch]; | |
305 | QDECREF(lexer->token); | |
306 | lexer->token = qstring_new(); | |
307 | break; | |
308 | case ERROR: | |
309 | return -EINVAL; | |
310 | default: | |
311 | break; | |
312 | } | |
313 | ||
314 | qstring_append_chr(lexer->token, ch); | |
315 | ||
316 | return 0; | |
317 | } | |
318 | ||
319 | int json_lexer_feed(JSONLexer *lexer, const char *buffer, size_t size) | |
320 | { | |
321 | size_t i; | |
322 | ||
323 | for (i = 0; i < size; i++) { | |
324 | int err; | |
325 | ||
326 | err = json_lexer_feed_char(lexer, buffer[i]); | |
327 | if (err < 0) { | |
328 | return err; | |
329 | } | |
330 | } | |
331 | ||
332 | return 0; | |
333 | } | |
334 | ||
335 | int json_lexer_flush(JSONLexer *lexer) | |
336 | { | |
337 | return json_lexer_feed_char(lexer, 0); | |
338 | } | |
339 | ||
340 | void json_lexer_destroy(JSONLexer *lexer) | |
341 | { | |
342 | QDECREF(lexer->token); | |
343 | } |