strings do not have to be valid UTF-8 any more

This commit is contained in:
Ralf Jung 2020-05-12 09:46:41 +02:00
parent 09c817eeb2
commit 6f9810c981
3 changed files with 12 additions and 11 deletions

View File

@ -15,7 +15,7 @@ use rustc_middle::mir::interpret::{InterpError, InterpErrorInfo};
use rustc_middle::ty;
use rustc_middle::ty::layout::TyAndLayout;
use rustc_span::symbol::{sym, Symbol};
use rustc_target::abi::{Abi, LayoutOf, Scalar, VariantIdx, Variants};
use rustc_target::abi::{Abi, LayoutOf, Scalar, Size, VariantIdx, Variants};
use std::hash::Hash;
@ -744,10 +744,11 @@ impl<'rt, 'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> ValueVisitor<'mir, 'tcx, M>
match op.layout.ty.kind {
ty::Str => {
let mplace = op.assert_mem_place(self.ecx); // strings are never immediate
let len = mplace.len(self.ecx)?;
try_validation!(
self.ecx.read_str(mplace),
self.ecx.memory.read_bytes(mplace.ptr, Size::from_bytes(len)),
self.path,
err_ub!(InvalidStr(..)) => { "uninitialized or non-UTF-8 data in str" },
err_ub!(InvalidUninitBytes(..)) => { "uninitialized data in `str`" },
);
}
ty::Array(tys, ..) | ty::Slice(tys)

View File

@ -42,11 +42,11 @@ const MY_STR_LENGTH_PTR: &MyStr = unsafe { mem::transmute((&42u8, &3)) };
const MY_STR_MUCH_TOO_LONG: &MyStr = unsafe { mem::transmute((&42u8, usize::MAX)) };
//~^ ERROR it is undefined behavior to use this value
// invalid UTF-8
const STR_NO_UTF8: &str = unsafe { mem::transmute::<&[u8], _>(&[0xFF]) };
// uninitialized byte
const STR_NO_INIT: &str = unsafe { mem::transmute::<&[_], _>(&[MaybeUninit::<u8> { uninit: () }]) };
//~^ ERROR it is undefined behavior to use this value
// invalid UTF-8 in user-defined str-like
const MYSTR_NO_UTF8: &MyStr = unsafe { mem::transmute::<&[u8], _>(&[0xFF]) };
// uninitialized byte in user-defined str-like
const MYSTR_NO_INIT: &MyStr = unsafe { mem::transmute::<&[_], _>(&[MaybeUninit::<u8> { uninit: () }]) };
//~^ ERROR it is undefined behavior to use this value
// # slice

View File

@ -41,16 +41,16 @@ LL | const MY_STR_MUCH_TOO_LONG: &MyStr = unsafe { mem::transmute((&42u8, usize:
error[E0080]: it is undefined behavior to use this value
--> $DIR/ub-wide-ptr.rs:46:1
|
LL | const STR_NO_UTF8: &str = unsafe { mem::transmute::<&[u8], _>(&[0xFF]) };
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ type validation failed: encountered uninitialized or non-UTF-8 data in str at .<deref>
LL | const STR_NO_INIT: &str = unsafe { mem::transmute::<&[_], _>(&[MaybeUninit::<u8> { uninit: () }]) };
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ type validation failed: encountered uninitialized data in `str` at .<deref>
|
= note: The rules on what exactly is undefined behavior aren't clear, so this check might be overzealous. Please open an issue on the rustc repository if you believe it should not be considered undefined behavior.
error[E0080]: it is undefined behavior to use this value
--> $DIR/ub-wide-ptr.rs:49:1
|
LL | const MYSTR_NO_UTF8: &MyStr = unsafe { mem::transmute::<&[u8], _>(&[0xFF]) };
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ type validation failed: encountered uninitialized or non-UTF-8 data in str at .<deref>.0
LL | const MYSTR_NO_INIT: &MyStr = unsafe { mem::transmute::<&[_], _>(&[MaybeUninit::<u8> { uninit: () }]) };
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ type validation failed: encountered uninitialized data in `str` at .<deref>.0
|
= note: The rules on what exactly is undefined behavior aren't clear, so this check might be overzealous. Please open an issue on the rustc repository if you believe it should not be considered undefined behavior.