bootstrap: Fix stack printing when a step cycle is detected
When bootstrap detects a step dependency cycle (which represents a bootstrap bug), it is supposed to print out the contents of the step stack as part of its panic message.
However, while investigating #138205 it was found that bootstrap was actually printing out several copies of `Any { .. }`, because that is the Debug implementation for `dyn Any`. This is sadly not very helpful.
This PR fixes that problem by introducing a `trait AnyDebug: Any + Debug` that delegates to the underlying type's Debug implementation, while still allowing downcasting via Any.
---
The fixed behaviour can be verified manually (and is tested automatically) via a new dummy command, `./x run cyclic-step`:
```
$ x run cyclic-step
Building bootstrap
Finished `dev` profile [unoptimized] target(s) in 0.02s
thread 'main' panicked at src/bootstrap/src/core/builder/mod.rs:1521:17:
Cycle in build detected when adding CyclicStep { n: 0 }
CyclicStep { n: 0 }
CyclicStep { n: 1 }
CyclicStep { n: 2 }
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
Build completed unsuccessfully in 0:00:00
```
depend more on attr_data_structures and move find_attr! there
r? ``@oli-obk``
This should be an easy one. It just moves some imports around. This is necessary for other changes that I'm working on not to have import cycles. However, it's an easy one to just merge on its own.
Move more layouting logic to `rustc_abi`
Move all `LayoutData`-constructing code to `rustc_abi`:
- Infaillible operations get a new `LayoutData` constructor method;
- Faillible ones get a new method on `LayoutCalculator`.
compiler: Use `size_of` from the prelude instead of imported
Use `std::mem::{size_of, size_of_val, align_of, align_of_val}` from the prelude instead of importing or qualifying them. Apply this change across the compiler.
These functions were added to all preludes in Rust 1.80.
r? ``@compiler-errors``
Change TaskDeps to start preallocated with 128 capacity
This is a tiny change that makes `TaskDeps::read_set` start preallocated with capacity for 128 elements.
From local profiling, it looks like `TaskDeps::read_set` is one of the most-often resized hash-sets in `rustc`.
add `tool::CargoClippy` and `tool::Cargofmt` binary to target sysroot
When running `x build clippy`, we expect `stage1-tool-bin/cargo-clippy` and `stage2/bin/cargo-clippy` to be the same, but they aren't. This happens because `tool::CargoClippy` doesn't place its binary in the `stage2` directory. As a result, `stage1-tool-bin/cargo-clippy` comes from `tool::CargoClippy`, while `stage2/bin/cargo-clippy` comes from `tool::Cargo`. Same applies for `tool::Cargofmt`.
This PR fixes the issue by adding `tool::CargoClippy` and ``tool::Cargofmt`` binaries to the expected sysroot and makes sure both directories share the same binary.
To test this, run `x build --stage 2 compiler clippy rustfmt`, link the stage2 sysroot with rustup, and then call `cargo +stage2 fmt` and `cargo +stage2 clippy` on any rust project (it wouldn't work without this PR).
Don't re-`assume` in `transmute`s that don't change niches
I noticed in nightly 2025-02-21 that `transmute` is emitting way more `assume`s than necessary for newtypes.
For example, the three transmutes in <https://rust.godbolt.org/z/fW1KaTc4o> emits
```rust
define noundef range(i32 1, 0) i32 `@repeatedly_transparent_transmute(i32` noundef range(i32 1, 0) %_1) unnamed_addr {
start:
%0 = sub i32 %_1, 1
%1 = icmp ule i32 %0, -2
call void `@llvm.assume(i1` %1)
%2 = sub i32 %_1, 1
%3 = icmp ule i32 %2, -2
call void `@llvm.assume(i1` %3)
%4 = sub i32 %_1, 1
%5 = icmp ule i32 %4, -2
call void `@llvm.assume(i1` %5)
%6 = sub i32 %_1, 1
%7 = icmp ule i32 %6, -2
call void `@llvm.assume(i1` %7)
%8 = sub i32 %_1, 1
%9 = icmp ule i32 %8, -2
call void `@llvm.assume(i1` %9)
%10 = sub i32 %_1, 1
%11 = icmp ule i32 %10, -2
call void `@llvm.assume(i1` %11)
ret i32 %_1
}
```
But those are all just newtypes that don't change size or niches, so none of it's needed.
After this PR it's down to just
```rust
define noundef range(i32 1, 0) i32 `@repeatedly_transparent_transmute(i32` noundef range(i32 1, 0) %_1) unnamed_addr {
start:
ret i32 %_1
}
```
because none of those `assume`s in the original actually did anything.
