Initial importv0.1.0

1 files changed, 168 insertions, 0 deletions

diff --git a/src/csv.c b/src/csv.c
new file mode 100644
index 0000000..ea9859b
--- /dev/null
+++ b/src/csv.c
@@ -0,0 +1,168 @@
+#include "csv.h"
+
+#include <stdbool.h>
+#include <stdlib.h>
+#include <string.h>
+
+struct csv_reader {
+        const char *p;
+        const char *end;
+};
+
+typedef struct {
+        char *data;
+        size_t len;
+        size_t cap;
+} csv_buf_t;
+
+csv_reader_t *csv_reader_new(const char *data, size_t len)
+{
+        csv_reader_t *reader;
+
+        if (data == NULL)
+                return NULL;
+        reader = calloc(1, sizeof(*reader));
+        if (reader == NULL)
+                return NULL;
+        reader->p = data;
+        reader->end = data + len;
+        if (len >= 3 && memcmp(data, "\xEF\xBB\xBF", 3) == 0)
+                reader->p += 3;
+        return reader;
+}
+
+void csv_reader_free(csv_reader_t *reader)
+{
+        free(reader);
+}
+
+void csv_row_free(csv_row_t *row)
+{
+        size_t i;
+
+        if (row == NULL)
+                return;
+        for (i = 0; i < row->count; i++)
+                free(row->fields[i]);
+        free(row->fields);
+        row->fields = NULL;
+        row->count = 0;
+}
+
+static bool buf_push(csv_buf_t *buf, char c)
+{
+        if (buf->len + 2 > buf->cap) {
+                size_t cap = buf->cap ? buf->cap * 2 : 32;
+                char *grown = realloc(buf->data, cap);
+                if (grown == NULL)
+                        return false;
+                buf->data = grown;
+                buf->cap = cap;
+        }
+        buf->data[buf->len++] = c;
+        buf->data[buf->len] = '\0';
+        return true;
+}
+
+static bool row_push(csv_row_t *row, size_t *cap, char *field)
+{
+        if (row->count + 1 > *cap) {
+                size_t next = *cap ? *cap * 2 : 16;
+                char **grown = realloc(row->fields, next * sizeof(*grown));
+                if (grown == NULL)
+                        return false;
+                row->fields = grown;
+                *cap = next;
+        }
+        row->fields[row->count++] = field;
+        return true;
+}
+
+static char *parse_field(csv_reader_t *reader)
+{
+        csv_buf_t field = {NULL, 0, 0};
+        bool quoted = false;
+
+        if (!buf_push(&field, '\0'))
+                return NULL;
+        field.len = 0;
+        field.data[0] = '\0';
+
+        if (reader->p < reader->end && *reader->p == '"') {
+                quoted = true;
+                reader->p++;
+        }
+
+        while (reader->p < reader->end) {
+                char c = *reader->p;
+
+                if (quoted) {
+                        reader->p++;
+                        if (c != '"') {
+                                if (!buf_push(&field, c))
+                                        goto fail;
+                                continue;
+                        }
+                        if (reader->p < reader->end && *reader->p == '"') {
+                                reader->p++;
+                                if (!buf_push(&field, '"'))
+                                        goto fail;
+                                continue;
+                        }
+                        quoted = false;
+                        continue;
+                }
+                if (c == ',' || c == '\n' || c == '\r')
+                        break;
+                reader->p++;
+                if (!buf_push(&field, c))
+                        goto fail;
+        }
+
+        if (quoted)
+                goto fail;
+        return field.data;
+
+fail:
+        free(field.data);
+        return NULL;
+}
+
+int csv_reader_next(csv_reader_t *reader, csv_row_t *out)
+{
+        csv_row_t row = {NULL, 0};
+        size_t cap = 0;
+
+        if (reader == NULL || out == NULL)
+                return -1;
+        if (reader->p >= reader->end)
+                return 0;
+
+        for (;;) {
+                char *field = parse_field(reader);
+
+                if (field == NULL) {
+                        csv_row_free(&row);
+                        return -1;
+                }
+                if (!row_push(&row, &cap, field)) {
+                        free(field);
+                        csv_row_free(&row);
+                        return -1;
+                }
+                if (reader->p >= reader->end)
+                        break;
+                if (*reader->p == ',') {
+                        reader->p++;
+                        continue;
+                }
+                if (*reader->p == '\r')
+                        reader->p++;
+                if (reader->p < reader->end && *reader->p == '\n')
+                        reader->p++;
+                break;
+        }
+
+        *out = row;
+        return 1;
+}