Move to intra-doc links for /library/core/src/intrinsics.rs
Helps with #75080.
@rustbot modify labels: T-doc, A-intra-doc-links, T-rustdoc
Known issues:
* The following f32 and f64 primitive methods cannot be resolved:
f32/f64::powi
f32/f64::sqrt
f32/f64::sin
f32/f64::cos
f32/f64::powf
f32/f64::exp
f32/f64::exp2
f32/f64::ln
f32/f64::log2
f32/f64::log10
f32/f64::mul_add
f32/f64::abs
f32/f64::copysign
f32/f64::floor
f32/f64::ceil
f32/f64::trunc
f32/f64::round
* Links from core to std:
[`std::pointer::*`]
[`std::process::abort`]
[`from_raw_parts`]
[`Vec::append`]
* Links with anchors?
I provided a separate commit that replaced links with anchors by intra-doc links.
Here the anchor location information gets lost, so its questionable whether to
actually replace those links.
enable align_to tests in Miri
With https://github.com/rust-lang/miri/issues/1074 resolved, we can enable these tests in Miri.
I also tweaked the test sized to get reasonable execution times with decent test coverage.
Use min_specialization in libcore
Getting `TrustedRandomAccess` to work is the main interesting thing here.
- `get_unchecked` is now an unstable, hidden method on `Iterator`
- The contract for `TrustedRandomAccess` is made clearer in documentation
- Fixed a bug where `Debug` would create aliasing references when using the specialized zip impl
- Added tests for the side effects of `next_back` and `nth`.
closes#68536
`stride == 1` case can be computed more efficiently through `-p (mod
a)`. That, then translates to a nice and short sequence of LLVM
instructions:
%address = ptrtoint i8* %p to i64
%negptr = sub i64 0, %address
%offset = and i64 %negptr, %a_minus_one
And produces pretty much ideal code-gen when this function is used in
isolation.
Typical use of this function will, however, involve use of
the result to offset a pointer, i.e.
%aligned = getelementptr inbounds i8, i8* %p, i64 %offset
This still looks very good, but LLVM does not really translate that to
what would be considered ideal machine code (on any target). For example
that's the codegen we obtain for an unknown alignment:
; x86_64
dec rsi
mov rax, rdi
neg rax
and rax, rsi
add rax, rdi
In particular negating a pointer is not something that’s going to be
optimised for in the design of CISC architectures like x86_64. They
are much better at offsetting pointers. And so we’d love to utilize this
ability and produce code that's more like this:
; x86_64
lea rax, [rsi + rdi - 1]
neg rsi
and rax, rsi
To achieve this we need to give LLVM an opportunity to apply its
various peep-hole optimisations that it does during DAG selection. In
particular, the `and` instruction appears to be a major inhibitor here.
We cannot, sadly, get rid of this load-bearing operation, but we can
reorder operations such that LLVM has more to work with around this
instruction.
One such ordering is proposed in #75579 and results in LLVM IR that
looks broadly like this:
; using add enables `lea` and similar CISCisms
%offset_ptr = add i64 %address, %a_minus_one
%mask = sub i64 0, %a
%masked = and i64 %offset_ptr, %mask
; can be folded with `gepi` that may follow
%offset = sub i64 %masked, %address
…and generates the intended x86_64 machine code. One might also wonder
how the increased amount of code would impact a RISC target. Turns out
not much:
; aarch64 previous ; aarch64 new
sub x8, x1, #1 add x8, x1, x0
neg x9, x0 sub x8, x8, #1
and x8, x9, x8 neg x9, x1
add x0, x0, x8 and x0, x8, x9
(and similarly for ppc, sparc, mips, riscv, etc)
The only target that seems to do worse is… wasm32.
Onto actual measurements – the best way to evaluate snippets like these
is to use llvm-mca. Much like Aarch64 assembly would allow to suspect,
there isn’t any performance difference to be found. Both snippets
execute in same number of cycles for the CPUs I tried. On x86_64,
we get throughput improvement of >50%, however!
