resolve: Relax shadowing restrictions on macro-expanded macros
Previously any macro-expanded macros weren't allowed to shadow macros from outer scopes.
Now only "more macro-expanded" macros cannot shadow "less macro-expanded" macros.
See comments to `fn may_appear_after` and added tests for more details and examples.
The functional changes are a21f6f588fc28c97533130ae44a6957b579ab58c and 46dd365ce9ca0a6b8653849b80267763c542842a, other commits are refactorings.
Stabilize edition 2018; also updates Clippy, RLS and Cargo
Supersedes https://github.com/rust-lang/rust/pull/53999 , https://github.com/rust-lang/rust/pull/53935
Clippy build was failing there because crate_visibility_modifier feature was taken out of edition 2018 and clippy used it.
The clippy update enables the corresponding feature explicitly.
r? @Mark-Simulacrum
proc_macro::Group::span_open and span_close
Before this addition, every delimited group like `(`...`)` `[`...`]` `{`...`}` has only a single Span that covers the full source location from opening delimiter to closing delimiter. This makes it impossible for a procedural macro to trigger an error pointing to just the opening or closing delimiter. The Rust compiler does not seem to have the same limitation:
```rust
mod m {
type T =
}
```
```console
error: expected type, found `}`
--> src/main.rs:3:1
|
3 | }
| ^
```
On that same input, a procedural macro would be forced to trigger the error on the last token inside the block, on the entire block, or on the next token after the block, none of which is really what you want for an error like above.
This commit adds `group.span_open()` and `group.span_close()` which access the Span associated with just the opening delimiter and just the closing delimiter of the group. Relevant to Syn as we implement real error messages for when parsing fails in a procedural macro: https://github.com/dtolnay/syn/issues/476.
```diff
impl Group {
fn span(&self) -> Span;
+ fn span_open(&self) -> Span;
+ fn span_close(&self) -> Span;
}
```
Fixes#48187
r? @alexcrichton
rustc_codegen_llvm: don't assume offsets are always aligned.
Fixes#53728 by taking into account not just overall type alignment and the field's alignment when determining whether a field is aligned or not ("packed"), but also the field's offset within the type.
Previously, rustc assumed that the offset was always at least as aligned as `min(struct.align, field.align)`. However, there's no real reason to have that assumption, and it obviously can't always be true after we implement `#[repr(align(N), pack(K))]`. There's also a case today where that assumption is not true, involving niche discriminants in enums:
Suppose that we have the code in #53728:
```Rust
#[repr(u16)]
enum DeviceKind {
Nil = 0,
}
#[repr(packed)]
struct DeviceInfo {
endianness: u8,
device_kind: DeviceKind,
}
struct Wrapper {
device_info: DeviceInfo,
data: u32
}
```
Observe the layout of `Option<Wrapper>`. It has an alignment of 4 because of the `u32`. `device_info.device_kind` is a good niche field to use, which means the enum ends up with this layout:
```
size = 8
align = 4
fields = [
{ offset=1, type=u16 } // discriminant, .<Some>.device_info.device_kind
]
```
And here we have an discriminant with alignment 2 (`u16`) but offset 1.
rustc_resolve: only prepend CrateRoot to a non-keyword segment.
Fixes#53770 by treating `use` paths as absolute in a finer-grained manner, specifically:
```rust
use {a, crate::b, self::c, super::d};
```
Used to be interpreted as if it were (when `uniform_paths` is not enabled):
```rust
use ::{a, crate::b, self::c, super::d};
```
With this PR, the `CrateRoot` pseudo-keyword indicating an absolute path is only inserted when the first path segment is found (if it's not a keyword), i.e. the example behaves like:
```rust
use {::a, crate::b, self::c, super::d};
```
This should (finally) make `use {path};` fully equivalent to `use path;`.
r? @petrochenkov cc @cramertj @joshtriplett @nikomatsakis
Optimize miri checking of integer array/slices
This pull request implements the optimization described in #53845 (the `E-easy` part of that issue, not the refactoring). Instead of checking every element of an integral array, we can check the whole memory range at once.
r? @RalfJung
Rollup of 10 pull requests
Successful merges:
- #53315 (use `NonZeroU32` in `newtype_index!`macro, change syntax)
- #53932 ([NLL] Remove base_place)
- #53942 (Rewrite `precompute_borrows_out_of_scope` for fewer hash table lookups.)
- #53973 (Have rust-lldb look for the rust-enabled lldb)
- #53981 (Implement initializer() for FileDesc)
- #53987 (rustbuild: allow configuring llvm version suffix)
- #53993 (rustc_resolve: don't record uniform_paths canaries as reexports.)
- #54007 (crates that provide a `panic_handler` are exempt from the `unused_extern_crates` lint)
- #54040 (update books for next release)
- #54050 (Update `petgraph` dependency to 0.4.13 to fix build with nightly)
use `NonZeroU32` in `newtype_index!`macro, change syntax
Various improvements to the `newtype_index!` macro:
- Use `NonZeroU32` so that `Option<T>` is cheap
- More ergonomic helper method, no need to import `Idx` trait all the time
- Improve syntax to use `struct` keyword so that ripgrep works to find type def'n
Fixes https://github.com/rust-lang/rust/issues/50337
I'm curious to see if this passes tests =)
Update `petgraph` dependency to 0.4.13 to fix build with nightly
I wanted to build Rust from source using a local nightly compiler, but I was unable to get `bootstrap` to compile due to a naming conflict with the `find_map` function.
