Does exceeding unwrapping affect a perfomance?

code review
1

I needed a method like is_executable, so I went to crates.io and discovered a magic crate for that. However Windows implementation was questionable for my needs and I simplified it. However my implementation has too many unwraps:

metadata.is_dir() || (self.extension().is_some() &&
              (self.extension().unwrap() == "exe" || self.extension().unwrap() == "bat"))

Should I worry about that, or Rust compiler is smart enough and will eliminate the extra unwrapping? How would you implemented the code?

2 
metadata.is_dir() || matches!(self.extension, Some("exe") | Some("bat"))
7 Likes
3

Slight improvement:

metadata.is_dir() || matches!(self.extension(), Some("exe" | "bat"))
9 Likes
4

Nice, thank you guys.

5

Because it's actually an OsStr, you can't pattern match it. But you can still avoid most of the duplication with is_some_and.

metadata.is_dir() || self.extension().is_some_and(|s| s == "exe" || s == "bat")

And to answer the title: I tried some stuff in godbolt, and unwrap by itself got optimized away just fine. However, when extension is involved, it leaves 3 uninlined function calls and the unwraps come back. Unwraps are much faster than extension anyway, so extension should be your main concern.

1 Like
6

Thank you for the comprehensive analysis. I assume that the fastest solution will look like :

If self.name.len() < 5 -> no,
If self.name[len-4] != '.' -> no,
If self.name[len-1] != 'e' goto check-t,
If self.name[len-3] != 'e' -> no,
If self.name[len-2] != 'x' -> no,
-> yes,
check-t: If self.name[len-1] != 't' -> no,
If self.name[len-3] != 'b' -> no,
If self.name[len-2] != 'a' -> no,
-> yes.

And check-t can be done in a parallel branch. However, a readability of the code will be worse, and it stops me from using the approach.

7

This topic was automatically closed 90 days after the last reply. We invite you to open a new topic if you have further questions or comments.