Improve codegen for `align_offset`
In this PR the `align_offset` implementation is changed/improved to produce better code in certain scenarios such as when pointer type is has a stride of 1 or when building for low optimisation levels.
While these changes do not achieve the "ideal" codegen referenced in #75579, it gets significantly closer to it. I’m not actually sure if the codegen can actually be much better with this function returning the offset, rather than the aligned pointer.
See the descriptions for separate commits for further information.
docs(marker/copy): provide example for `&T` being `Copy`
### Edited 2020-08-16 (most recent)
In the current documentation about the `Copy` marker trait, there is a section
with examples of structs that can implement `Copy`. Currently there is no example for
showing that shared references (`&T`) are also `Copy`.
It is worth to have a dedicated example for `&T` being `Copy`, because shared
references are an integral part of the language and it being `Copy` is not as
intuitive as other types that share this behaviour like `i32` or `bool`.
The example picks up on the previous non-`Copy` struct and shows that
structs can be `Copy`, even when they hold a shared reference to a non-`Copy` type.
-----------------------------------------
### Edited 2020-08-02, 3:28 p.m.
I've just realized that it says "in addition to the **implementors listed below**", which makes this PR kind of "wrong", because `&T` is indeed in the "implementors listed below".
Maybe we can instead show an example with `&T` in the [When can my type be Copy](https://doc.rust-lang.org/std/marker/trait.Copy.html#when-can-my-type-be-copy) section.
What I really want to achieve is that it becomes more obvious that `&T` is also `Copy`, because, I think, it is very valuable to know and it wasn't obvious for me, until I read something about it in a forum post.
What do you think? I would create another PR for that.
**Please feel free to close this PR.**
-----------------------------------
### Original post
In the current documentation about the `Copy` marker trait, there is a section
about "additional implementors", which list additional implementors of the `Copy` trait.
The fact that shared references are also `Copy` is mixed with another point,
which makes it hard to recognize and make it seem not as important.
This clarifies the fact that shared references are also `Copy`, by mentioning it as a
separate item in the list of "additional implementors".
See also X-Link mem::{swap, take, replace}
Since it's easy to end up at one of these functions when you really wanted the other one, cross link them with descriptions of why you'd want to use them.
Add `as_uninit`-like methods to pointer types and unify documentation of `as_ref` methods
This adds a convenient method to retrieve a `&(mut) [MaybeUninit<T>]` from slice pointers (`*const [T]`, `*mut [T]`, `NonNull<[T]>`). See also https://github.com/rust-lang/wg-allocators/issues/66#issuecomment-671789105.
~I'll add a tracking issue as soon as it's reviewed and CI passed.~
Tracking Issue: #75402
r? @RalfJung
Reference lang items during AST lowering
Fixes#60607 and fixes#61019.
This PR introduces `QPath::LangItem` to the HIR and uses it in AST lowering instead of constructing a `hir::Path` from a slice of symbols:
- Credit for much of this work goes to @matthewjasper, I basically just [rebased their earlier work](a227c706b7 (diff-c0f791ead38d2d02916faaad0f56f41d)).
- ~~Changes to Clippy might not be correct, they compile but attempting to run tests through `./x.py` produced failures which appeared spurious, so I didn't run any clippy tests.~~
- Changes to save analysis might not be correct - tests pass but I don't have a lot of confidence in those changes being correct.
- I've used `GenericBounds::LangItemTrait` rather than changing `PolyTraitRef`, as suggested by @matthewjasper [in this comment](a227c706b7 (r40107992)) but I'd prefer that be left for a follow-up.
