small iter.intersperse.fold() optimization
No need to call into fold when the first item is already None, this avoids some redundant work for empty iterators.
"But it uses Fuse" one might want to protest, but Fuse is specialized and may call into the inner iterator anyway.
Change `core::iter::Fuse`'s `Default` impl to do what its docs say it does
The [docs on `impl<I: Default> Default for core::iter::Fuse<I>`](https://doc.rust-lang.org/nightly/std/iter/struct.Fuse.html#impl-Default-for-Fuse%3CI%3E) say (as the `I: Default` bound implies) that `Fuse::<I>::default` "Creates a `Fuse` iterator from the default value of `I`". However, the implementation creates a `Fuse` with `Fuse { iter: Default::default() }`, and since the `iter` field is an `Option<I>`, this is actually `Fuse { iter: None }`, not `Fuse { iter: Some(I::default()) }`, so `Fuse::<I>::default()` always returns an empty iterator, even if `I::default()` would not be empty.
This PR changes `Fuse`'s `Default` implementation to match the documentation. This will be a behavior change for anyone currently using `Fuse::<I>::default()` where `I::default()` is not an empty iterator[^1], as `Fuse::<I>::default()` will now also not be an empty iterator.
(Alternately, the docs could be updated to reflect what the current implementation actually does, i.e. returns an always-exhausted iterator that never yields any items (even if `I::default()` would have yielded items). With this option, the `I: Default` bound could also be removed to reflect that no `I` is ever created.)
[Current behavior example](https://play.rust-lang.org/?version=stable&mode=debug&edition=2024&gist=a1e0adc4badca3dc11bfb70a99213249) (maybe an example like this should be added to the docs either way?)
This PR changes publicly observable behavior, so I think requires at least a T-libs-api FCP?
r? libs-api
cc https://github.com/rust-lang/rust/issues/140961
`impl<I: Default> Default for Fuse<I>` was added in 1.70.0 (https://github.com/rust-lang/rust/pull/99929), and it's docs and behavior do not appear to have changed since (`Fuse`'s `iter` field has been an `Option` since before the impl was added).
[^1]: IIUC it is a "de facto" guideline for the stdlib that an iterator type's `default()` should be empty (and for iterators where that would not make sense, they should not implement `Default`): cc https://github.com/rust-lang/libs-team/issues/77#issuecomment-1194681709 , so for stdlib iterators, I don't think this would change anything. However, if a user has a custom `Iterator` type `I`, *and* they are using `Fuse<I>`, *and* they call `Fuse::<I>::default()`, this may change the behavior of their code.
No need to call into fold when the first item is already None,
this avoids some redundant work for empty iterators.
"But it uses Fuse" one might want to protest, but Fuse is specialized
and may call into the inner iterator anyway.
This adds an `iter!` macro that can be used to create movable
generators.
This also adds a yield_expr feature so the `yield` keyword can be used
within iter! macro bodies. This was needed because several unstable
features each need `yield` expressions, so this allows us to stabilize
them separately from any individual feature.
Co-authored-by: Oli Scherer <github35764891676564198441@oli-obk.de>
Co-authored-by: Jieyou Xu <jieyouxu@outlook.com>
Co-authored-by: Travis Cross <tc@traviscross.com>
Implement `advance_by` via `try_fold` for `Sized` iterators
When `try_fold` is overriden, it is usually easier for compilers to optimize.
Example difference: https://iter.godbolt.org/z/z8cEfnKro
docs: Add example to `Iterator::take` with `by_ref`
If you want to logically split an iterator after `n` items, you might first discover `take`. Before this change, you'd find that `take` consumes the iterator, and you'd probably be stuck. The answer involves `by_ref`, but that's hard to discover, especially since `by_ref` is a bit abstract and `Iterator` has many methods.
After this change, you'd see the example showing `take` along with `by_ref`, which allows you to continue using the rest of the iterator. `by_ref` had a good example involving `take` already, so this change just duplicates that existing example under `take`.
If you want to logically split an iterator after `n` items, you might first
discover `take`. Before this change, you'd find that `take` consumes the
iterator, and you'd probably be stuck. The answer involves `by_ref`, but that's
hard to discover, especially since `by_ref` is a bit abstract and `Iterator`
has many methods.
After this change, you'd see the example showing `take` along with `by_ref`,
which allows you to continue using the rest of the iterator. `by_ref` had a
good example involving `take` already, so this change just duplicates that
existing example under `take`.
Avoid cloning in `Cloned<I>` or copying in `Copied<I>` when elements are
only needed by reference or not at all. There is already some precedent
for this, given that `__iterator_get_unchecked` is implemented, which
can skip elements. The reduced clones are technically observable by a
user impl of `Clone`.
...by adding an optimized implementation of `try_fold` and `fold` as well as replacing some unnecessary `mem::replace` calls with `MaybeUninit` helper methods.
debug-assert that the size_hint is well-formed in `collect`
Closes#137919
In the hopes of helping to catch any future accidentally-incorrect rustc or stdlib iterators (like the ones #137908 accidentally found), this has `Iterator::collect` call `size_hint` and check its `low` doesn't exceed its `Some(high)`.
There's of course a bazillion more places this *could* be checked, but the hope is that this one is a good tradeoff of being likely to catch lots of things while having minimal maintenance cost (especially compared to putting it in *every* container's `from_iter`).
Remove `#[cfg(not(test))]` gates in `core`
These gates are unnecessary now that unit tests for `core` are in a separate package, `coretests`, instead of in the same files as the source code. They previously prevented the two `core` versions from conflicting with each other.
library: 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.
These functions were added to all preludes in Rust 1.80.
try-job: test-various
try-job: x86_64-gnu
try-job: x86_64-msvc-1
Use `std::mem::{size_of, size_of_val, align_of, align_of_val}` from the
prelude instead of importing or qualifying them.
These functions were added to all preludes in Rust 1.80.
These gates are unnecessary now that unit tests for `core` are in a
separate package, `coretests`, instead of in the same files as the
source code. They previously prevented the two `core` versions from
conflicting with each other.
Document workings of successors more clearly
This is an attempt to fix#135087 together with https://github.com/rust-lang/rust/pull/135886, but I am not sure if I've succeeded in adding much clarity here, so don't be shy with your comments.
Because the neutral element of `<fNN as iter::Sum>` was changed to
`neg_zero`, the documentation needed to be updated, as it was reporting
inadequate information about what should be expected from the return.
Co-authored-by: Jubilee <workingjubilee@gmail.com>
