Add iterator method .rfold(init, function); the reverse of fold
rfold is the reverse version of fold.
Fold allows iterators to implement a different (non-resumable) internal
iteration when it is more efficient than the external iteration implemented
through the next method. (Common examples are VecDeque and .chain()).
Introduce rfold() so that the same customization is available for reverse
iteration. This is achieved by both adding the method, and by having the
Rev\<I> adaptor connect Rev::rfold → I::fold and Rev::fold → I::rfold.
On the surface, rfold(..) is just .rev().fold(..), but the special case
implementations allow a data structure specific fold to be used through for
example .iter().rev(); we thus have gains even for users never calling exactly
rfold themselves.
Optimize drain_filter
This PR cuts out two copies from each iteration of `drain_filter` by exchanging the swap operation for a copy_nonoverlapping function call instead. Since the data being swapped is not needed anymore we can just overwrite it instead.
Implement <Rc<Any>>::downcast
* Implement `<Rc<Any>>::downcast::<T>`
* New unstable method. Works just like Box\<Any\>, but for Rc.
* Any has two cases for its methods: Any and Any + Send; Rc is never Send, so that case is skipped for Rc.
* Motivation for being a method with self is to match Box and there is no user-supplied type; the inner type is Any and downcast does not conflict with any method of Any.
* Arc was skipped because Any itself has no downcast for the case that makes most sense: Any + Send + Sync
Fix example in transmute; add safety requirement to Vec::from_raw_parts
This fixes the second bullet point on #44281 and also removes some incorrect information.
Fix drain_filter doctest.
Fixes#44499.
Also change some of the hidden logic in the doctest as a regression test; two bugs in the original would now cause test failure.
impl Hasher for {&mut Hasher, Box<Hasher>}
**Rationale:** The `Hash` trait has `fn hash<H: Hasher>(&self, state: &mut H)`, which can only accept a `Sized` hasher, even if the `Hasher` trait is object-safe. We cannot retroactively add the `?Sized` bound without breaking stability, thus implementing `Hasher` to a trait object reference is the next best solution.
**Warning:** These `impl` are insta-stable, and should need an FCP. I don't think a full RFC is necessary.
Add or_default to Entry APIs
As argued for in #44324, this PR adds a new `or_default` method to the various `Entry` APIs (currently just for `BTreeMap` and `HashMap`) when `V: Default`. This method is effectively a shorthand for `or_insert_with(Default::default)`.
Implement downcast the like it exists for Box.
The implementation avoids using into_raw/from_raw, because the pointer
arithmetic which should cancel does not seem to optimize out at the
moment.
Since Rc<T> is never Send, only Rc<Any> and not Rc<Any + Send>
implements downcast.
std: Mark allocation functions as nounwind
This commit flags all allocation-related functions in liballoc as "this can't
unwind" which should largely resolve the size-related issues found on #42808.
The documentation on the trait was updated with such a restriction (they can't
panic) as well as some other words about the relative instability about
implementing a bullet-proof allocator.
Closes#42808
