fix(bench): Result interpretation problems (#5798)

Co-authored-by: Lucas Nogueira <lucas@tauri.studio>

.github/workflows/bench.yml

@@ -10,6 +10,7 @@ on:
 env:
   RUST_BACKTRACE: 1
   CARGO_PROFILE_DEV_DEBUG: 0 # This would add unnecessary bloat to the target folder, decreasing cache efficiency.
+  LC_ALL: en_US.UTF-8 # This prevents strace from changing it's number format to use commas.
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.ref }}

tooling/bench/src/build_benchmark_jsons.rs

@@ -22,20 +22,19 @@ fn main() {
 
   // all data's
   let all_data_buffer =
-    BufReader::new(File::open(&tauri_data).expect("Unable to read all data file"));
+    BufReader::new(File::open(tauri_data).expect("Unable to read all data file"));
   let mut all_data: Vec<utils::BenchResult> =
     serde_json::from_reader(all_data_buffer).expect("Unable to read all data buffer");
 
   // add current data to alls data
   all_data.push(current_data);
 
   // use only latest 20 elements from alls data
-  let recent: Vec<utils::BenchResult>;
-  if all_data.len() > 20 {
-    recent = all_data[all_data.len() - 20..].to_vec();
+  let recent: Vec<utils::BenchResult> = if all_data.len() > 20 {
+    all_data[all_data.len() - 20..].to_vec()
   } else {
-    recent = all_data.clone();
-  }
+    all_data.clone()
+  };
 
   // write json's
   utils::write_json(
@@ -44,13 +43,13 @@ fn main() {
       .expect("Something wrong with tauri_data"),
     &serde_json::to_value(&all_data).expect("Unable to build final json (alls)"),
   )
-  .expect(format!("Unable to write {:?}", tauri_data).as_str());
+  .unwrap_or_else(|_| panic!("Unable to write {:?}", tauri_data));
 
   utils::write_json(
     tauri_recent
       .to_str()
       .expect("Something wrong with tauri_recent"),
     &serde_json::to_value(&recent).expect("Unable to build final json (recent)"),
   )
-  .expect(format!("Unable to write {:?}", tauri_recent).as_str());
+  .unwrap_or_else(|_| panic!("Unable to write {:?}", tauri_recent));
 }

tooling/bench/src/run_benchmark.rs

@@ -49,7 +49,7 @@ fn run_strace_benchmarks(new_data: &mut utils::BenchResult) -> Result<()> {
     let mut file = tempfile::NamedTempFile::new()?;
 
     Command::new("strace")
-      .args(&[
+      .args([
         "-c",
         "-f",
         "-o",
@@ -64,7 +64,10 @@ fn run_strace_benchmarks(new_data: &mut utils::BenchResult) -> Result<()> {
     file.as_file_mut().read_to_string(&mut output)?;
 
     let strace_result = utils::parse_strace_output(&output);
-    let clone = strace_result.get("clone").map(|d| d.calls).unwrap_or(0) + 1;
+    // Note, we always have 1 thread. Use cloneX calls as counter for additional threads created.
+    let clone = 1
+      + strace_result.get("clone").map(|d| d.calls).unwrap_or(0)
+      + strace_result.get("clone3").map(|d| d.calls).unwrap_or(0);
     let total = strace_result.get("total").unwrap().calls;
     thread_count.insert(name.to_string(), clone);
     syscall_count.insert(name.to_string(), total);
@@ -84,7 +87,7 @@ fn run_max_mem_benchmark() -> Result<HashMap<String, u64>> {
     let benchmark_file = benchmark_file.to_str().unwrap();
 
     let proc = Command::new("mprof")
-      .args(&[
+      .args([
         "run",
         "-C",
         "-o",
@@ -99,7 +102,7 @@ fn run_max_mem_benchmark() -> Result<HashMap<String, u64>> {
     println!("{:?}", proc_result);
     results.insert(
       name.to_string(),
-      utils::parse_max_mem(&benchmark_file).unwrap(),
+      utils::parse_max_mem(benchmark_file).unwrap(),
     );
   }
 
@@ -132,7 +135,7 @@ fn rlib_size(target_dir: &std::path::Path, prefix: &str) -> u64 {
 fn get_binary_sizes(target_dir: &Path) -> Result<HashMap<String, u64>> {
   let mut sizes = HashMap::<String, u64>::new();
 
-  let wry_size = rlib_size(&target_dir, "libwry");
+  let wry_size = rlib_size(target_dir, "libwry");
   println!("wry {} bytes", wry_size);
   sizes.insert("wry_rlib".to_string(), wry_size);
 
