speedup directory traversal on windows

[klensy]

Optimizes walking over dirs on windows by replacing FindFirstFileW with FindFirstFileExW with FindExInfoBasic option, that allows skipping filling unused struct field which should be faster.

Also adds the same change for fallback call of FindFirstFileExW in metadata call.

Locally shows small speedup, but bench results from other users are welcome.

[rustbot]

r? @ChrisDenton

rustbot has assigned @ChrisDenton.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[klensy]

For app:

use walkdir::WalkDir;

fn main() {
    let mut c = 0;
    for _entry in WalkDir::new(".") {
        c += 1;
    }
    
    println!("{}",c );
}

running over llvm-project folder on HDD.
dirwalk.old: rustc 1.84.0-nightly (da93539 2024-10-19) (i'll try master checkout later, maybe something other affecting perf the same as master)
dirwalk.new: this PR

$ hyperfine --warmup 3 'dirwalk.old.exe'  'dirwalk.new.exe'
Benchmark 1: dirwalk.old.exe
  Time (mean ± σ):     549.3 ms ±   6.6 ms    [User: 147.5 ms, System: 395.9 ms]
  Range (min … max):   532.9 ms … 559.6 ms    10 runs

Benchmark 2: dirwalk.new.exe
  Time (mean ± σ):     453.7 ms ±   7.2 ms    [User: 186.6 ms, System: 263.1 ms]
  Range (min … max):   441.7 ms … 460.3 ms    10 runs

Summary
  dirwalk.new.exe ran
    1.21 ± 0.02 times faster than dirwalk.old.exe

Well, idk why results difference so big, this looks wrong. Messed up source file, all ok.

[ChrisDenton]

Seems fine to me. It would be interesting to see the effect of FIND_FIRST_EX_LARGE_FETCH given modern systems. But in any case I do think this is an improvement.

[ChrisDenton]

@bors r+

[bors]

📌 Commit 2920ed0 has been approved by ChrisDenton

It is now in the queue for this repository.

[bors]

⌛ Testing commit 2920ed0 with merge 3ec4308...

[bors]

☀️ Test successful - checks-actions
Approved by: ChrisDenton
Pushing 3ec4308 to master...

[rust-timer]

Finished benchmarking commit (3ec4308): comparison URL.

Overall result: no relevant changes - no action needed

@rustbot label: -perf-regression

Instruction count

This benchmark run did not return any relevant results for this metric.

Max RSS (memory usage)

Results (secondary -2.9%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.9%	[-3.7%, -2.1%]	2
All ❌✅ (primary)	-	-	0

Cycles

Results (secondary 4.1%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	7.8%	[6.6%, 8.8%]	3
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-7.0%	[-7.0%, -7.0%]	1
All ❌✅ (primary)	-	-	0

Binary size

This benchmark run did not return any relevant results for this metric.

Bootstrap: 782.557s -> 781.937s (-0.08%)
Artifact size: 333.66 MiB -> 333.66 MiB (-0.00%)