json-lexer.c

   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_DONE,
  58     IN_WHITESPACE,
  59     IN_OPERATOR_DONE,
  60     IN_START,
  61 };
  62
  63 #define TERMINAL(state) [0 ... 0x7F] = (state)
  64
  65 static const uint8_t json_lexer[][256] =  {
  66     [IN_DONE_STRING] = {
  67         TERMINAL(JSON_STRING),
  68     },
  69
  70     /* double quote string */
  71     [IN_DQ_UCODE3] = {
  72         ['0' ... '9'] = IN_DQ_STRING,
  73         ['a' ... 'f'] = IN_DQ_STRING,
  74         ['A' ... 'F'] = IN_DQ_STRING,
  75     },
  76     [IN_DQ_UCODE2] = {
  77         ['0' ... '9'] = IN_DQ_UCODE3,
  78         ['a' ... 'f'] = IN_DQ_UCODE3,
  79         ['A' ... 'F'] = IN_DQ_UCODE3,
  80     },
  81     [IN_DQ_UCODE1] = {
  82         ['0' ... '9'] = IN_DQ_UCODE2,
  83         ['a' ... 'f'] = IN_DQ_UCODE2,
  84         ['A' ... 'F'] = IN_DQ_UCODE2,
  85     },
  86     [IN_DQ_UCODE0] = {
  87         ['0' ... '9'] = IN_DQ_UCODE1,
  88         ['a' ... 'f'] = IN_DQ_UCODE1,
  89         ['A' ... 'F'] = IN_DQ_UCODE1,
  90     },
  91     [IN_DQ_STRING_ESCAPE] = {
  92         ['b'] = IN_DQ_STRING,
  93         ['f'] =  IN_DQ_STRING,
  94         ['n'] =  IN_DQ_STRING,
  95         ['r'] =  IN_DQ_STRING,
  96         ['t'] =  IN_DQ_STRING,
  97         ['\''] = IN_DQ_STRING,
  98         ['\"'] = IN_DQ_STRING,
  99         ['u'] = IN_DQ_UCODE0,
 100     },
 101     [IN_DQ_STRING] = {
 102         [1 ... 0xFF] = IN_DQ_STRING,
 103         ['\\'] = IN_DQ_STRING_ESCAPE,
 104         ['"'] = IN_DONE_STRING,
 105     },
 106
 107     /* single quote string */
 108     [IN_SQ_UCODE3] = {
 109         ['0' ... '9'] = IN_SQ_STRING,
 110         ['a' ... 'f'] = IN_SQ_STRING,
 111         ['A' ... 'F'] = IN_SQ_STRING,
 112     },
 113     [IN_SQ_UCODE2] = {
 114         ['0' ... '9'] = IN_SQ_UCODE3,
 115         ['a' ... 'f'] = IN_SQ_UCODE3,
 116         ['A' ... 'F'] = IN_SQ_UCODE3,
 117     },
 118     [IN_SQ_UCODE1] = {
 119         ['0' ... '9'] = IN_SQ_UCODE2,
 120         ['a' ... 'f'] = IN_SQ_UCODE2,
 121         ['A' ... 'F'] = IN_SQ_UCODE2,
 122     },
 123     [IN_SQ_UCODE0] = {
 124         ['0' ... '9'] = IN_SQ_UCODE1,
 125         ['a' ... 'f'] = IN_SQ_UCODE1,
 126         ['A' ... 'F'] = IN_SQ_UCODE1,
 127     },
 128     [IN_SQ_STRING_ESCAPE] = {
 129         ['b'] = IN_SQ_STRING,
 130         ['f'] =  IN_SQ_STRING,
 131         ['n'] =  IN_SQ_STRING,
 132         ['r'] =  IN_SQ_STRING,
 133         ['t'] =  IN_SQ_STRING,
 134         ['\''] = IN_SQ_STRING,
 135         ['\"'] = IN_SQ_STRING,
 136         ['u'] = IN_SQ_UCODE0,
 137     },
 138     [IN_SQ_STRING] = {
 139         [1 ... 0xFF] = IN_SQ_STRING,
 140         ['\\'] = IN_SQ_STRING_ESCAPE,
 141         ['\''] = IN_DONE_STRING,
 142     },
 143
 144     /* Zero */
 145     [IN_ZERO] = {
 146         TERMINAL(JSON_INTEGER),
 147         ['0' ... '9'] = ERROR,
 148         ['.'] = IN_MANTISSA,
 149     },
 150
 151     /* Float */
 152     [IN_DIGITS] = {
 153         TERMINAL(JSON_FLOAT),
 154         ['0' ... '9'] = IN_DIGITS,
 155     },
 156
 157     [IN_DIGIT] = {
 158         ['0' ... '9'] = IN_DIGITS,
 159     },
 160
 161     [IN_EXP_E] = {
 162         ['-'] = IN_DIGIT,
 163         ['+'] = IN_DIGIT,
 164         ['0' ... '9'] = IN_DIGITS,
 165     },
 166
 167     [IN_MANTISSA_DIGITS] = {
 168         TERMINAL(JSON_FLOAT),
 169         ['0' ... '9'] = IN_MANTISSA_DIGITS,
 170         ['e'] = IN_EXP_E,
 171         ['E'] = IN_EXP_E,
 172     },
 173
 174     [IN_MANTISSA] = {
 175         ['0' ... '9'] = IN_MANTISSA_DIGITS,
 176     },
 177
 178     /* Number */
 179     [IN_NONZERO_NUMBER] = {
 180         TERMINAL(JSON_INTEGER),
 181         ['0' ... '9'] = IN_NONZERO_NUMBER,
 182         ['e'] = IN_EXP_E,