@@ -174,9 +177,9 @@ fn cargo_deps() -> HashMap<String, usize> {
       let mut cmd = Command::new("cargo");
       cmd.arg("tree");
       cmd.arg("--no-dedupe");
-      cmd.args(&["--edges", "normal"]);
-      cmd.args(&["--prefix", "none"]);
-      cmd.args(&["--target", target]);
+      cmd.args(["--edges", "normal"]);
+      cmd.args(["--prefix", "none"]);
+      cmd.args(["--target", target]);
       cmd.current_dir(&utils::tauri_root_path());
 
       let full_deps = cmd.output().expect("failed to run cargo tree").stdout;
@@ -268,7 +271,7 @@ fn main() -> Result<()> {
     time::format_description::parse("[year]-[month]-[day]T[hour]:[minute]:[second]Z").unwrap();
   let now = time::OffsetDateTime::now_utc();
   let mut new_data = utils::BenchResult {
-    created_at: format!("{}", now.format(&format).unwrap()),
+    created_at: now.format(&format).unwrap(),
     sha1: utils::run_collect(&["git", "rev-parse", "HEAD"])
       .0
       .trim()

tooling/bench/src/utils.rs

@@ -45,12 +45,11 @@ pub fn get_target() -> &'static str {
 }
 
 pub fn target_dir() -> PathBuf {
-  let target_dir = bench_root_path()
+  bench_root_path()
     .join("tests")
     .join("target")
     .join(get_target())
-    .join("release");
-  target_dir.into()
+    .join("release")
 }
 
 pub fn bench_root_path() -> PathBuf {
@@ -105,16 +104,14 @@ pub fn parse_max_mem(file_path: &str) -> Option<u64> {
   let output = BufReader::new(file);
   let mut highest: u64 = 0;
   // MEM 203.437500 1621617192.4123
-  for line in output.lines() {
-    if let Ok(line) = line {
-      // split line by space
-      let split = line.split(" ").collect::<Vec<_>>();
-      if split.len() == 3 {
-        // mprof generate result in MB
-        let current_bytes = str::parse::<f64>(split[1]).unwrap() as u64 * 1024 * 1024;
-        if current_bytes > highest {
-          highest = current_bytes;
-        }
+  for line in output.lines().flatten() {
+    // split line by space
+    let split = line.split(' ').collect::<Vec<_>>();
+    if split.len() == 3 {
+      // mprof generate result in MB
+      let current_bytes = str::parse::<f64>(split[1]).unwrap() as u64 * 1024 * 1024;
+      if current_bytes > highest {
+        highest = current_bytes;
       }
     }
   }
@@ -169,14 +166,26 @@ pub fn parse_strace_output(output: &str) -> HashMap<String, StraceOutput> {
   }
 
   let total_fields = total_line.split_whitespace().collect::<Vec<_>>();
+
   summary.insert(
     "total".to_string(),
-    StraceOutput {
-      percent_time: str::parse::<f64>(total_fields[0]).unwrap(),
-      seconds: str::parse::<f64>(total_fields[1]).unwrap(),
-      usecs_per_call: None,
-      calls: str::parse::<u64>(total_fields[2]).unwrap(),
-      errors: str::parse::<u64>(total_fields[3]).unwrap(),
+    match total_fields.len() {
+      // Old format, has no usecs/call
+      5 => StraceOutput {
+        percent_time: str::parse::<f64>(total_fields[0]).unwrap(),
+        seconds: str::parse::<f64>(total_fields[1]).unwrap(),
+        usecs_per_call: None,
+        calls: str::parse::<u64>(total_fields[2]).unwrap(),
+        errors: str::parse::<u64>(total_fields[3]).unwrap(),
+      },
+      6 => StraceOutput {
+        percent_time: str::parse::<f64>(total_fields[0]).unwrap(),
+        seconds: str::parse::<f64>(total_fields[1]).unwrap(),
+        usecs_per_call: Some(str::parse::<u64>(total_fields[2]).unwrap()),
+        calls: str::parse::<u64>(total_fields[3]).unwrap(),
+        errors: str::parse::<u64>(total_fields[4]).unwrap(),
+      },
+      _ => panic!("Unexpected total field count: {}", total_fields.len()),
     },
   );
 
@@ -222,7 +231,7 @@ pub fn download_file(url: &str, filename: PathBuf) {
     .arg("-s")
     .arg("-o")
     .arg(&filename)
-    .arg(&url)
+    .arg(url)
     .status()
     .unwrap();