comparison: src/json.c
src/json.c
- changeset 1122
- 49ab92de9a13
- parent 1121
- 7fd2672199d7
- child 1123
- 2b83302d595a
equal
deleted
inserted
replaced
| 1121:7fd2672199d7 | 1122:49ab92de9a13 |
|---|---|
| 250 CX_JSON_NO_DATA : CX_JSON_INCOMPLETE_DATA; | 250 CX_JSON_NO_DATA : CX_JSON_INCOMPLETE_DATA; |
| 251 } | 251 } |
| 252 | 252 |
| 253 // current token type and start index | 253 // current token type and start index |
| 254 CxJsonTokenType ttype = json->uncompleted.tokentype; | 254 CxJsonTokenType ttype = json->uncompleted.tokentype; |
| 255 size_t token_start = json->buffer.pos; | 255 size_t token_part_start = json->buffer.pos; |
| 256 | 256 |
| 257 for (size_t i = json->buffer.pos; i < json->buffer.size; i++) { | 257 for (size_t i = json->buffer.pos; i < json->buffer.size; i++) { |
| 258 char c = json->buffer.space[i]; | 258 char c = json->buffer.space[i]; |
| 259 if (ttype != CX_JSON_TOKEN_STRING) { | 259 if (ttype != CX_JSON_TOKEN_STRING) { |
| 260 // currently non-string token | 260 // currently non-string token |
| 264 json->buffer.pos++; | 264 json->buffer.pos++; |
| 265 continue; | 265 continue; |
| 266 } else if (ctype == CX_JSON_TOKEN_STRING) { | 266 } else if (ctype == CX_JSON_TOKEN_STRING) { |
| 267 // begin string | 267 // begin string |
| 268 ttype = CX_JSON_TOKEN_STRING; | 268 ttype = CX_JSON_TOKEN_STRING; |
| 269 token_start = i; | 269 token_part_start = i; |
| 270 } else if (ctype != CX_JSON_NO_TOKEN) { | 270 } else if (ctype != CX_JSON_NO_TOKEN) { |
| 271 // single-char token | 271 // single-char token |
| 272 json->buffer.pos = i + 1; | 272 json->buffer.pos = i + 1; |
| 273 *result = (CxJsonToken){ctype, false, {NULL, 0}}; | 273 *result = (CxJsonToken){ctype, false, {NULL, 0}}; |
| 274 return CX_JSON_NO_ERROR; | 274 return CX_JSON_NO_ERROR; |
| 275 } else { | 275 } else { |
| 276 ttype = CX_JSON_TOKEN_LITERAL; // number or literal | 276 ttype = CX_JSON_TOKEN_LITERAL; // number or literal |
| 277 token_start = i; | 277 token_part_start = i; |
| 278 } | 278 } |
| 279 } else { | 279 } else { |
| 280 // finish token | 280 // finish token |
| 281 if (ctype != CX_JSON_NO_TOKEN) { | 281 if (ctype != CX_JSON_NO_TOKEN) { |
| 282 *result = token_create(json, false, token_start, i); | 282 *result = token_create(json, false, token_part_start, i); |
| 283 if (result->tokentype == CX_JSON_NO_TOKEN) { | 283 if (result->tokentype == CX_JSON_NO_TOKEN) { |
| 284 return CX_JSON_BUFFER_ALLOC_FAILED; // LCOV_EXCL_LINE | 284 return CX_JSON_BUFFER_ALLOC_FAILED; // LCOV_EXCL_LINE |
| 285 } | 285 } |
| 286 if (result->tokentype == CX_JSON_TOKEN_ERROR) { | 286 if (result->tokentype == CX_JSON_TOKEN_ERROR) { |
| 287 return CX_JSON_FORMAT_ERROR_NUMBER; | 287 return CX_JSON_FORMAT_ERROR_NUMBER; |
| 294 // currently inside a string | 294 // currently inside a string |
| 295 if (json->tokenizer_escape) { | 295 if (json->tokenizer_escape) { |
| 296 json->tokenizer_escape = false; | 296 json->tokenizer_escape = false; |
| 297 } else { | 297 } else { |
| 298 if (c == '"') { | 298 if (c == '"') { |
| 299 *result = token_create(json, true, token_start, i + 1); | 299 *result = token_create(json, true, token_part_start, i + 1); |
| 300 if (result->tokentype == CX_JSON_NO_TOKEN) { | 300 if (result->tokentype == CX_JSON_NO_TOKEN) { |
| 301 return CX_JSON_BUFFER_ALLOC_FAILED; // LCOV_EXCL_LINE | 301 return CX_JSON_BUFFER_ALLOC_FAILED; // LCOV_EXCL_LINE |
| 302 } | 302 } |
| 303 json->buffer.pos = i + 1; | 303 json->buffer.pos = i + 1; |
| 304 return CX_JSON_NO_ERROR; | 304 return CX_JSON_NO_ERROR; |
| 309 } | 309 } |
| 310 } | 310 } |
| 311 | 311 |
| 312 if (ttype != CX_JSON_NO_TOKEN) { | 312 if (ttype != CX_JSON_NO_TOKEN) { |
| 313 // uncompleted token | 313 // uncompleted token |
| 314 size_t uncompleted_len = json->buffer.size - token_start; | 314 size_t uncompleted_len = json->buffer.size - token_part_start; |
| 315 if (json->uncompleted.tokentype == CX_JSON_NO_TOKEN) { | 315 if (json->uncompleted.tokentype == CX_JSON_NO_TOKEN) { |
| 316 // current token is uncompleted | 316 // current token is uncompleted |