This PR updates the `petgraph` dependency of `bootstrap` to 0.4.13, fixing the issue.
Implement initializer() for FileDesc
Here was my initial issue:
```rust
use std::process::{Command};
fn main() {
let output = Command::new("curl").arg("-s").arg("http://ovh.net/files/100Mio.dat").output();
println!("{:?}", output.unwrap().stdout.len());
}
```
```
~/stuff ❯❯❯ time ./dwl
104857600
./dwl 16.22s user 1.80s system 23% cpu 1:15.24 total
```
```rust
use std::process::{Command, Stdio};
fn main() {
let child = Command::new("curl").arg("-s").arg("http://ovh.net/files/100Mio.dat").stdout(Stdio::piped()).spawn();
let output = child.unwrap().wait_with_output().unwrap();
println!("{:?}", output.stdout.len());
}
```
```
~/stuff ❯❯❯ time ./dwl2
104857600
./dwl2 0.64s user 2.18s system 5% cpu 53.072 total
```
As you can see the first version is spending much more time in userland and also uses more cpu. With the help of @programble, @talchas and @habnabit on the rust IRC, we discovered that the slow version uses two pipes, one for `stdin` and one for `stderr` and in that case it polls when going through [this function](https://github.com/rust-lang/rust/blob/master/src/libstd/sys/unix/pipe.rs#L82). The polling calls `read_to_end` on the pipes repetitively and this results in zeroing its internal buffer each time. To avoid this zeroing, `FileDesc` needs to implement `initializer`. We see no reason why it [wouldn't work with uninitialized memory](https://doc.rust-lang.org/1.26.1/src/std/io/mod.rs.html#534) so this PR fixes that.
Here is some tracing of the slow program:
versus the fast program:
I have not tested the change yet but will try to build it tomorrow.
Have rust-lldb look for the rust-enabled lldb
We're shipping a rust-enabled lldb, but the "lldb" executable is not
installed into the "bin" directory by rustup. See the discussion in
https://github.com/rust-lang-nursery/rustup.rs/pull/1492 for
background on this decision. There, we agreed to have rust-lldb
prefer the rust-enabled lldb if it is installed. This patch changes
rust-lldb to look in the sysroot and use the lldb found there, if any.
See issue #48168
Rewrite `precompute_borrows_out_of_scope` for fewer hash table lookups.
It now does one hash table lookup per basic block, instead of one per
statement. This is worthwhile because this function is hot for NLL
builds of `ucd`.
I haven't measured the effect of this yet because I'm having trouble doing optimized builds of rustc that are suitable for profiling (#53916). I will do an online perf run instead.
r? @nikomatsakis
[NLL] Remove base_place
This function was supposed to make `Box` less special. But
* I think that the consensus is that MIR borrowck is going to fully special case `Box`
* It wasn't implemented correctly, it's looking at the type of the wrong `Place`, resulting in weird behaviour:
```rust
#![feature(nll)]
type A = Box<i32>; // If this is changed to another type then this will compile.
pub fn foo(x: Box<(String, A)>) {
let a = x.0; // This will compile if these lines are swapped
let b = x.1;
}
```
r? @nikomatsakis
stabilize #[panic_handler]
closes#44489
### Update(2018-09-07)
This was proposed for stabilization in https://github.com/rust-lang/rust/issues/44489#issuecomment-398965881 and its FCP with disposition to merge / accept is nearly over. The summary of what's being stabilized can be found in https://github.com/rust-lang/rust/issues/44489#issuecomment-416645946
Documentation PRs:
- Reference. https://github.com/rust-lang-nursery/reference/pull/362
- Nomicon. https://github.com/rust-lang-nursery/nomicon/pull/75
---
`#[panic_implementation]` was implemented recently in #50338. `#[panic_implementation]` is basically the old `panic_fmt` language item but in a less error prone (\*) shape. There are still some issues and questions to sort out around this feature (cf. #44489) but this PR is meant to start a discussion about those issues / questions with the language team.
(\*) `panic_fmt` was not type checked; changes in its function signature caused serious, silent binary size regressions like the one observed in #43054
Some unresolved questions from #44489:
> Should the Display of PanicInfo format the panic information as "panicked at 'reason',
> src/main.rs:27:4", as "'reason', src/main.rs:27:4", or simply as "reason".
The current implementation formats `PanicInfo` as the first alternative, which is how panic messages are formatted by the `std` panic handler. The `Display` implementation is more than a convenience: `PanicInfo.message` is unstable so it's not possible to replicate the `Display` implementation on stable.
> Is this design compatible, or can it be extended to work, with unwinding implementations for
> no-std environments?
I believe @whitequark made more progress with unwinding in no-std since their last comment in #44489. Perhaps they can give us an update?
---
Another unresolved question is where this feature should be documented. The feature currently doesn't have any documentation.
cc @rust-lang/lang
cc @jackpot51 @alevy @phil-opp
do not propagate closure requirements if we can prove them locally
Fixes#53570
cc @mikhail-m1 -- you're the one who last touched this code
r? @pnkfelix
This allows it to print out the "late-bound regions" from the closure
context more easily. Besides, all the state that is being printed it
is private to the `UniversalRegions`.
