motiejus/zig
1
Fork 0
Code Releases Activity

std.mem: use indexOfScalarPos when indexOf is called where needle.len == 1

Browse Source 
When `std.mem.indexOf` is called with a single-item needle, use `indexOfScalarPos` which is significantly faster than the more general `indexOfPosLinear`.  This can be done without introducing overhead to normal cases (where `needle.len > 1`).
This commit is contained in:
Karl Seguin
2023-10-09 21:50:16 +08:00
committed by GitHub
parent 57874ce619
commit 75b48ef503
1 changed files with 5 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
lib/std/mem.zig
View File

@@ -1344,7 +1344,11 @@ pub fn lastIndexOf(comptime T: type, haystack: []const T, needle: []const T) ?us
/// Uses Boyer-Moore-Horspool algorithm on large inputs; `indexOfPosLinear` on small inputs.
pub fn indexOfPos(comptime T: type, haystack: []const T, start_index: usize, needle: []const T) ?usize {
if (needle.len > haystack.len) return null;
if (needle.len == 0) return start_index;
if (needle.len < 2) {
if (needle.len == 0) return start_index;
// indexOfScalarPos is significantly faster than indexOfPosLinear
return indexOfScalarPos(T, haystack, start_index, needle[0]);
}
if (!meta.trait.hasUniqueRepresentation(T) or haystack.len < 52 or needle.len <= 4)
return indexOfPosLinear(T, haystack, start_index, needle);