31e46ac0a9
The current behavior leads to adjustments like `&&*` being applied instead of just `&` (when the unmodified receiver is a `&T` or an `&mut T`). This causes both safety errors and unexpected behavior. The safety errors result from regionck not being prepared for auto-ref-ref-like adjustments; this is worth fixing on its own, but I think the best way to do it is to modify regionck to use expr-use-visitor (and fix expr-use-visitor as well, which I don't think properly invokes `borrow` for each level of auto-ref), and for now it's simpler to just not produce the adjustment in question. (I have a separate patch porting regionck to use exprusevisitor for a different bug, so that is coming.)
55 lines
1.6 KiB
Rust
55 lines
1.6 KiB
Rust
// Copyright 2014 The Rust Project Developers. See the COPYRIGHT
|
|
// file at the top-level directory of this distribution and at
|
|
// http://rust-lang.org/COPYRIGHT.
|
|
//
|
|
// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
|
|
// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
|
|
// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
|
|
// option. This file may not be copied, modified, or distributed
|
|
// except according to those terms.
|
|
|
|
// Test that an `&mut self` method, when invoked on an lvalue whose
|
|
// type is `&mut [u8]`, passes in a pointer to the lvalue and not a
|
|
// temporary. Issue #19147.
|
|
|
|
use std::raw;
|
|
use std::mem;
|
|
use std::slice;
|
|
use std::io::IoResult;
|
|
|
|
trait MyWriter {
|
|
fn my_write(&mut self, buf: &[u8]) -> IoResult<()>;
|
|
}
|
|
|
|
impl<'a> MyWriter for &'a mut [u8] {
|
|
fn my_write(&mut self, buf: &[u8]) -> IoResult<()> {
|
|
slice::bytes::copy_memory(*self, buf);
|
|
|
|
let write_len = buf.len();
|
|
unsafe {
|
|
*self = mem::transmute(raw::Slice {
|
|
data: self.as_ptr().offset(write_len as int),
|
|
len: self.len() - write_len,
|
|
});
|
|
}
|
|
|
|
Ok(())
|
|
}
|
|
}
|
|
|
|
fn main() {
|
|
let mut buf = [0_u8, .. 6];
|
|
|
|
{
|
|
let mut writer = buf.as_mut_slice();
|
|
writer.my_write(&[0, 1, 2]).unwrap();
|
|
writer.my_write(&[3, 4, 5]).unwrap();
|
|
}
|
|
|
|
// If `my_write` is not modifying `buf` in place, then we will
|
|
// wind up with `[3, 4, 5, 0, 0, 0]` because the first call to
|
|
// `my_write()` doesn't update the starting point for the write.
|
|
|
|
assert_eq!(buf, [0, 1, 2, 3, 4, 5]);
|
|
}
|