- I've split things into smaller commits fairly arbitrarily to make the diff easier to review, each commit should compile but might not pass tests until the final commit.
r? @oli-obk
cc @matthewjasper
Fix example in `NonNull::as_uninit_slice`
Rename feature gate to "ptr_as_uninit"
Make methods more consistent with already stable methods
Make `pointer::as_uninit_slice` return an `Option`
Fix placement for `// SAFETY` section
Add `as_uninit_ref` and `as_uninit_mut` to pointers
Fix doctest
Update tracking issue
Fix doc links
Apply suggestions from review
Make wording about counterparts consistent
Fix doc links
Improve documentation
Fix doc-tests
Fix doc links ... again
Apply suggestions from review
Apply suggestions from Review
Apply suggestion from review to all affected files
Add missing words in safety sections in `as_uninit_slice_mut`
Fix safety-comment in `NonNull::as_uninit_slice_mut`
Move to intra-doc links for /library/core/src/any.rs
Helps with #75080.
@rustbot modify labels: T-doc, A-intra-doc-links, T-rustdoc
Known issues:
* Links from `core` to `std` (#74481):
* `[Box]: ../../std/boxed/struct.Box.html`
pin docs: add some forward references
@nagisa had some questions about pinning that were answered in the docs, which they did not realize because that discussion is below the examples. I still think it makes sense to introduce the examples before that discussion, since it give the discussion something concrete to refer to, but this PR adds some forward references so people don't think the examples conclude the docs.
@nagisa do you think this would have helped?
Code blocks in doc comments are compiled and run, so we show `Copy` works in this example.
Co-authored-by: Poliorcetics <poliorcetics@users.noreply.github.com>
Previously checking for `pmoda == 0` would get LLVM to generate branchy
code, when, for `stride = 1` the offset can be computed without such a
branch by doing effectively a `-p % a`.
For well-known (constant) alignments, with the new ordering of these
conditionals, we end up generating 2 to 3 cheap instructions on x86_64:
movq %rdi, %rax
negl %eax
andl $7, %eax
instead of 5+ as previously.
For unknown alignments the new code also generates just 3 instructions:
negq %rdi
leaq -1(%rsi), %rax
andq %rdi, %rax
At opt-level <= 1, the methods such as `wrapping_mul` are not being
inlined, causing significant bloating and slowdowns of the
implementation at these optimisation levels.
With use of these intrinsics, the codegen of this function at
-Copt_level=1 is the same as it is at -Copt_level=3.
In the current documentation about the `Copy` marker trait, there is a section
with examples of structs that can implement `Copy`. Currently there is no example for
showing that shared references (`&T`) are also `Copy`.
It is worth to have a dedicated example for `&T` being `Copy`, because shared
references are an integral part of the language and it being `Copy` is not as
intuitive as other types that share this behaviour like `i32` or `bool`.
The example picks up on the previous non-`Copy` struct and shows that
structs can be `Copy`, even when they hold a shared reference to a non-`Copy` type.
This commit adds new lang items which will be used in AST lowering once
`QPath::LangItem` is introduced.
Co-authored-by: Matthew Jasper <mjjasper1@gmail.com>
Signed-off-by: David Wood <david@davidtw.co>
Doc: String isn't a collection
On forums one user was confused by this text, interpreting it as saying that `String` is a `Vec<char>` literally, rather than figuratively for the purpose of collect. I've reworded that paragraph.
Switch to intra-doc links in `core::option`
Part of #75080.
I didn't change some of the links since they link into `std` and you can't link from `core` to `std` (#74481).
Also, at least one other link can't be switched to an intra-doc link because it's not supported yet (#74489).
Constified str::from_utf8_unchecked
This would be useful for const code to use an array to construct a string using guaranteed utf8 inputs, and then create a `&str` from it.
Switch from indexing to zip, and also use `write` on `MaybeUninit`.
Add array_map feature to core/src/lib
Attempt to fix issue of no such feature
Update w/ pickfire's review
This changes a couple of names around, adds another small test of variable size,
and hides the rustdoc #![feature(..)].
