tests: Benchmarks for --files

.github/workflows/main.yml

@@ -55,6 +55,23 @@ jobs:
           tool: cargo-msrv
       - run: cargo msrv verify
 
+  bench-files:
+    name: Run benchmarks for `--files` option
+
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          submodules: recursive
+      - uses: swatinem/rust-cache@v2
+      - uses: taiki-e/install-action@v2
+        with:
+          tool: hyperfine
+      - name: Run benchmarks
+        run: >
+          ./benches/bench-files.sh
+
   build-test-feature-powerset:
     name: Build and test feature powerset
     runs-on: ubuntu-latest

.gitignore

@@ -1,5 +1,8 @@
 # https://github.com/github/gitignore/blob/main/Rust.gitignore
 
+# From a custom benchmark script
+benches/srgn
+
 # While working on new README samples, ignore corresponding files
 sample.*
 

.gitmodules

@@ -0,0 +1,9 @@
+[submodule "benches/django"]
+	path = benches/django
+	url = [email protected]:django/django.git
+[submodule "benches/kubernetes"]
+	path = benches/kubernetes
+	url = [email protected]:kubernetes/kubernetes.git
+[submodule "benches/pydantic"]
+	path = benches/pydantic
+	url = [email protected]:pydantic/pydantic.git

benches/bench-files.sh

@@ -0,0 +1,52 @@
+#!/usr/bin/env bash
+
+set -o errexit
+set -o nounset
+set -o pipefail
+
+setup() {
+    sudo --validate
+
+    command -v hyperfine >/dev/null || cargo install hyperfine
+
+    git submodule init
+    git submodule update
+
+    cargo build --release
+    cp target/release/srgn benches/
+}
+
+bench() {
+    # https://www.kernel.org/doc/Documentation/sysctl/vm.txt:
+    local wipe_caches='sync; echo 3 | sudo tee /proc/sys/vm/drop_caches'
+
+    (
+        cd benches
+
+        for i in \
+            "python comments py django,pydantic" \
+            "go comments go kubernetes"
+        do
+            # We WANT splitting here: https://stackoverflow.com/a/52228219/11477374
+            # shellcheck disable=SC2086
+            set -- $i
+
+            local lang="$1"
+            local query_type="$2"
+            local file_suffix="$3"
+            local repos="$4"  # Can be a comma-separated list
+
+            hyperfine \
+                --max-runs 3 \
+                --prepare "$wipe_caches" \
+                --cleanup "git restore --recurse-submodules {repo}" \
+                --parameter-list repo "$repos" \
+                --parameter-list find "e+,[Tt]he" \
+                --parameter-list replace "_,🙂" \
+                "./srgn --fail-empty-glob --$lang $query_type --files '{repo}/**/*.$file_suffix' '{find}' '{replace}'"
+        done
+    )
+}
+
+setup
+bench