treat incomplete array as zero length array

[flier]

fix issue #455

[emilio]

Otherwise looks good to me, thanks for the fix!

[emilio]

I'm not sure this is legal C/C++? (the zero_length_array and incomplete_array fields would have the same address).

Could you remove the zero_length_array field from the test?

[emilio]

(or better, test them in two different structs)

[flier]

Yes, some times, people use zero length array as a mark (another story should be solved), and use incomplete array as the last field

/* define a set of marker types that can be used to refer to set points in the
 * mbuf */
__extension__
typedef void    *MARKER[0];   /**< generic marker for a point in a structure */
__extension__
typedef uint8_t  MARKER8[0];  /**< generic marker with 1B alignment */
__extension__
typedef uint64_t MARKER64[0]; /**< marker that allows us to overwrite 8 bytes
                               * with a single assignment */

/**
 * The generic rte_mbuf, containing a packet mbuf.
 */
struct rte_mbuf {
	MARKER cacheline0;

	void *buf_addr;           /**< Virtual address of segment buffer. */
	phys_addr_t buf_physaddr; /**< Physical address of segment buffer. */

	uint16_t buf_len;         /**< Length of segment buffer. */

	/* next 6 bytes are initialised on RX descriptor rearm */
	MARKER8 rearm_data;
	uint16_t data_off;

struct cmdline_inst {
	/* f(parsed_struct, data) */
	void (*f)(void *, struct cmdline *, void *);
	void *data;
	const char *help_str;
	cmdline_parse_token_hdr_t *tokens[];
};

I have submitted another patch base on a helper class like __BindgenUnionField did

#[repr(C)]
pub struct __IncompleteArrayField<T>(::std::marker::PhantomData<T>);
impl <T> __IncompleteArrayField<T> {
    #[inline]
    pub fn new() -> Self {
        __IncompleteArrayField(::std::marker::PhantomData)
    }
    #[inline]
    pub unsafe fn as_slice(&self, len: usize) -> &[T] {
        ::std::slice::from_raw_parts(::std::mem::transmute(self), len)
    }
    #[inline]
    pub unsafe fn as_mut_slice(&mut self, len: usize) -> &mut [T] {
        ::std::slice::from_raw_parts_mut(::std::mem::transmute(self), len)
    }
}
impl <T> ::std::fmt::Debug for __IncompleteArrayField<T> {
    fn fmt(&self, fmt: &mut ::std::fmt::Formatter) -> ::std::fmt::Result {
        fmt.write_str("__IncompleteArrayField")
    }
}
#[repr(C)]
pub struct C {
    pub a: ::std::os::raw::c_int,
    pub big_array: [::std::os::raw::c_char; 33usize],
    pub zero_length_array: __IncompleteArrayField<::std::os::raw::c_char>,
    pub incomplete_array: __IncompleteArrayField<::std::os::raw::c_char>,
}
#[test]
fn bindgen_test_layout_C() {
    assert_eq!(::std::mem::size_of::<C>() , 40usize);
    assert_eq!(::std::mem::align_of::<C>() , 4usize);
}

[emilio]

So I'm not super-fond of adding a new type just for this, but that being said I don't have any other complaints about this patch, and if you have a use case for the helper methods that's fine for me.

I'm slightly concerned about the array type in the static, I need to write a test case using it to verify it works, and then you have my sign-off on this.

[emilio]

@bors-servo r+

Actually marking them as zero-size array is the correct thing to do, will land a test for this in a few minutes.

[bors-servo]

📌 Commit edf7f35 has been approved by emilio

[bors-servo]

⚡ Test exempted - status

[emilio]

Thanks for the patch!

[upsuper]

In theory, both zero-length array and flexible array member are invalid in C++. Probably better putting them in a C header rather than hpp.