Fmt doctest
Add suggestions from lcnr
Add basic test
And also run fmt which is where the other changes are from
Fix mut issues
These only appear when running tests, so resolved by adding mut
Swap order of forget
Add pub and rm guard impl
Add explicit type to guard
Add safety note
Change guard type from T to S
It should never have been T, as it guards over [MaybeUninit<S>; N]
Also add feature to test
This creates the language item for arrays, and adds the map fn which is like map in options or
iterators. It currently allocates an extra array, unfortunately.
Added fixme for transmuting
Fix typo
Add drop guard
The previous `assert_eq` generated quite some code, which is especially
problematic when this call is inlined. This commit also slightly
improves the panic message from:
assertion failed: `(left == right)`
left: `3`,
right: `2`: destination and source slices have different lengths
...to:
source slice length (2) does not match destination slice length (3)
Use intra-doc links in /library/core/src/cmp.rs
Helps with #75080.
@rustbot modify labels: T-doc, A-intra-doc-links, T-rustdoc
Known issues:
* Links from `core` to `std` (#74481):
* [`Vec::sort_by_key`]
Rollup of 15 pull requests
Successful merges:
- #74712 (Update E0271 explanation)
- #74842 (adjust remaining targets)
- #75151 (Consistent variable name alloc for raw_vec)
- #75162 (Fix the documentation for move about Fn traits implementations)
- #75248 (Add `as_mut_ptr` to `NonNull<[T]>`)
- #75262 (Show multi extension example for Path in doctests)
- #75266 (Add safety section to `NonNull::as_*` method docs)
- #75284 (Show relative example for Path ancestors)
- #75285 (Separate example for Path strip_prefix)
- #75287 (Show Path extension example change multi extension)
- #75288 (Use assert! for Path exists example to check bool)
- #75289 (Remove ambiguity from PathBuf pop example)
- #75290 (fix `min_const_generics` version)
- #75291 (Clean up E0750)
- #75292 (Clean up E0502)
Failed merges:
r? @ghost
Add safety section to `NonNull::as_*` method docs
This basically adds the safety section of `*mut T::as_{ref,mut}` to the
same methods on `NonNull` with minor modifications to fit the
differences.
Part of #48929.
Simplify array::IntoIter
- Initialization can use `transmute_copy` to do the bitwise copy.
- `as_slice` can use `get_unchecked` and `MaybeUninit::slice_get_ref`,
and `as_mut_slice` can do similar.
- `next` and `next_back` can use the corresponding `Range` methods.
- `Clone` doesn't need any unsafety, and we can dynamically update the
new range to get partial drops if `T::clone` panics.
r? @LukasKalbertodt
This basically adds the safety section of `*mut T::as_{ref,mut}` to the
same methods on `NonNull` with minor modifications to fit the
differences.
Part of #48929.
- Initialization can use `transmute_copy` to do the bitwise copy.
- `as_slice` can use `get_unchecked` and `MaybeUninit::slice_get_ref`,
and `as_mut_slice` can do similar.
- `next` and `next_back` can use the corresponding `Range` methods.
- `Clone` doesn't need any unsafety, and we can dynamically update the
new range to get partial drops if `T::clone` panics.
Rollup of 4 pull requests
Successful merges:
- #74774 (adds [*mut|*const] ptr::set_ptr_value)
- #75079 (Disallow linking to items with a mismatched disambiguator)
- #75203 (Make `IntoIterator` lifetime bounds of `&BTreeMap` match with `&HashMap` )
- #75227 (Fix ICE when using asm! on an unsupported architecture)
Failed merges:
r? @ghost
adds [*mut|*const] ptr::set_ptr_value
I propose the addition of these two functions to `*mut T` and `*const T`, respectively. The motivation for this is primarily byte-wise pointer arithmetic on (potentially) fat pointers, i.e. for types with a `T: ?Sized` bound. A concrete use-case has been discussed in [this](https://internals.rust-lang.org/t/byte-wise-fat-pointer-arithmetic/12739) thread.
