csv_create: 58| 309|{ 59| 309| memset(csv, 0, sizeof(struct csv)); 60| 309| csv->delimiter= ','; 61| 309| csv->quote_char = '\"'; 62| 309| csv->realloc = realloc; 63| 309| csv->emit_field = csv_emit_field_empty; 64| 309| csv->emit_row = csv_emit_row_empty; 65| 309|} csv_destroy: 69| 309|{ 70| 309| if (csv->buf) { ------------------ | Branch (70:6): [True: 56, False: 253] ------------------ 71| 56| csv->realloc(csv->buf, 0); 72| 56| csv->buf = NULL; 73| 56| } 74| 309|} csv_isvalid: 78| 309|{ 79| 309| if (csv->error_status == CSV_ER_OK && csv->state == CSV_IN_QUOTES) ------------------ | Branch (79:6): [True: 309, False: 0] | Branch (79:40): [True: 56, False: 253] ------------------ 80| 56| csv->error_status = CSV_ER_INVALID; 81| 309| return !csv->error_status; 82| 309|} csv_parse_impl: 127| 309|{ 128| 309| if (end - s == 0) ------------------ | Branch (128:6): [True: 0, False: 309] ------------------ 129| 0| return NULL; 130| 309| assert(end - s > 0); 131| 309| assert(csv->emit_field); 132| 309| assert(csv->emit_row); 133| 8.20M| for (const char *p = s; p != end; p++) { ------------------ | Branch (133:26): [True: 8.20M, False: 309] ------------------ 134| 8.20M| bool is_line_end = (*p == '\n' || *p == '\r'); ------------------ | Branch (134:23): [True: 60.8k, False: 8.14M] | Branch (134:37): [True: 1.67k, False: 8.14M] ------------------ 135| | /* realloc buffer */ 136| 8.20M| if (csv->buf == NULL || ------------------ | Branch (136:7): [True: 309, False: 8.20M] ------------------ 137| 8.20M| (csv->bufp && csv->buf_len < (size_t)(csv->bufp - csv->buf + 1))) { ------------------ | Branch (137:7): [True: 8.14M, False: 61.4k] | Branch (137:20): [True: 237, False: 8.14M] ------------------ 138| 546| size_t new_size = csv->buf_len * 2; 139| 546| if (csv->buf_len == 0 || csv->buf == NULL) ------------------ | Branch (139:8): [True: 309, False: 237] | Branch (139:29): [True: 0, False: 237] ------------------ 140| 309| new_size = 256; 141| 546| char *new_buf = (char *)csv->realloc(csv->buf, new_size); 142| 546| if (new_buf == NULL) { ------------------ | Branch (142:8): [True: 0, False: 546] ------------------ 143| 0| csv->error_status = CSV_ER_MEMORY_ERROR; 144| 0| return NULL; 145| 0| } 146| 546| csv->buf_len = new_size; 147| 546| csv->bufp = csv->bufp - csv->buf + new_buf; 148| 546| csv->buf = new_buf; 149| 546| } 150| | /* \r\n (or \n\r) linebreak, not in quotes */ 151| 8.20M| if (is_line_end && csv->state != CSV_IN_QUOTES && ------------------ | Branch (151:7): [True: 62.5k, False: 8.14M] | Branch (151:22): [True: 61.5k, False: 1.00k] ------------------ 152| 8.20M| *p != csv->prev_symbol && ------------------ | Branch (152:6): [True: 60.9k, False: 575] ------------------ 153| 8.20M| (csv->prev_symbol == '\n' || csv->prev_symbol == '\r')) { ------------------ | Branch (153:7): [True: 381, False: 60.5k] | Branch (153:36): [True: 322, False: 60.2k] ------------------ 154| 703| csv->prev_symbol = '\0'; 155| 703| continue; 156| 703| } 157| 8.20M| csv->prev_symbol = *p; 158| | /* 2 switches to avoid code dublicates */ 159| 8.20M| switch (csv->state) { ------------------ | Branch (159:11): [True: 8.13M, False: 64.5k] ------------------ 160| 61.9k| case CSV_LEADING_SPACES: ------------------ | Branch (160:3): [True: 61.9k, False: 8.14M] ------------------ 161| 61.9k| csv->bufp = csv->buf; 162| 61.9k| if (*p == ' ') /* skip spaces */ ------------------ | Branch (162:8): [True: 263, False: 61.6k] ------------------ 163| 263| continue; 164| 61.6k| csv->state = CSV_OUT_OF_QUOTES; 165| | /* symbol not handled, continue to the next switch */ 166| 61.6k| break; 167| 1.36k| case CSV_QUOTE_OPENING: ------------------ | Branch (167:3): [True: 1.36k, False: 8.20M] ------------------ 168| 1.36k| if (*p == csv->quote_char && csv->bufp) { ------------------ | Branch (168:8): [True: 398, False: 962] | Branch (168:33): [True: 398, False: 0] ------------------ 169| | /* double-quote "" */ 170| 398| *csv->bufp++ = csv->quote_char; 171| 398| csv->state = CSV_OUT_OF_QUOTES; 172| 398| continue; 173| 398| } 174| 962| csv->state = CSV_IN_QUOTES; 175| | /* symbol not handled, continue to the next switch */ 176| 962| break; 177| 1.28k| case CSV_QUOTE_CLOSING: ------------------ | Branch (177:3): [True: 1.28k, False: 8.20M] ------------------ 178| 1.28k| if (*p == csv->quote_char) { ------------------ | Branch (178:8): [True: 384, False: 898] ------------------ 179| | /* double-quote "" */ 180| 384| *csv->bufp++ = csv->quote_char; 181| 384| csv->state = CSV_IN_QUOTES; 182| 384| continue; 183| 384| } 184| 898| csv->state = CSV_OUT_OF_QUOTES; 185| | /* symbol not handled, continue to the next switch */ 186| 898| break; 187| 8.20M| } 188| | 189| 8.20M| switch (csv->state) { ------------------ | Branch (189:11): [True: 0, False: 8.20M] ------------------ 190| 7.12M| case CSV_OUT_OF_QUOTES: ------------------ | Branch (190:3): [True: 7.12M, False: 1.07M] ------------------ 191| 7.12M| if (is_line_end || *p == csv->delimiter) { ------------------ | Branch (191:8): [True: 60.8k, False: 7.06M] | Branch (191:23): [True: 616, False: 7.06M] ------------------ 192| | /* end of field */ 193| 61.4k| csv->state = CSV_LEADING_SPACES; 194| 61.4k| csv->bufp -= csv->ending_spaces; 195| 61.4k| csv->ending_spaces = 0; 196| 61.4k| if (firstonly) { ------------------ | Branch (196:9): [True: 0, False: 61.4k] ------------------ 197| 0| csv->state = CSV_NEWFIELD; 198| 0| return p; 199| 61.4k| } else { 200| 61.4k| csv->emit_field(csv->emit_ctx, 201| 61.4k| csv->buf, csv->bufp); 202| 61.4k| } 203| | 204| 61.4k| csv->bufp = csv->buf; 205| | 206| 7.06M| } else if (*p == csv->quote_char) { ------------------ | Branch (206:15): [True: 1.36k, False: 7.06M] ------------------ 207| 1.36k| csv->state = CSV_QUOTE_OPENING; 208| 7.06M| } else { 209| 7.06M| *csv->bufp++ = *p; 210| 7.06M| } 211| | 212| 7.12M| if (*p == ' ') { ------------------ | Branch (212:8): [True: 3.08k, False: 7.12M] ------------------ 213| 3.08k| csv->ending_spaces++; 214| 7.12M| } else { 215| 7.12M| csv->ending_spaces = 0; 216| 7.12M| } 217| 7.12M| if (is_line_end) { ------------------ | Branch (217:8): [True: 60.8k, False: 7.06M] ------------------ 218| | /* 219| | * bufp == buf means an empty field, 220| | * but bufp == 0 means no field at the moment, 221| | * it may be an end of the line or file 222| | */ 223| 60.8k| csv->bufp = 0; 224| 60.8k| csv->emit_row(csv->emit_ctx); 225| 60.8k| } 226| 7.12M| break; 227| 1.07M| case CSV_IN_QUOTES: ------------------ | Branch (227:3): [True: 1.07M, False: 7.12M] ------------------ 228| | /* 229| | * Bufp can became NULL in two cases: 230| | * CSV_NEWFIELD and CSV_OUT_OF_QUOTES. 