Function Strs_str

python/stringzilla.c:6744–6823

* @brief Array to string conversion method, that concatenates all the strings in the array. * Will output an object that looks like `['item1', 'item2', 'item3']`, containing all * the strings. */

 *          the strings.
 */
static PyObject *Strs_str(Strs *self) {
    get_string_at_offset_t getter = str_at_offset_getter(self);
    if (!getter) {
        PyErr_SetString(PyExc_TypeError, "Unknown Strs kind");
        return NULL;
    }

    // Aggregate the total length of all the slices and count the number of bytes we need to allocate:
    sz_size_t count = Strs_len(self);
    PyObject *parent_string;
    sz_size_t total_bytes = 2; // opening and closing square brackets
    for (sz_size_t i = 0; i < count; i++) {
        sz_cptr_t cstr_start = NULL;
        sz_size_t cstr_length = 0;
        getter(self, i, count, &parent_string, &cstr_start, &cstr_length);

        if (i != 0) total_bytes += 2; // For the preceding comma and space

        // Check if string is valid UTF-8 to determine format
        if (sz_utf8_valid(cstr_start, cstr_length)) {
            // Valid UTF-8: format as '...' with escaped quotes
            total_bytes += 2;           // Opening and closing quotes
            total_bytes += cstr_length; // Base string length

            // Count the number of single quotes that need escaping
            sz_cptr_t scan_ptr = cstr_start;
            sz_size_t scan_length = cstr_length;
            while (scan_length) {
                char quote = '\'';
                sz_cptr_t next_quote = sz_find_byte(scan_ptr, scan_length, &quote);
                if (next_quote == NULL) break;
                total_bytes++; // Extra byte for escaping
                scan_length -= next_quote - scan_ptr + 1;
                scan_ptr = next_quote + 1;
            }
        }
        else {
            // Invalid UTF-8: format as b'\x...'
            total_bytes += 3;               // "b'" prefix
            total_bytes += cstr_length * 4; // Each byte becomes \xNN (4 chars)
            total_bytes += 1;               // Closing quote
        }
    }

    // Now allocate the memory for the concatenated string
    sz_ptr_t const result_buffer = malloc(total_bytes);
    if (!result_buffer) {
        PyErr_SetString(PyExc_MemoryError, "Failed to allocate memory for the concatenated string");
        return NULL;
    }

    // Copy the strings into the result buffer
    sz_ptr_t result_ptr = result_buffer;
    *result_ptr++ = '[';
    for (sz_size_t i = 0; i < count; i++) {
        if (i != 0) {
            *result_ptr++ = ',';
            *result_ptr++ = ' ';