Optimize catch_unwind to match C++ try/catch
This refactors the implementation of catching unwinds to allow LLVM to inline the "try" closure directly into the happy path, avoiding indirection. This means that the catch_unwind implementation is (after this PR) zero-cost unless a panic is thrown.
https://rust.godbolt.org/z/cZcUSB is an example of the current codegen in a simple case. Notably, the codegen is *exactly the same* if `-Cpanic=abort` is passed, which is clearly not great.
This PR, on the other hand, generates the following assembly:
```asm
# -Cpanic=unwind:
push rbx
mov ebx,0x2a
call QWORD PTR [rip+0x1c53c] # <happy>
mov eax,ebx
pop rbx
ret
mov rdi,rax
call QWORD PTR [rip+0x1c537] # cleanup function call
call QWORD PTR [rip+0x1c539] # <unfortunate>
mov ebx,0xd
mov eax,ebx
pop rbx
ret
# -Cpanic=abort:
push rax
call QWORD PTR [rip+0x20a1] # <happy>
mov eax,0x2a
pop rcx
ret
```
Fixes#64224, and resolves#64222.
Implement Error for TryReserveError
I noticed that the Error trait wasn't implemented for TryReserveError. (#48043)
Not sure if the error messages and code style are 100% correct, it's my first time contributing to the Rust std.
Allow ZSTs in `AllocRef`
Allows ZSTs in all `AllocRef` methods. The implementation of `AllocRef` for `Global` and `System` were adjusted to reflect those changes.
This is the second item on the roadmap to support ZSTs in `AllocRef`: https://github.com/rust-lang/wg-allocators/issues/38#issuecomment-595861542
After this has landed, I will adapt `RawVec`, but since this will be a pretty big overhaul, it makes sense to do a different PR for it.
~~Requires #69794 to land first~~
r? @Amanieu
Remove spotlight
I had a few comments saying that this feature was at best misunderstood or not even used so I decided to organize a poll about on [twitter](https://twitter.com/imperioworld_/status/1232769353503956994). After 87 votes, the result is very clear: it's not useful. Considering the amount of code we have just to run it, I think it's definitely worth it to remove it.
r? @kinnison
cc @ollie27
Although `stack_overflow::init` runs very early in the process, even
before `main`, there may already be signal handlers installed for things
like the address sanitizer. In that case, just leave it alone, and don't
bother trying to allocate our own signal stacks either.
even more clippy cleanups
* Don't pass &mut where immutable reference (&) is sufficient (clippy::unnecessary_mut_passed)
* Use more efficient &&str to String conversion (clippy::inefficient_to_string)
* Don't always eval arguments inside .expect(), use unwrap_or_else and closure. (clippy::expect_fun_call)
* Use righthand '&' instead of lefthand "ref". (clippy::toplevel_ref_arg)
* Use simple 'for i in x' loops instead of 'while let Some(i) = x.next()' loops on iterators. (clippy::while_let_on_iterator)
* Const items have by default a static lifetime, there's no need to annotate it. (clippy::redundant_static_lifetimes)
* Remove redundant patterns when matching ( x @ _ to x) (clippy::redundant_pattern)
Stabilize assoc_int_consts associated int/float constants
The next step in RFC https://github.com/rust-lang/rfcs/pull/2700 (tracking issue #68490). Stabilizing the associated constants that were added in #68325.
* Stabilize all constants under the `assoc_int_consts` feature flag.
* Update documentation on old constants to say they are soft-deprecated and the new ones should be preferred.
* Update documentation examples to use new constants.
* Remove `uint_macro` and use `int_macro` for all integer types since the macros were identical anyway.
r? @LukasKalbertodt
more cleanups
* use starts_with() instead of chars().next() == Some(x)
* use subsec_micros() instead of subsec_nanos() / 1000
* use for (idx, item) in iter.enumerate() instead of manually counting loop iterations with variables
* use values() or keys() respectively when iterating only over keys or values of maps.
Remove `usable_size` APIs
This removes the usable size APIs:
- remove `usable_size` (obv)
- change return type of allocating methods to include the allocated size
- remove `_excess` API
r? @Amanieu
closesrust-lang/wg-allocators#17
This allows LLVM to inline the happy path, such that catching unwinding is
zero-cost when no panic occurs. This also allows us to match the code generated
by C++ try/catch.
Correct comment to match behavior
Corrects the header comment on `saturating_duration_since` to match the behavior of returning 0 if the other timestamp is _later_ than the invocant, not earlier,
This is purely a documentation change, so hopefully it doesn't require an issue; if it does, I'll open one and resubmit.
