qlist: add qlist_first()/qlist_next()
[qemu/mdroth.git] / json-lexer.c
blobc21338f66db75814e504ca86bf29c7b7487c3903
1 /*
2 * JSON lexer
4 * Copyright IBM, Corp. 2009
6 * Authors:
7 * Anthony Liguori <aliguori@us.ibm.com>
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.
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"
21 #define MAX_TOKEN_SIZE (64ULL << 20)
24 * \"([^\\\"]|(\\\"\\'\\\\\\/\\b\\f\\n\\r\\t\\u[0-9a-fA-F][0-9a-fA-F][0-9a-fA-F][0-9a-fA-F]))*\"
25 * '([^\\']|(\\\"\\'\\\\\\/\\b\\f\\n\\r\\t\\u[0-9a-fA-F][0-9a-fA-F][0-9a-fA-F][0-9a-fA-F]))*'
26 * 0|([1-9][0-9]*(.[0-9]+)?([eE]([-+])?[0-9]+))
27 * [{}\[\],:]
28 * [a-z]+
32 enum json_lexer_state {
33 IN_ERROR = 0,
34 IN_DQ_UCODE3,
35 IN_DQ_UCODE2,
36 IN_DQ_UCODE1,
37 IN_DQ_UCODE0,
38 IN_DQ_STRING_ESCAPE,
39 IN_DQ_STRING,
40 IN_SQ_UCODE3,
41 IN_SQ_UCODE2,
42 IN_SQ_UCODE1,
43 IN_SQ_UCODE0,
44 IN_SQ_STRING_ESCAPE,
45 IN_SQ_STRING,
46 IN_ZERO,
47 IN_DIGITS,
48 IN_DIGIT,
49 IN_EXP_E,
50 IN_MANTISSA,
51 IN_MANTISSA_DIGITS,
52 IN_NONZERO_NUMBER,
53 IN_NEG_NONZERO_NUMBER,
54 IN_KEYWORD,
55 IN_ESCAPE,
56 IN_ESCAPE_L,
57 IN_ESCAPE_LL,
58 IN_ESCAPE_I,
59 IN_ESCAPE_I6,
60 IN_ESCAPE_I64,
61 IN_WHITESPACE,
62 IN_START,
65 #define TERMINAL(state) [0 ... 0x7F] = (state)
67 /* Return whether TERMINAL is a terminal state and the transition to it
68 from OLD_STATE required lookahead. This happens whenever the table
69 below uses the TERMINAL macro. */
70 #define TERMINAL_NEEDED_LOOKAHEAD(old_state, terminal) \
71 (json_lexer[(old_state)][0] == (terminal))
73 static const uint8_t json_lexer[][256] = {
74 /* double quote string */
75 [IN_DQ_UCODE3] = {
76 ['0' ... '9'] = IN_DQ_STRING,
77 ['a' ... 'f'] = IN_DQ_STRING,
78 ['A' ... 'F'] = IN_DQ_STRING,
80 [IN_DQ_UCODE2] = {
81 ['0' ... '9'] = IN_DQ_UCODE3,
82 ['a' ... 'f'] = IN_DQ_UCODE3,
83 ['A' ... 'F'] = IN_DQ_UCODE3,
85 [IN_DQ_UCODE1] = {
86 ['0' ... '9'] = IN_DQ_UCODE2,
87 ['a' ... 'f'] = IN_DQ_UCODE2,
88 ['A' ... 'F'] = IN_DQ_UCODE2,
90 [IN_DQ_UCODE0] = {
91 ['0' ... '9'] = IN_DQ_UCODE1,
92 ['a' ... 'f'] = IN_DQ_UCODE1,
93 ['A' ... 'F'] = IN_DQ_UCODE1,
95 [IN_DQ_STRING_ESCAPE] = {
96 ['b'] = IN_DQ_STRING,
97 ['f'] = IN_DQ_STRING,
98 ['n'] = IN_DQ_STRING,
99 ['r'] = IN_DQ_STRING,
100 ['t'] = IN_DQ_STRING,
101 ['/'] = IN_DQ_STRING,
102 ['\\'] = IN_DQ_STRING,
103 ['\''] = IN_DQ_STRING,
104 ['\"'] = IN_DQ_STRING,
105 ['u'] = IN_DQ_UCODE0,
107 [IN_DQ_STRING] = {
108 [1 ... 0xBF] = IN_DQ_STRING,
109 [0xC2 ... 0xF4] = IN_DQ_STRING,
110 ['\\'] = IN_DQ_STRING_ESCAPE,
111 ['"'] = JSON_STRING,
114 /* single quote string */
115 [IN_SQ_UCODE3] = {
116 ['0' ... '9'] = IN_SQ_STRING,
117 ['a' ... 'f'] = IN_SQ_STRING,
118 ['A' ... 'F'] = IN_SQ_STRING,
120 [IN_SQ_UCODE2] = {
121 ['0' ... '9'] = IN_SQ_UCODE3,
122 ['a' ... 'f'] = IN_SQ_UCODE3,
