Implement LazyBTreeMap and use it in a few places.
This is a thin wrapper around BTreeMap that avoids allocating upon creation.
I would prefer to change BTreeMap directly to make it lazy (like I did with HashSet in #36734) and I initially attempted that by making BTreeMap::root an Option<>. But then I also had to change Iter and Range to handle trees with no root, and those types have stability markers on them and I wasn't sure if that was acceptable. Also, BTreeMap has a lot of complex code and changing it all was challenging, and I didn't have high confidence about my general approach.
So I prototyped this wrapper instead and used it in the hottest locations to get some measurements about the effect. The measurements are pretty good!
- Doing a debug build of serde, it reduces the total number of heap allocations from 17,728,709 to 13,359,384, a 25% reduction. The number of bytes allocated drops from 7,474,672,966 to 5,482,308,388, a 27% reduction.
- It gives speedups of up to 3.6% on some rustc-perf benchmark jobs. crates.io, futures, and serde benefit most.
```
futures-check
avg: -1.9% min: -3.6% max: -0.5%
serde-check
avg: -2.1% min: -3.5% max: -0.7%
crates.io-check
avg: -1.7% min: -3.5% max: -0.3%
serde
avg: -2.0% min: -3.0% max: -0.9%
serde-opt
avg: -1.8% min: -2.9% max: -0.3%
futures
avg: -1.5% min: -2.8% max: -0.4%
tokio-webpush-simple-check
avg: -1.1% min: -2.2% max: -0.1%
futures-opt
avg: -1.2% min: -2.1% max: -0.4%
piston-image-check
avg: -0.8% min: -1.1% max: -0.3%
crates.io
avg: -0.6% min: -1.0% max: -0.3%
```
@Gankro, how do you think I should proceed here? Is leaving this as a wrapper reasonable? Or should I try to make BTreeMap itself lazy? If so, can I change the representation of Iter and Range?
Thanks!
Replace manual iterator exhaust with for_each(drop)
This originally added a dedicated method, `Iterator::exhaust`, and has since been replaced with `for_each(drop)`, which is more idiomatic.
<del>This is just shorthand for `for _ in &mut self {}` or `while let Some(_) = self.next() {}`. This states the intent a lot more clearly than the identical code: run the iterator to completion.
<del>At least personally, my eyes tend to gloss over `for _ in &mut self {}` without fully paying attention to what it does; having a `Drop` implementation akin to:
<del>`for _ in &mut self {}; unsafe { free(self.ptr); }`</del>
<del>Is not as clear as:
<del>`self.exhaust(); unsafe { free(self.ptr); }`
<del>Additionally, I've seen debate over whether `while let Some(_) = self.next() {}` or `for _ in &mut self {}` is more clear, whereas `self.exhaust()` is clearer than both.
This makes sure that all bits in each IdxSet between the universe length
and the end of the word are all zero instead of being in an indeterminate state.
This fixes a crash with RUST_LOG=rustc_mir, and is probably a good idea
anyway.
Implement some trivial size_hints for various iterators
This also implements ExactSizeIterator where applicable.
Addresses most of the Iterator traits mentioned in #23708.
I intend to do more, but I don't want to make the PR too large.
Remove some unnecessary IdxSet methods
This replaces `IdxSet:: reset_to_empty` with `IdxSet:: clear`, and `IdxSet::elems`/`IdxSet::each_bit` with `IdxSet::iter`. Based on some [comments on #rustc](https://botbot.me/mozilla/rustc/2018-01-23/?msg=96063396).
r? @pnkfelix
Replace iterator structures with `impl Trait`.
Two commits:
* Replace iterator structures with `impl Trait`.
* Run rustfmt on `src/librustc_data_structures/graph/mod.rs`.
