3 files changed, 191 insertions, 0 deletions
diff --git a/third_party/rust/unicode-segmentation/benches/graphemes.rs b/third_party/rust/unicode-segmentation/benches/graphemes.rs
new file mode 100644
index 0000000000..3a0b9b76a3
--- /dev/null
+++ b/third_party/rust/unicode-segmentation/benches/graphemes.rs
@@ -0,0 +1,63 @@
+use criterion::{black_box, criterion_group, criterion_main, Criterion};
+use unicode_segmentation;
+
+use std::fs;
+use unicode_segmentation::UnicodeSegmentation;
+
+fn graphemes(c: &mut Criterion, lang: &str, path: &str) {
+    let text = fs::read_to_string(path).unwrap();
+
+    c.bench_function(&format!("graphemes_{}", lang), |bench| {
+        bench.iter(|| {
+            for g in UnicodeSegmentation::graphemes(black_box(&*text), true) {
+                black_box(g);
+            }
+        })
+    });
+}
+
+fn graphemes_arabic(c: &mut Criterion) {
+    graphemes(c, "arabic", "benches/texts/arabic.txt");
+}
+
+fn graphemes_english(c: &mut Criterion) {
+    graphemes(c, "english", "benches/texts/english.txt");
+}
+
+fn graphemes_hindi(c: &mut Criterion) {
+    graphemes(c, "hindi", "benches/texts/hindi.txt");
+}
+
+fn graphemes_japanese(c: &mut Criterion) {
+    graphemes(c, "japanese", "benches/texts/japanese.txt");
+}
+
+fn graphemes_korean(c: &mut Criterion) {
+    graphemes(c, "korean", "benches/texts/korean.txt");
+}
+
+fn graphemes_mandarin(c: &mut Criterion) {
+    graphemes(c, "mandarin", "benches/texts/mandarin.txt");
+}
+
+fn graphemes_russian(c: &mut Criterion) {
+    graphemes(c, "russian", "benches/texts/russian.txt");
+}
+
+fn graphemes_source_code(c: &mut Criterion) {
+    graphemes(c, "source_code", "benches/texts/source_code.txt");
+}
+
+criterion_group!(
+    benches,
+    graphemes_arabic,
+    graphemes_english,
+    graphemes_hindi,
+    graphemes_japanese,
+    graphemes_korean,
+    graphemes_mandarin,
+    graphemes_russian,
+    graphemes_source_code,
+);
+
+criterion_main!(benches);
diff --git a/third_party/rust/unicode-segmentation/benches/unicode_words.rs b/third_party/rust/unicode-segmentation/benches/unicode_words.rs
new file mode 100644
index 0000000000..c87851a376
--- /dev/null
+++ b/third_party/rust/unicode-segmentation/benches/unicode_words.rs
@@ -0,0 +1,64 @@
+#[macro_use]
+extern crate bencher;
+extern crate unicode_segmentation;
+
+use bencher::Bencher;
+use std::fs;
+use unicode_segmentation::UnicodeSegmentation;
+
+fn unicode_words(bench: &mut Bencher, path: &str) {
+    let text = fs::read_to_string(path).unwrap();
+    bench.iter(|| {
+        for w in text.unicode_words() {
+            bencher::black_box(w);
+        }
+    });
+
+    bench.bytes = text.len() as u64;
+}
+
+fn unicode_words_arabic(bench: &mut Bencher) {
+    unicode_words(bench, "benches/texts/arabic.txt");
+}
+
+fn unicode_words_english(bench: &mut Bencher) {
+    unicode_words(bench, "benches/texts/english.txt");
+}
+
+fn unicode_words_hindi(bench: &mut Bencher) {
+    unicode_words(bench, "benches/texts/hindi.txt");
+}
+
+fn unicode_words_japanese(bench: &mut Bencher) {
+    unicode_words(bench, "benches/texts/japanese.txt");
+}
+
+fn unicode_words_korean(bench: &mut Bencher) {
+    unicode_words(bench, "benches/texts/korean.txt");
+}
+
+fn unicode_words_mandarin(bench: &mut Bencher) {
+    unicode_words(bench, "benches/texts/mandarin.txt");
+}
+
+fn unicode_words_russian(bench: &mut Bencher) {
+    unicode_words(bench, "benches/texts/russian.txt");
+}
+
+fn unicode_words_source_code(bench: &mut Bencher) {
+    unicode_words(bench, "benches/texts/source_code.txt");
+}
+
+benchmark_group!(
+    benches,
+    unicode_words_arabic,
+    unicode_words_english,
+    unicode_words_hindi,
+    unicode_words_japanese,
+    unicode_words_korean,
+    unicode_words_mandarin,
+    unicode_words_russian,
+    unicode_words_source_code,
+);
+
+benchmark_main!(benches);
diff --git a/third_party/rust/unicode-segmentation/benches/word_bounds.rs b/third_party/rust/unicode-segmentation/benches/word_bounds.rs
new file mode 100644
index 0000000000..6b01ddb109
--- /dev/null
+++ b/third_party/rust/unicode-segmentation/benches/word_bounds.rs
@@ -0,0 +1,64 @@
+#[macro_use]
+extern crate bencher;
+extern crate unicode_segmentation;
+
+use bencher::Bencher;
+use std::fs;
+use unicode_segmentation::UnicodeSegmentation;
+
+fn word_bounds(bench: &mut Bencher, path: &str) {
+    let text = fs::read_to_string(path).unwrap();
+    bench.iter(|| {
+        for w in text.split_word_bounds() {
+            bencher::black_box(w);
+        }
+    });
+
+    bench.bytes = text.len() as u64;
+}
+
+fn word_bounds_arabic(bench: &mut Bencher) {
+    word_bounds(bench, "benches/texts/arabic.txt");
+}
+
+fn word_bounds_english(bench: &mut Bencher) {
+    word_bounds(bench, "benches/texts/english.txt");
+}
+
+fn word_bounds_hindi(bench: &mut Bencher) {
+    word_bounds(bench, "benches/texts/hindi.txt");
+}
+
+fn word_bounds_japanese(bench: &mut Bencher) {
+    word_bounds(bench, "benches/texts/japanese.txt");
+}
+
+fn word_bounds_korean(bench: &mut Bencher) {
+    word_bounds(bench, "benches/texts/korean.txt");
+}
+
+fn word_bounds_mandarin(bench: &mut Bencher) {
+    word_bounds(bench, "benches/texts/mandarin.txt");
+}
+
+fn word_bounds_russian(bench: &mut Bencher) {
+    word_bounds(bench, "benches/texts/russian.txt");
+}
+
+fn word_bounds_source_code(bench: &mut Bencher) {
+    word_bounds(bench, "benches/texts/source_code.txt");
+}
+
+benchmark_group!(
+    benches,
+    word_bounds_arabic,
+    word_bounds_english,
+    word_bounds_hindi,
+    word_bounds_japanese,
+    word_bounds_korean,
+    word_bounds_mandarin,
+    word_bounds_russian,
+    word_bounds_source_code,
+);
+
+benchmark_main!(benches);