Introduce `MixedBitSet`
`ChunkedBitSet` is good at avoiding excessive memory usage for programs with very large functgions where dataflow bitsets have very large domain sizes. But it's overly heavyweight for small bitsets, because any non-empty `ChunkedBitSet` takes up at least 256 bytes.
This PR introduces `MixedBitSet`, which is a simple bitset that uses `BitSet` for small/medium bitsets and `ChunkedBitSet` for large bitsets. It's a speed and memory usage win.
r? `@Mark-Simulacrum`
Don't use a SyntheticProvider for literally every type
Replaces a glob regex with individualized imports for each standard library type. This improves debugger performance by quite a bit when populating lots of values with lots of fields
With the glob, afaik every single value of every single type that the debugger renders is run through a python function that does quite a few string comparisons (i plan to fix those next) to determine the SyntheticProvider to use. It looks like DefaultSyntheticProvider's functions internally call the liblldb c++ functions, which ends up with identical behavior to not using a SyntheticProvider at all, except you have extra python round trips slowing things down.
These sample vidoes were run on x86-64-pc-windows-gnu. `vect` is a 1000 element `Vec<Big>`, `Big` contains a dozen or so `Small`, and `Small` contains a dozen or so `[i32; 5]`
Before:
https://github.com/user-attachments/assets/07c31fe7-e126-4c2e-8ae9-cfe36e351d3f
After:
https://github.com/user-attachments/assets/6c0d1a45-1ffe-46de-95a0-5dbe59a173b5
---
try-job: aarch64-apple
Update cargo
6 commits in 05f54fdc34310f458033af8a63ce1d699fae8bf6..20a443231846b81c7b909691ec3f15eb173f2b18
2024-12-03 03:14:12 +0000 to 2024-12-06 21:56:56 +0000
- fix(fingerprint): Don't throwaway the cache on RUSTFLAGS changes (rust-lang/cargo#14830)
- fix(build-rs)!: Remove meaningless 'cargo_cfg_debug_assertions' (rust-lang/cargo#14901)
- docs(fingerprint): cargo-rustc extra flags do not affect the metadata (rust-lang/cargo#14898)
- fix(add): Don't select yanked versions when normalizing names (rust-lang/cargo#14895)
- fix(fix): Migrate workspace dependencies (rust-lang/cargo#14890)
- test(build-std): make mock-std closer to real world (rust-lang/cargo#14896)
Remove a lit_to_const call
We have so many special cases of `match expr.kind { Lit() => {}, Unary(Neg, Lit()) => {} }`... I'm trying to figure out how to get these all unified, but outright removing some is good, too. So let's try it.
Tho we don't have many `const {}` blocks in the perf test suite... But I also don't know how common `const { 42 }` blocks are, I'd expect these to occur mostly from macros (like `thread_local!`)
handle `--json-output` properly
Because `rustfmt` doesn't support JSON output, `x test --json-output` doesn't respect the `--json-output` flag during formatting step. This change makes that `x test` skips the formatting step if `--json-output` is specified. In addition, resolves#133855 with the 2nd commit.
Stabilize `std::io::ErrorKind::CrossesDevices`
FCP in #130191
cc #86442
See #130191 for more info and a recap of what has happened up until now.
TLDR: This had been FCP'd in December 2022 with some other `ErrorKind`s, but the stabilization got postponed due to some concerns voiced about several of the variants. However, the only concern ever voiced for this variant in particular was a wish to rename this to `NotSameDevice` analogous to Windows's `ERROR_NOT_SAME_DEVICE` (as opposed to Unix's `EXDEV`). This suggestion did not receive any support. So let's try to FCP this as is.
r? libs-api