 183         ['E'] = IN_EXP_E,
 184         ['.'] = IN_MANTISSA,
 185     },
 186
 187     [IN_NEG_NONZERO_NUMBER] = {
 188         ['0'] = IN_ZERO,
 189         ['1' ... '9'] = IN_NONZERO_NUMBER,
 190     },
 191
 192     /* keywords */
 193     [IN_KEYWORD] = {
 194         TERMINAL(JSON_KEYWORD),
 195         ['a' ... 'z'] = IN_KEYWORD,
 196     },
 197
 198     /* whitespace */
 199     [IN_WHITESPACE] = {
 200         TERMINAL(JSON_SKIP),
 201         [' '] = IN_WHITESPACE,
 202         ['\t'] = IN_WHITESPACE,
 203         ['\r'] = IN_WHITESPACE,
 204         ['\n'] = IN_WHITESPACE,
 205     },
 206
 207     /* operator */
 208     [IN_OPERATOR_DONE] = {
 209         TERMINAL(JSON_OPERATOR),
 210     },
 211
 212     /* escape */
 213     [IN_ESCAPE_DONE] = {
 214         TERMINAL(JSON_ESCAPE),
 215     },
 216
 217     [IN_ESCAPE_LL] = {
 218         ['d'] = IN_ESCAPE_DONE,
 219     },
 220
 221     [IN_ESCAPE_L] = {
 222         ['d'] = IN_ESCAPE_DONE,
 223         ['l'] = IN_ESCAPE_LL,
 224     },
 225
 226     [IN_ESCAPE] = {
 227         ['d'] = IN_ESCAPE_DONE,
 228         ['i'] = IN_ESCAPE_DONE,
 229         ['p'] = IN_ESCAPE_DONE,
 230         ['s'] = IN_ESCAPE_DONE,
 231         ['f'] = IN_ESCAPE_DONE,
 232         ['l'] = IN_ESCAPE_L,
 233     },
 234
 235     /* top level rule */
 236     [IN_START] = {
 237         ['"'] = IN_DQ_STRING,
 238         ['\''] = IN_SQ_STRING,
 239         ['0'] = IN_ZERO,
 240         ['1' ... '9'] = IN_NONZERO_NUMBER,
 241         ['-'] = IN_NEG_NONZERO_NUMBER,
 242         ['{'] = IN_OPERATOR_DONE,
 243         ['}'] = IN_OPERATOR_DONE,
 244         ['['] = IN_OPERATOR_DONE,
 245         [']'] = IN_OPERATOR_DONE,
 246         [','] = IN_OPERATOR_DONE,
 247         [':'] = IN_OPERATOR_DONE,
 248         ['a' ... 'z'] = IN_KEYWORD,
 249         ['%'] = IN_ESCAPE,
 250         [' '] = IN_WHITESPACE,
 251         ['\t'] = IN_WHITESPACE,
 252         ['\r'] = IN_WHITESPACE,
 253         ['\n'] = IN_WHITESPACE,
 254     },
 255 };
 256
 257 void json_lexer_init(JSONLexer *lexer, JSONLexerEmitter func)
 258 {
 259     lexer->emit = func;
 260     lexer->state = IN_START;
 261     lexer->token = qstring_new();
 262 }
 263
 264 static int json_lexer_feed_char(JSONLexer *lexer, char ch)
 265 {
 266     char buf[2];
 267
 268     lexer->x++;
 269     if (ch == '\n') {
 270         lexer->x = 0;
 271         lexer->y++;
 272     }
 273
 274     lexer->state = json_lexer[lexer->state][(uint8_t)ch];
 275
 276     switch (lexer->state) {
 277     case JSON_OPERATOR:
 278     case JSON_ESCAPE:
 279     case JSON_INTEGER:
 280     case JSON_FLOAT:
 281     case JSON_KEYWORD:
 282     case JSON_STRING:
 283         lexer->emit(lexer, lexer->token, lexer->state, lexer->x, lexer->y);
 284     case JSON_SKIP:
 285         lexer->state = json_lexer[IN_START][(uint8_t)ch];
 286         QDECREF(lexer->token);
 287         lexer->token = qstring_new();
 288         break;
 289     case ERROR:
 290         return -EINVAL;
 291     default:
 292         break;
 293     }
 294
 295     buf[0] = ch;
 296     buf[1] = 0;
 297
 298     qstring_append(lexer->token, buf);
 299
 300     return 0;
 301 }
 302
 303 int json_lexer_feed(JSONLexer *lexer, const char *buffer, size_t size)
 304 {
 305     size_t i;
 306
 307     for (i = 0; i < size; i++) {
 308         int err;
 309
 310         err = json_lexer_feed_char(lexer, buffer[i]);
 311         if (err < 0) {
 312             return err;
 313         }
 314     }
 315
 316     return 0;
 317 }
 318
 319 int json_lexer_flush(JSONLexer *lexer)
 320 {
 321     return json_lexer_feed_char(lexer, 0);
 322 }
 323
 324 void json_lexer_destroy(JSONLexer *lexer)
 325 {
 326     QDECREF(lexer->token);
 327 }