231| | * 232| | * In a case of 'newfield' the csv after 233| | * nullifying bufp and returning to the 234| | * iteration starts from 235| | * CSV_LEADING_SPACES. Here bufp is set 236| | * to not NULL (see 'leading_spaces'). 237| | * 238| | * In a case of 'out_of_quotes' it can be 239| | * set to NULL only if 240| | * is_line_end == true. So at the 241| | * beginning of 'out_of_quotes' 242| | * is_line_end was true also 243| | * (see "if (is_line_end || ..." above). 244| | * In this 'if' the state of the csv is 245| | * set to CSV_LEADING_SPACES, so on a 246| | * next iteration the bufp are set to not 247| | * NULL. 248| | */ 249| 1.07M| assert(csv->bufp != NULL); 250| 1.07M| if (*p == csv->quote_char) { ------------------ | Branch (250:8): [True: 1.29k, False: 1.07M] ------------------ 251| 1.29k| csv->state = CSV_QUOTE_CLOSING; 252| 1.07M| } else { 253| 1.07M| *csv->bufp++ = *p; 254| 1.07M| } 255| 1.07M| break; 256| 0| case CSV_NEWFIELD: ------------------ | Branch (256:3): [True: 0, False: 8.20M] ------------------ 257| 0| csv->bufp = csv->buf; 258| 0| csv->state = CSV_LEADING_SPACES; 259| 0| if (is_line_end) { ------------------ | Branch (259:8): [True: 0, False: 0] ------------------ 260| 0| csv->bufp = 0; 261| 0| if (p + 1 == end) ------------------ | Branch (261:9): [True: 0, False: 0] ------------------ 262| 0| return NULL; 263| 0| else 264| 0| return p + 1; 265| | 266| 0| } 267| 0| break; 268| 8.20M| } 269| 8.20M| } 270| 309| return end; 271| 309|} csv_parse_chunk: 275| 309|csv_parse_chunk(struct csv *csv, const char *s, const char *end) { 276| 309| csv_parse_impl(csv, s, end, false); 277| 309|} csv_finish_parsing: 281| 309|{ 282| 309| if (csv_isvalid(csv)){ ------------------ | Branch (282:6): [True: 253, False: 56] ------------------ 283| 253| if (csv->bufp) { ------------------ | Branch (283:7): [True: 179, False: 74] ------------------ 284| 179| csv->bufp -= csv->ending_spaces; 285| 179| csv->emit_field(csv->emit_ctx, 286| 179| csv->buf, csv->bufp); 287| 179| csv->emit_row(csv->emit_ctx); 288| 179| } 289| 253| if (csv->buf) ------------------ | Branch (289:7): [True: 253, False: 0] ------------------ 290| 253| csv->realloc(csv->buf, 0); 291| 253| csv->bufp = NULL; 292| 253| csv->buf = NULL; 293| 253| csv->buf_len = 0; 294| 253| } 295| 309|} csv.c:csv_emit_field_empty: 50| 61.6k|{ 51| 61.6k| (void) ctx; 52| 61.6k| (void) field; 53| 61.6k| (void) end; 54| 61.6k|} csv.c:csv_emit_row_empty: 44| 61.0k|{ 45| 61.0k| (void) ctx; 46| 61.0k|} LLVMFuzzerTestOneInput: 8| 309|{ 9| 309| struct csv csv; 10| 309| csv_create(&csv); 11| 309| char *buf = calloc(size + 1, sizeof(char)); 12| 309| if (buf == NULL) ------------------ | Branch (12:6): [True: 0, False: 309] ------------------ 13| 0| return 0; 14| 309| memcpy(buf, data, size); 15| 309| buf[size] = '\0'; 16| 309| char *end = buf + size; 17| 309| csv_parse_chunk(&csv, buf, end); 18| 309| csv_finish_parsing(&csv); 19| 309| csv_destroy(&csv); 20| 309| free(buf); 21| | 22| 309| return 0; 23| 309|}