4 * Copyright IBM, Corp. 2009
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.
18 #include "qemu-common.h"
19 #include "json-lexer.h"
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]+))
30 enum json_lexer_state {
52 IN_NEG_NONZERO_NUMBER,
66 #define TERMINAL(state) [0 ... 0x7F] = (state)
68 static const uint8_t json_lexer[][256] = {
70 TERMINAL(JSON_STRING),
73 /* double quote string */
75 ['0' ... '9'] = IN_DQ_STRING,
76 ['a' ... 'f'] = IN_DQ_STRING,
77 ['A' ... 'F'] = IN_DQ_STRING,
80 ['0' ... '9'] = IN_DQ_UCODE3,
81 ['a' ... 'f'] = IN_DQ_UCODE3,
82 ['A' ... 'F'] = IN_DQ_UCODE3,
85 ['0' ... '9'] = IN_DQ_UCODE2,
86 ['a' ... 'f'] = IN_DQ_UCODE2,
87 ['A' ... 'F'] = IN_DQ_UCODE2,
90 ['0' ... '9'] = IN_DQ_UCODE1,
91 ['a' ... 'f'] = IN_DQ_UCODE1,
92 ['A' ... 'F'] = IN_DQ_UCODE1,
94 [IN_DQ_STRING_ESCAPE] = {
100 ['\''] = IN_DQ_STRING,
101 ['\"'] = IN_DQ_STRING,
102 ['u'] = IN_DQ_UCODE0,
105 [1 ... 0xFF] = IN_DQ_STRING,
106 ['\\'] = IN_DQ_STRING_ESCAPE,
107 ['"'] = IN_DONE_STRING,
110 /* single quote string */
112 ['0' ... '9'] = IN_SQ_STRING,
113 ['a' ... 'f'] = IN_SQ_STRING,
114 ['A' ... 'F'] = IN_SQ_STRING,
117 ['0' ... '9'] = IN_SQ_UCODE3,
118 ['a' ... 'f'] = IN_SQ_UCODE3,
119 ['A' ... 'F'] = IN_SQ_UCODE3,
122 ['0' ... '9'] = IN_SQ_UCODE2,
123 ['a' ... 'f'] = IN_SQ_UCODE2,
124 ['A' ... 'F'] = IN_SQ_UCODE2,
127 ['0' ... '9'] = IN_SQ_UCODE1,
128 ['a' ... 'f'] = IN_SQ_UCODE1,
129 ['A' ... 'F'] = IN_SQ_UCODE1,
131 [IN_SQ_STRING_ESCAPE] = {
132 ['b'] = IN_SQ_STRING,
133 ['f'] = IN_SQ_STRING,
134 ['n'] = IN_SQ_STRING,
135 ['r'] = IN_SQ_STRING,
136 ['t'] = IN_SQ_STRING,
137 ['\''] = IN_SQ_STRING,
138 ['\"'] = IN_SQ_STRING,
139 ['u'] = IN_SQ_UCODE0,
142 [1 ... 0xFF] = IN_SQ_STRING,
143 ['\\'] = IN_SQ_STRING_ESCAPE,
144 ['\''] = IN_DONE_STRING,
149 TERMINAL(JSON_INTEGER),
150 ['0' ... '9'] = ERROR,
156 TERMINAL(JSON_FLOAT),
157 ['0' ... '9'] = IN_DIGITS,
161 ['0' ... '9'] = IN_DIGITS,
167 ['0' ... '9'] = IN_DIGITS,
170 [IN_MANTISSA_DIGITS] = {
171 TERMINAL(JSON_FLOAT),
172 ['0' ... '9'] = IN_MANTISSA_DIGITS,
178 ['0' ... '9'] = IN_MANTISSA_DIGITS,
182 [IN_NONZERO_NUMBER] = {
183 TERMINAL(JSON_INTEGER),
184 ['0' ... '9'] = IN_NONZERO_NUMBER,
190 [IN_NEG_NONZERO_NUMBER] = {
192 ['1' ... '9'] = IN_NONZERO_NUMBER,
197 TERMINAL(JSON_KEYWORD),
198 ['a' ... 'z'] = IN_KEYWORD,
204 [' '] = IN_WHITESPACE,
205 ['\t'] = IN_WHITESPACE,
206 ['\r'] = IN_WHITESPACE,
207 ['\n'] = IN_WHITESPACE,
211 [IN_OPERATOR_DONE] = {
212 TERMINAL(JSON_OPERATOR),
217 TERMINAL(JSON_ESCAPE),
221 ['d'] = IN_ESCAPE_DONE,
225 ['d'] = IN_ESCAPE_DONE,
226 ['l'] = IN_ESCAPE_LL,
230 ['d'] = IN_ESCAPE_DONE,
234 ['4'] = IN_ESCAPE_I64,
238 ['6'] = IN_ESCAPE_I6,
242 ['d'] = IN_ESCAPE_DONE,
243 ['i'] = IN_ESCAPE_DONE,
244 ['p'] = IN_ESCAPE_DONE,
245 ['s'] = IN_ESCAPE_DONE,
246 ['f'] = IN_ESCAPE_DONE,
253 ['"'] = IN_DQ_STRING,
254 ['\''] = IN_SQ_STRING,
256 ['1' ... '9'] = IN_NONZERO_NUMBER,
257 ['-'] = IN_NEG_NONZERO_NUMBER,
258 ['{'] = IN_OPERATOR_DONE,
259 ['}'] = IN_OPERATOR_DONE,
260 ['['] = IN_OPERATOR_DONE,
261 [']'] = IN_OPERATOR_DONE,
262 [','] = IN_OPERATOR_DONE,
263 [':'] = IN_OPERATOR_DONE,
264 ['a' ... 'z'] = IN_KEYWORD,
266 [' '] = IN_WHITESPACE,
267 ['\t'] = IN_WHITESPACE,
268 ['\r'] = IN_WHITESPACE,
269 ['\n'] = IN_WHITESPACE,
273 void json_lexer_init(JSONLexer *lexer, JSONLexerEmitter func)
276 lexer->state = IN_START;
277 lexer->token = qstring_new();
280 static int json_lexer_feed_char(JSONLexer *lexer, char ch)
290 lexer->state = json_lexer[lexer->state][(uint8_t)ch];
292 switch (lexer->state) {
299 lexer->emit(lexer, lexer->token, lexer->state, lexer->x, lexer->y);
301 lexer->state = json_lexer[IN_START][(uint8_t)ch];
302 QDECREF(lexer->token);
303 lexer->token = qstring_new();
314 qstring_append(lexer->token, buf);
319 int json_lexer_feed(JSONLexer *lexer, const char *buffer, size_t size)
323 for (i = 0; i < size; i++) {
326 err = json_lexer_feed_char(lexer, buffer[i]);
335 int json_lexer_flush(JSONLexer *lexer)
337 return json_lexer_feed_char(lexer, 0);
340 void json_lexer_destroy(JSONLexer *lexer)
342 QDECREF(lexer->token);