123 ['A' ... 'F'] = IN_SQ_UCODE3,
125 [IN_SQ_UCODE1] = {
126 ['0' ... '9'] = IN_SQ_UCODE2,
127 ['a' ... 'f'] = IN_SQ_UCODE2,
128 ['A' ... 'F'] = IN_SQ_UCODE2,
130 [IN_SQ_UCODE0] = {
131 ['0' ... '9'] = IN_SQ_UCODE1,
132 ['a' ... 'f'] = IN_SQ_UCODE1,
133 ['A' ... 'F'] = IN_SQ_UCODE1,
135 [IN_SQ_STRING_ESCAPE] = {
136 ['b'] = IN_SQ_STRING,
137 ['f'] = IN_SQ_STRING,
138 ['n'] = IN_SQ_STRING,
139 ['r'] = IN_SQ_STRING,
140 ['t'] = IN_SQ_STRING,
141 ['/'] = IN_DQ_STRING,
142 ['\\'] = IN_DQ_STRING,
143 ['\''] = IN_SQ_STRING,
144 ['\"'] = IN_SQ_STRING,
145 ['u'] = IN_SQ_UCODE0,
147 [IN_SQ_STRING] = {
148 [1 ... 0xBF] = IN_SQ_STRING,
149 [0xC2 ... 0xF4] = IN_SQ_STRING,
150 ['\\'] = IN_SQ_STRING_ESCAPE,
151 ['\''] = JSON_STRING,
154 /* Zero */
155 [IN_ZERO] = {
156 TERMINAL(JSON_INTEGER),
157 ['0' ... '9'] = IN_ERROR,
158 ['.'] = IN_MANTISSA,
161 /* Float */
162 [IN_DIGITS] = {
163 TERMINAL(JSON_FLOAT),
164 ['0' ... '9'] = IN_DIGITS,
167 [IN_DIGIT] = {
168 ['0' ... '9'] = IN_DIGITS,
171 [IN_EXP_E] = {
172 ['-'] = IN_DIGIT,
173 ['+'] = IN_DIGIT,
174 ['0' ... '9'] = IN_DIGITS,
177 [IN_MANTISSA_DIGITS] = {
178 TERMINAL(JSON_FLOAT),
179 ['0' ... '9'] = IN_MANTISSA_DIGITS,
180 ['e'] = IN_EXP_E,
181 ['E'] = IN_EXP_E,
184 [IN_MANTISSA] = {
185 ['0' ... '9'] = IN_MANTISSA_DIGITS,
188 /* Number */
189 [IN_NONZERO_NUMBER] = {
190 TERMINAL(JSON_INTEGER),
191 ['0' ... '9'] = IN_NONZERO_NUMBER,
192 ['e'] = IN_EXP_E,
193 ['E'] = IN_EXP_E,
194 ['.'] = IN_MANTISSA,
197 [IN_NEG_NONZERO_NUMBER] = {
198 ['0'] = IN_ZERO,
199 ['1' ... '9'] = IN_NONZERO_NUMBER,
202 /* keywords */
203 [IN_KEYWORD] = {
204 TERMINAL(JSON_KEYWORD),
205 ['a' ... 'z'] = IN_KEYWORD,
208 /* whitespace */
209 [IN_WHITESPACE] = {
210 TERMINAL(JSON_SKIP),
211 [' '] = IN_WHITESPACE,
212 ['\t'] = IN_WHITESPACE,
213 ['\r'] = IN_WHITESPACE,
214 ['\n'] = IN_WHITESPACE,
217 /* escape */
218 [IN_ESCAPE_LL] = {
219 ['d'] = JSON_ESCAPE,
222 [IN_ESCAPE_L] = {
223 ['d'] = JSON_ESCAPE,
224 ['l'] = IN_ESCAPE_LL,
227 [IN_ESCAPE_I64] = {
228 ['d'] = JSON_ESCAPE,
231 [IN_ESCAPE_I6] = {
232 ['4'] = IN_ESCAPE_I64,
235 [IN_ESCAPE_I] = {
236 ['6'] = IN_ESCAPE_I6,
239 [IN_ESCAPE] = {
240 ['d'] = JSON_ESCAPE,
241 ['i'] = JSON_ESCAPE,
242 ['p'] = JSON_ESCAPE,
243 ['s'] = JSON_ESCAPE,
244 ['f'] = JSON_ESCAPE,
245 ['l'] = IN_ESCAPE_L,
246 ['I'] = IN_ESCAPE_I,
249 /* top level rule */
250 [IN_START] = {
251 ['"'] = IN_DQ_STRING,
252 ['\''] = IN_SQ_STRING,
253 ['0'] = IN_ZERO,
254 ['1' ... '9'] = IN_NONZERO_NUMBER,
255 ['-'] = IN_NEG_NONZERO_NUMBER,
256 ['{'] = JSON_OPERATOR,
257 ['}'] = JSON_OPERATOR,
258 ['['] = JSON_OPERATOR,
259 [']'] = JSON_OPERATOR,
260 [','] = JSON_OPERATOR,
261 [':'] = JSON_OPERATOR,
262 ['a' ... 'z'] = IN_KEYWORD,
263 ['%'] = IN_ESCAPE,
264 [' '] = IN_WHITESPACE,