(Transmuting to something with a difference niche, though, still has the assumes -- the other tests continue to pass checking that.)
Don't include global asm in `mir_keys`, fix error body synthesis
r? oli-obk
Fixes#137470Fixes#137471Fixes#137472Fixes#137473
try-job: test-various
try-job: x86_64-apple-2
skip `compile::Std` and `compile::Rustc` on `forced_compiler`
Fixes https://github.com/rust-lang/rust/issues/138220
Set `download-rustc = true/false` and run `x test tests/ui/meta/no_std-extern-libc.rs --force-rerun` to debug it.
this prevents us from trying unsizing coercion in cases like
`*const W<dyn T>` -> `*const dyn T`, where it would later cause a
compilation error since `W<dyn T>: Sized` and `W<dyn T>: T` do not hold.
Rollup of 5 pull requests
Successful merges:
- #136642 (Put the alloc unit tests in a separate alloctests package)
- #137528 (Windows: Fix error in `fs::rename` on Windows 1607)
- #137685 (self-contained linker: conservatively default to `-znostart-stop-gc` on x64 linux)
- #137757 (On long spans, trim the middle of them to make them fit in the terminal width)
- #138189 (Mention `env` and `option_env` macros in `std::env::var` docs)
r? `@ghost`
`@rustbot` modify labels: rollup
Mention `env` and `option_env` macros in `std::env::var` docs
Fixes https://github.com/rust-lang/rust/issues/138159.
Just like there are mentions in `env!` and `option_env!` docs to `std::env::var`, it'd be nice to have a "mention back" as well.
On long spans, trim the middle of them to make them fit in the terminal width
When encountering a single line span that is wider than the terminal, we keep context at the start and end of the span but otherwise remove the code from the middle. This is somewhat independent from whether the left and right margins of the output have been trimmed as well.
```
error[E0308]: mismatched types
--> $DIR/long-span.rs:6:15
|
LL | ... = [0, 0, 0, 0, ..., 0, 0];
| ^^^^^^^^^^^^^...^^^^^^^ expected `u8`, found `[{integer}; 1681]`
```
Address part of https://github.com/rust-lang/rust/issues/137680 (missing handling of the long suggestion). Fix https://github.com/rust-lang/rust/issues/125581.
---
Change the way that underline positions are calculated by delaying using the "visual" column position until the last possible moment, instead using the "file"/byte position in the file, and then calculating visual positioning as late as possible. This should make the underlines more resilient to non-1-width unicode chars.
Unfortunately, as part of this change (which fixes some visual bugs) comes with the loss of some eager tab codepoint handling, but the output remains legible despite some minor regression on the "margin trimming" logic.
---
`-Zteach` is perma-unstable, barely used, the highlighting logic buggy and the flag being passed around is tech-debt. We should likely remove `-Zteach` in its entirely.
self-contained linker: conservatively default to `-znostart-stop-gc` on x64 linux
To help stabilization, this PR disables an LLD optimization on x64 linux with respect to `--gc-sections` and encapsulation symbols: it will reduce the number of crates needing to opt-out of lld due to this bfd / lld difference. For example, all the people using [linkme](https://github.com/dtolnay/linkme), which [doesn't work with lld](https://github.com/dtolnay/linkme/issues/63) or on nightly, need to disable lld.
More information about all this, and the historical differences, can be found in:
- https://maskray.me/blog/2021-01-31-metadata-sections-comdat-and-shf-link-order
- https://lld.llvm.org/ELF/start-stop-gc
This optimization has [no visible impact](https://github.com/rust-lang/rust/pull/137685#issuecomment-2686116312) on our benchmarks, so we can use it by default and have a safer/more conservative starting point to remove friction during migration. We can them emit an FCW for the cases where lld detects reliance on encapsulation symbols without `-znostart-stop-gc`, and then revert back to lld's default after a while. No one compiling on nightly relies on this difference, obviously, so doing an FCW is not necessary until after lld is used on stable.
I've tested that this correctly links on `linkme` examples. I've also quickly tried to crate an rmake test but the setup with encapsulation symbols is annoying to reproduce: a few link section/name attributes is not enough, we also need to collect symbols between the encapsulation symbols, without referencing them in code, for `-znostart-stop-gc` to only impact this... It should of course be doable though, maybe ````@Kobzol```` will look into it if they have time.
r? ````@petrochenkov````