TL;DR: Currently, byte-wise pointer arithmetic with potentially fat pointers in not possible in either stable or nightly Rust without making assumptions about the layout of fat pointers, which is currently still an implementation detail and not formally stabilized. This PR adds one function to `*mut T` and `*const T` each, allowing to circumvent this restriction without exposing any internal implementation details.
One possible alternative would be to add specific byte-wise pointer arithmetic functions to the two pointer types in addition to the already existing count-wise functions. However, I feel this fairly niche use case does not warrant adding a whole set of new functions like `add_bytes`, `offset_bytes`, `wrapping_offset_bytes`, etc. (times two, one for each pointer type) to `libcore`.
Completes support for coverage in external crates
Follow-up to #74959 :
The prior PR corrected for errors encountered when trying to generate
the coverage map on source code inlined from external crates (including
macros and generics) by avoiding adding external DefIds to the coverage
map.
This made it possible to generate a coverage report including external
crates, but the external crate coverage was incomplete (did not include
coverage for the DefIds that were eliminated.
The root issue was that the coverage map was converting Span locations
to source file and locations, using the SourceMap for the current crate,
and this would not work for spans from external crates (compliled with a
different SourceMap).
The solution was to convert the Spans to filename and location during
MIR generation instead, so precompiled external crates would already
have the correct source code locations embedded in their MIR, when
imported into another crate.
@wesleywiser FYI
r? @tmandry
The prior PR corrected for errors encountered when trying to generate
the coverage map on source code inlined from external crates (including
macros and generics) by avoiding adding external DefIds to the coverage
map.
This made it possible to generate a coverage report including external
crates, but the external crate coverage was incomplete (did not include
coverage for the DefIds that were eliminated.
The root issue was that the coverage map was converting Span locations
to source file and locations, using the SourceMap for the current crate,
and this would not work for spans from external crates (compliled with a
different SourceMap).
The solution was to convert the Spans to filename and location during
MIR generation instead, so precompiled external crates would already
have the correct source code locations embedded in their MIR, when
imported into another crate.
add `unsigned_abs` to signed integers
Mentioned on rust-lang/rfcs#2914
This PR simply adds an `unsigned_abs` to signed integers function which returns the correct absolute value as a unsigned integer.
Stabilize `Result::as_deref` and `as_deref_mut`
FCP completed in https://github.com/rust-lang/rust/issues/50264#issuecomment-645681400.
This PR stabilizes two new APIs for `std::result::Result`:
```rust
fn as_deref(&self) -> Result<&T::Target, &E> where T: Deref;
fn as_deref_mut(&mut self) -> Result<&mut T::Target, &mut E> where T: DerefMut;
```
This PR also removes two rarely used unstable APIs from `Result`:
```rust
fn as_deref_err(&self) -> Result<&T, &E::Target> where E: Deref;
fn as_deref_mut_err(&mut self) -> Result<&mut T, &mut E::Target> where E: DerefMut;
```
Closes#50264
In the current documentation about the `Copy` marker trait, there is a section
about "additional implementors", which list additional implementors of the `Copy` trait.
The fact that shared references are also `Copy` is mixed with another point,
which makes it hard to recognize and make it seem not as important.
This clarifies the fact that shared references are also `Copy`, by mentioning it as a
separate item in the list of "additional implementors".
add `slice::array_chunks` to std
Now that #74113 has landed, these methods are suddenly usable. A rebirth of #72334
Tests are directly copied from `chunks_exact` and some additional tests for type inference.
r? @withoutboats as you are both part of t-libs and working on const generics. closes#60735
Make `Option::unwrap` unstably const
This is lumped into the `const_option` feature gate (#67441), which enables a potpourri of `Option` methods.
cc @rust-lang/wg-const-eval
r? @oli-obk
`Result::unwrap` is not eligible becuase it formats the contents of the
`Err` variant. `unwrap_or`, `unwrap_or_else` and friends are not
eligible because they drop things or invoke closures.
