diff --git a/py/objstr.c b/py/objstr.c index 714a170d0a..6ddba18a3b 100644 --- a/py/objstr.c +++ b/py/objstr.c @@ -344,11 +344,14 @@ uncomparable: return MP_OBJ_NULL; // op not supported } +#if !MICROPY_PY_BUILTINS_STR_UNICODE +// objstrunicode defines own version const byte *str_index_to_ptr(const mp_obj_type_t *type, const byte *self_data, uint self_len, mp_obj_t index, bool is_slice) { machine_uint_t index_val = mp_get_index(type, self_len, index, is_slice); return self_data + index_val; } +#endif STATIC mp_obj_t str_subscr(mp_obj_t self_in, mp_obj_t index, mp_obj_t value) { mp_obj_type_t *type = mp_obj_get_type(self_in); diff --git a/py/objstr.h b/py/objstr.h index 75de50d29b..da2c9cd513 100644 --- a/py/objstr.h +++ b/py/objstr.h @@ -57,6 +57,9 @@ mp_obj_t mp_obj_new_str_of_type(const mp_obj_type_t *type, const byte* data, uin mp_obj_t str_binary_op(int op, mp_obj_t lhs_in, mp_obj_t rhs_in); machine_int_t str_get_buffer(mp_obj_t self_in, mp_buffer_info_t *bufinfo, int flags); +const byte *str_index_to_ptr(const mp_obj_type_t *type, const byte *self_data, uint self_len, + mp_obj_t index, bool is_slice); + MP_DECLARE_CONST_FUN_OBJ(str_encode_obj); MP_DECLARE_CONST_FUN_OBJ(str_find_obj); MP_DECLARE_CONST_FUN_OBJ(str_rfind_obj); diff --git a/py/objstrunicode.c b/py/objstrunicode.c index 334d9895e8..052028eee0 100644 --- a/py/objstrunicode.c +++ b/py/objstrunicode.c @@ -154,7 +154,8 @@ STATIC mp_obj_t str_make_new(mp_obj_t type_in, uint n_args, uint n_kw, const mp_ // Convert an index into a pointer to its lead byte. Out of bounds indexing will raise IndexError or // be capped to the first/last character of the string, depending on is_slice. -STATIC const char *str_index_to_ptr(const char *self_data, uint self_len, mp_obj_t index, bool is_slice) { +const byte *str_index_to_ptr(const mp_obj_type_t *type, const byte *self_data, uint self_len, + mp_obj_t index, bool is_slice) { machine_int_t i; // Copied from mp_get_index; I don't want bounds checking, just give me // the integer as-is. (I can't bounds-check without scanning the whole @@ -164,7 +165,7 @@ STATIC const char *str_index_to_ptr(const char *self_data, uint self_len, mp_obj } else if (!mp_obj_get_int_maybe(index, &i)) { nlr_raise(mp_obj_new_exception_msg_varg(&mp_type_TypeError, "string indices must be integers, not %s", mp_obj_get_type_str(index))); } - const char *s, *top = self_data + self_len; + const byte *s, *top = self_data + self_len; if (i < 0) { // Negative indexing is performed by counting from the end of the string. @@ -235,18 +236,18 @@ STATIC mp_obj_t str_subscr(mp_obj_t self_in, mp_obj_t index, mp_obj_t value) { } return mp_obj_new_str_of_type(type, self_data + start, stop - start); } - const char *pstart, *pstop; + const byte *pstart, *pstop; if (ostart != mp_const_none) { - pstart = str_index_to_ptr((const char *)self_data, self_len, ostart, true); + pstart = str_index_to_ptr(type, self_data, self_len, ostart, true); } else { - pstart = (const char *)self_data; + pstart = self_data; } if (ostop != mp_const_none) { // pstop will point just after the stop character. This depends on // the \0 at the end of the string. - pstop = str_index_to_ptr((const char *)self_data, self_len, ostop, true); + pstop = str_index_to_ptr(type, self_data, self_len, ostop, true); } else { - pstop = (const char *)self_data + self_len; + pstop = self_data + self_len; } if (pstop < pstart) { return MP_OBJ_NEW_QSTR(MP_QSTR_); @@ -258,7 +259,7 @@ STATIC mp_obj_t str_subscr(mp_obj_t self_in, mp_obj_t index, mp_obj_t value) { uint index_val = mp_get_index(type, self_len, index, false); return MP_OBJ_NEW_SMALL_INT((mp_small_int_t)self_data[index_val]); } - const char *s = str_index_to_ptr((const char *)self_data, self_len, index, false); + const byte *s = str_index_to_ptr(type, self_data, self_len, index, false); int len = 1; if (UTF8_IS_NONASCII(*s)) { // Count the number of 1 bits (after the first) @@ -266,7 +267,7 @@ STATIC mp_obj_t str_subscr(mp_obj_t self_in, mp_obj_t index, mp_obj_t value) { ++len; } } - return mp_obj_new_str(s, len, true); // This will create a one-character string + return mp_obj_new_str((const char*)s, len, true); // This will create a one-character string } else { return MP_OBJ_NULL; // op not supported }