No measurable performance improvement with native builds over LTO release builds.

Leaving the code for this commented out because it involved an unstable cargo feature without showing any benefit. I would like to revisit this later.

Cargo.toml

@@ -1,3 +1,5 @@
+# cargo-features = ["profile-rustflags"]
+
 [package]
 name = "organic"
 version = "0.1.10"
@@ -64,6 +66,13 @@ inherits = "release"
 lto = true
 strip = "symbols"
 
+# Optimized build for local execution.
+# [profile.native]
+# inherits = "release"
+# lto = true
+# strip = "symbols"
+# rustflags = ["-C", "target-cpu=native"]
+
 # Profile for performance testing with the "perf" tool. Notably keeps debug enabled and does not strip symbols to make reading the perf output easier.
 [profile.perf]
 inherits = "release"

src/parser/element_parser.rs

@@ -1,4 +1,3 @@
-
 use nom::multi::many0;
 
 use super::babel_call::babel_call;