Make `mem::size_of_val` and `mem::align_of_val` unstably const
Implements #46571 but does not stabilize it. I wanted this while working on something today.
The only reason not to immediately stabilize are concerns around [custom DSTs](https://github.com/rust-lang/rust/issues/46571#issuecomment-387669352). That proposal has made zero progress in the last two years and const eval is rich enough to support pretty much any user-defined `len` function as long as nightly features are allowed (`raw_ptr_deref`).
Currently, this raises a `const_err` lint when passed an `extern type`.
r? @oli-obk
cc @rust-lang/wg-const-eval
Stabilize const_type_id feature
The tracking issue for `const_type_id` points to the ill-fated #41875. So I'm re-energizing `TypeId` shenanigans by opening this one up to see if there's anything blocking us from stabilizing the constification of type ids.
Will wait for CI before pinging teams/groups.
-----
This PR stabilizes the `const_type_id` feature, which allows `TypeId::of` (and the underlying unstable intrinsic) to be called in constant contexts.
There are some [sanity tests](https://github.com/rust-lang/rust/blob/master/src/test/ui/consts/const-typeid-of-rpass.rs) that demonstrate its usage, but I’ve included some more below.
As a simple example, you could create a constant item that contains some type ids:
```rust
use std::any::TypeId;
const TYPE_IDS: [TypeId; 2] = [
TypeId::of::<u32>(),
TypeId::of::<i32>(),
];
assert_eq!(TypeId::of::<u32>(), TYPE_IDS[0]);
```
Type ids can also now appear in associated constants. You could create a trait that associates each type with its constant type id:
```rust
trait Any where Self: 'static {
const TYPE_ID: TypeId = TypeId::of::<Self>();
}
impl<T: 'static> Any for T { }
assert_eq!(TypeId::of::<usize>(), usize::TYPE_ID);
```
`TypeId::of` is generic, which we saw above in the way the generic `Self` argument was used. This has some implications for const evaluation. It means we can make trait impls evaluate differently depending on information that wasn't directly passed through the trait system. This violates the _parametricity_ property, which requires all instances of a generic function to behave the same way with respect to its generic parameters. That's not unique to `TypeId::of`, other generic const functions based on compiler intrinsics like `mem::align_of` can also violate parametricity. In practice Rust doesn't really have type parametricity anyway since it monomorphizes generics into concrete functions, so violating it using type ids isn’t new.
As an example of how impls can behave differently, you could combine constant type ids with the `const_if_match` feature to dispatch calls based on the type id of the generic `Self`, rather than based on information about `Self` that was threaded through trait bounds. It's like a rough-and-ready form of specialization:
```rust
#![feature(const_if_match)]
trait Specialized where Self: 'static {
// An associated constant that determines the function to call
// at compile-time based on `TypeId::of::<Self>`.
const CALL: fn(&Self) = {
const USIZE: TypeId = TypeId::of::<usize>();
match TypeId::of::<Self>() {
// Use a closure for `usize` that transmutes the generic `Self` to
// a concrete `usize` and dispatches to `Self::usize`.
USIZE => |x| Self::usize(unsafe { &*(x as *const Self as *const usize) }),
// For other types, dispatch to the generic `Self::default`.
_ => Self::default,
}
};
fn call(&self) {
// Call the function we determined at compile-time
(Self::CALL)(self)
}
fn default(x: &Self);
fn usize(x: &usize);
}
// Implement our `Specialized` trait for any `Debug` type.
impl<T: fmt::Debug + 'static> Specialized for T {
fn default(x: &Self) {
println!("default: {:?}", x);
}
fn usize(x: &usize) {
println!("usize: {:?}", x);
}
}
// Will print "usize: 42"
Specialized::call(&42usize);
// Will print "default: ()"
Specialized::call(&());
```
Type ids have some edges that this stabilization exposes to more contexts. It's possible for type ids to collide (but this is a bug). Since they can change between compiler versions, it's never valid to cast a type id to its underlying value.
