Currently, a coherence error based on overlapping impls simply mentions
the trait, and points to the two conflicting impls:
```
error: conflicting implementations for trait `Foo`
```
With this commit, the error will include all input types to the
trait (including the `Self` type) after unification between the
overlapping impls. In other words, the error message will provide
feedback with full type details, like:
```
error: conflicting implementations of trait `Foo<u32>` for type `u8`:
```
When the `Self` type for the two impls unify to an inference variable,
it is elided in the output, since "for type `_`" is just noise in that
case.
Closes#23980
This PR adds some safety checks to interning things in `ty::ctxt`. Accidentally re-interning an `AdtDef` has bitten me in the behind just last week (it messes up things in very subtle way only showing up later as an LLVM assertion).
Initially I had also added a check to `ty::ctxt::node_type_insert()` -- but there it seems to be expected that the same table slot is written to multiple times.
Roll-up candidate.
I believe that because Windows' unit of resolution is 100ns that this unit of
time will ensure that the assertions will hold true as it's representable in the
native format.
cc #29970
In #29932, I moved the location of TRPL, but I missed making the changes
in mk/tests.mk. This led to #30088 landing with a broken example.
As such, #30113 will need to land before this.
The `f` argument will reference the actual value in the `d` box, not the box in the `bar`'s stack frame.
I am just learning Rust, so I don't know how to explain this well, but just from `f`'s type it is clear that it will be a pointer to an `i32`, not a pointer to a pointer. Some `println!("{:p}", ...)`'s can easily confirm this.
I would actually suggest to remove/simplify this part of the example. This is a subtle issue that can easily confuse people at the early stages of familiarizing with the language. (As I got confused by it. :))
Fixes#30073. The input to `cfg!` is a meta attribute, but not _any_ meta attribute (e.g. `cfg!(allow(dead_code))` doesn't compile). But the macro_rules syntax can't quite express this, so I added a note to the doc.
In #29932, I moved the location of TRPL, but I missed making the changes
in mk/tests.mk. This led to #30088 landing with a broken example.
As such, #30113 will need to land before this.
I believe that because Windows' unit of resolution is 100ns that this unit of
time will ensure that the assertions will hold true as it's representable in the
native format.
cc #29970
I don't reproduce it on severals hosts (virtual or real), so I can't
debug it. As Bitrig has disabled this test too, I will follow the same
here.
r? @alexcrichton
