x.py sets it unconditionally, so want it for plain "cargo build".
We need to load one of the panic runtimes that is in src (vs. pre-built in the
compiler's sysroot) to ensure that we don't load libpanic_unwind from the
sysroot. That would lead to a load of libcore, also from the sysroot, and create
lots of errors about duplicate lang items.
The code was broken because it printed "llvm-config" instead of the
absolute path to the llvm-config executable, causing Cargo to always
rebuild librustc_llvm if using system LLVM.
Also, it's not the build system's job to rebuild when a system library
changes, so we simply don't emit "rerun-if-changed" if a path to LLVM
was not explicitly provided.
Rollup of 5 pull requests
Successful merges:
- #70644 (Clean up `ModuleConfig` initialization)
- #70937 (Fix staticlib name for *-pc-windows-gnu targets)
- #70996 (Add or_insert_with_key to Entry of HashMap/BTreeMap)
- #71020 (Store UNICODE_VERSION as a tuple)
- #71021 (Use write!-style syntax for MIR assert terminator)
Failed merges:
r? @ghost
Store UNICODE_VERSION as a tuple
Remove the UnicodeVersion struct containing
major, minor and update fields and replace it with
a 3-tuple containing the version number.
As the value of each field is limited to 255
use u8 to store them.
Add or_insert_with_key to Entry of HashMap/BTreeMap
Going along with `or_insert_with`, `or_insert_with_key` provides the `Entry`'s key to the lambda, avoiding the need to either clone the key or the need to reimplement this body of this method from scratch each time.
This is useful when the initial value for a map entry is derived from the key. For example, the introductory Rust book has an example Cacher struct that takes an expensive-to-compute lambda and then can, given an argument to the lambda, produce either the cached result or execute the lambda.
---
I'm fairly new to Rust, so any optimizations, corrections to types, better names, better documentation, or whatever else would be appreciated. I'd like to thank Arnavion on freenode for helping me to implement a very similar method when I found that `or_insert_with_key` was unavailable.
As a somewhat-related note, this implements https://github.com/rust-lang/rfcs/issues/1202 from 2015, so if this pull request is accepted, that should be closed.
Rollup of 5 pull requests
Successful merges:
- #69573 (tests encoding current behavior for various cases of "binding" to _.)
- #70881 (bootstrap: work around "unused attribute" errors in incremental stdlib rebuilds.)
- #70957 (Normalize MIR locals' types for generator layout computation.)
- #70962 (added machine hooks to track deallocations)
- #70982 (Normalize function signature in function casting check procedure)
Failed merges:
r? @ghost
Remove the UnicodeVersion struct containing
major, minor and update fields and replace it with
a 3-tuple containing the version number.
As the value of each field is limited to 255
use u8 to store them.
tests encoding current behavior for various cases of "binding" to _.
The `_` binding form is special, in that it encodes a "no-op": nothing is actually bound, and thus nothing is moved or borrowed in this scenario. Usually we do the "right" thing in all such cases. The exceptions are explicitly pointed out in this test case, so that we keep track of whether they are eventually fixed.
Cc #53114.
(This does not close the aforementioned issue; it just adds the tests encoding the current behavior, which we hope to eventually fix.)
Normalize function signature in function casting check procedure
Fixes#54094
```rust
trait Zoo {
type X;
}
impl Zoo for u16 {
type X = usize;
}
fn foo(abc: <u16 as Zoo>::X) {}
fn main() {
let x: *const u8 = foo as _;
}
```
Currently a `FnDef` need to be checked if it's able to cast to `FnPtr` before it is actually casted. But the signature of `FnPtr` target's associated types are not normalized:
96d77f0e5f/src/librustc_typeck/check/cast.rs (L536-L553)
However, during the coercion check, the signature of `FnPtr` target's associated types are normalized (The `<u16 as Zoo>::X` turns into `usize`).
96d77f0e5f/src/librustc_typeck/check/coercion.rs (L687-L729)
This inconsistency leads to the error:`Err(Sorts(ExpectedFound { expected: <u16 as Zoo>::X, found: usize }))`.
added machine hooks to track deallocations
This is part of rust-lang/miri#1314 in order to allow miri to show stack traces for on deallocation in order to debug use-after-free bugs
bootstrap: work around "unused attribute" errors in incremental stdlib rebuilds.
This should alleviate #58633 separately from a proper fix.
r? @Mark-Simulacrum
tests encoding current behavior for various cases of "binding" to _.
The `_` binding form is special, in that it encodes a "no-op": nothing is actually bound, and thus nothing is moved or borrowed in this scenario. Usually we do the "right" thing in all such cases. The exceptions are explicitly pointed out in this test case, so that we keep track of whether they are eventually fixed.
Cc #53114.
(This does not close the aforementioned issue; it just adds the tests encoding the current behavior, which we hope to eventually fix.)