265 ['\t'] = IN_WHITESPACE,
266 ['\r'] = IN_WHITESPACE,
267 ['\n'] = IN_WHITESPACE,
271 void json_lexer_init(JSONLexer *lexer, JSONLexerEmitter func)
273 lexer->emit = func;
274 lexer->state = IN_START;
275 lexer->token = qstring_new();
276 lexer->x = lexer->y = 0;
279 static int json_lexer_feed_char(JSONLexer *lexer, char ch, bool flush)
281 int char_consumed, new_state;
283 lexer->x++;
284 if (ch == '\n') {
285 lexer->x = 0;
286 lexer->y++;
289 do {
290 new_state = json_lexer[lexer->state][(uint8_t)ch];
291 char_consumed = !TERMINAL_NEEDED_LOOKAHEAD(lexer->state, new_state);
292 if (char_consumed) {
293 qstring_append_chr(lexer->token, ch);
296 switch (new_state) {
297 case JSON_OPERATOR:
298 case JSON_ESCAPE:
299 case JSON_INTEGER:
300 case JSON_FLOAT:
301 case JSON_KEYWORD:
302 case JSON_STRING:
303 lexer->emit(lexer, lexer->token, new_state, lexer->x, lexer->y);
304 case JSON_SKIP:
305 QDECREF(lexer->token);
306 lexer->token = qstring_new();
307 new_state = IN_START;
308 break;
309 case IN_ERROR:
310 /* XXX: To avoid having previous bad input leaving the parser in an
311 * unresponsive state where we consume unpredictable amounts of
312 * subsequent "good" input, percolate this error state up to the
313 * tokenizer/parser by forcing a NULL object to be emitted, then
314 * reset state.
316 * Also note that this handling is required for reliable channel
317 * negotiation between QMP and the guest agent, since chr(0xFF)
318 * is placed at the beginning of certain events to ensure proper
319 * delivery when the channel is in an unknown state. chr(0xFF) is
320 * never a valid ASCII/UTF-8 sequence, so this should reliably
321 * induce an error/flush state.
323 lexer->emit(lexer, lexer->token, JSON_ERROR, lexer->x, lexer->y);
324 QDECREF(lexer->token);
325 lexer->token = qstring_new();
326 new_state = IN_START;
327 lexer->state = new_state;
328 return 0;
329 default:
330 break;
332 lexer->state = new_state;
333 } while (!char_consumed && !flush);
335 /* Do not let a single token grow to an arbitrarily large size,
336 * this is a security consideration.
338 if (lexer->token->length > MAX_TOKEN_SIZE) {
339 lexer->emit(lexer, lexer->token, lexer->state, lexer->x, lexer->y);
340 QDECREF(lexer->token);
341 lexer->token = qstring_new();
342 lexer->state = IN_START;
345 return 0;
348 int json_lexer_feed(JSONLexer *lexer, const char *buffer, size_t size)
350 size_t i;
352 for (i = 0; i < size; i++) {
353 int err;
355 err = json_lexer_feed_char(lexer, buffer[i], false);
356 if (err < 0) {
357 return err;
361 return 0;
364 int json_lexer_flush(JSONLexer *lexer)
366 return lexer->state == IN_START ? 0 : json_lexer_feed_char(lexer, 0, true);
369 void json_lexer_destroy(JSONLexer *lexer)
371 QDECREF(lexer->token);