| 317 // save current token content | 317 // save current token content |
| 318 CxJsonToken uncompleted = { | 318 CxJsonToken uncompleted = { |
| 319 ttype, true, | 319 ttype, true, |
| 320 cx_strdup(cx_strn(json->buffer.space + token_start, uncompleted_len)) | 320 cx_strdup(cx_strn(json->buffer.space + token_part_start, uncompleted_len)) |
| 321 }; | 321 }; |
| 322 if (uncompleted.content.ptr == NULL) { | 322 if (uncompleted.content.ptr == NULL) { |
| 323 return CX_JSON_BUFFER_ALLOC_FAILED; // LCOV_EXCL_LINE | 323 return CX_JSON_BUFFER_ALLOC_FAILED; // LCOV_EXCL_LINE |
| 324 } | 324 } |
| 325 json->uncompleted = uncompleted; | 325 json->uncompleted = uncompleted; |
| 326 } else { | 326 } else { |
| 327 // previously we also had an uncompleted token | 327 // previously we also had an uncompleted token |
| 328 // combine the uncompleted token with the current token | 328 // combine the uncompleted token with the current token |
| 329 assert(json->uncompleted.allocated); | 329 assert(json->uncompleted.allocated); |
| 330 cxmutstr str = cx_strcat_m(json->uncompleted.content, 1, | 330 cxmutstr str = cx_strcat_m(json->uncompleted.content, 1, |
| 331 cx_strn(json->buffer.space + token_start, uncompleted_len)); | 331 cx_strn(json->buffer.space + token_part_start, uncompleted_len)); |
| 332 if (str.ptr == NULL) { | 332 if (str.ptr == NULL) { |
| 333 return CX_JSON_BUFFER_ALLOC_FAILED; // LCOV_EXCL_LINE | 333 return CX_JSON_BUFFER_ALLOC_FAILED; // LCOV_EXCL_LINE |
| 334 } | 334 } |
| 335 json->uncompleted.content = str; | 335 json->uncompleted.content = str; |
| 336 } | 336 } |
| 340 | 340 |
| 341 return CX_JSON_INCOMPLETE_DATA; | 341 return CX_JSON_INCOMPLETE_DATA; |
| 342 } | 342 } |
| 343 | 343 |
| 344 static cxmutstr unescape_string(const CxAllocator *a, cxmutstr str) { | 344 static cxmutstr unescape_string(const CxAllocator *a, cxmutstr str) { |
| 345 // TODO: support more escape sequences | 345 // note: this function expects that str contains the enclosing quotes! |
| 346 // TODO: to be consistent with escape_string() we might want to expect that the enclosing quotes were already removed | 346 |
| 347 cxmutstr result; | 347 cxmutstr result; |
| 348 result.length = 0; | 348 result.length = 0; |
| 349 result.ptr = cxMalloc(a, str.length - 1); | 349 result.ptr = cxMalloc(a, str.length - 1); |
| 350 if (result.ptr == NULL) return result; // LCOV_EXCL_LINE | 350 if (result.ptr == NULL) return result; // LCOV_EXCL_LINE |
| 351 | 351 |
| 356 u = false; | 356 u = false; |
| 357 if (c == 'n') { | 357 if (c == 'n') { |
| 358 c = '\n'; | 358 c = '\n'; |
| 359 } else if (c == 't') { | 359 } else if (c == 't') { |
| 360 c = '\t'; | 360 c = '\t'; |
| 361 } | 361 } else if (c == 'r') { |
| 362 c = '\r'; | |
| 363 } else if (c == '\\') { | |
| 364 c = '\\'; | |
| 365 } else if (c == '/') { | |
| 366 c = '/'; // always unescape, we don't need settings here | |
| 367 } else if (c == 'f') { | |
| 368 c = '\f'; | |
| 369 } else if (c == 'b') { | |
| 370 c = '\b'; | |
| 371 } | |
| 372 // TODO: support \uXXXX escape sequences | |
| 373 // TODO: discuss the behavior for unrecognized escape sequences | |
| 374 // most parsers throw an error here | |
| 362 result.ptr[result.length++] = c; | 375 result.ptr[result.length++] = c; |
| 363 } else { | 376 } else { |
| 364 if (c == '\\') { | 377 if (c == '\\') { |
| 365 u = true; | 378 u = true; |
| 366 } else { | 379 } else { |
| 372 | 385 |
| 373 return result; | 386 return result; |
| 374 } | 387 } |
| 375 | 388 |
| 376 static cxmutstr escape_string(cxmutstr str) { | 389 static cxmutstr escape_string(cxmutstr str) { |
| 390 // note: this function produces the string without enclosing quotes | |
| 391 // the reason is that we don't want to allocate memory just for that | |
| 377 CxBuffer buf = {0}; | 392 CxBuffer buf = {0}; |
| 378 | 393 |
| 379 bool all_printable = true; | 394 bool all_printable = true; |
| 380 for (size_t i = 0; i < str.length; i++) { | 395 for (size_t i = 0; i < str.length; i++) { |
| 381 bool escape = !isprint(str.ptr[i]) | 396 bool escape = !isprint(str.ptr[i]) |
