lintcheck: don't run clippy in parallel by default

clippy_dev/src/lintcheck.rs

@@ -588,7 +588,10 @@ pub fn run(clap_config: &ArgMatches) {
             .flatten()
             .collect()
     } else {
-        let counter = std::sync::atomic::AtomicUsize::new(0);
+        if config.max_jobs > 1 {
+            // run parallel with rayon
+
+            let counter = AtomicUsize::new(0);
 
             // Ask rayon for thread count. Assume that half of that is the number of physical cores
             // Use one target dir for each core so that we can run N clippys in parallel.
@@ -610,6 +613,16 @@ pub fn run(clap_config: &ArgMatches) {
                 .map(|krate| krate.run_clippy_lints(&cargo_clippy_path, &counter, num_cpus, num_crates))
                 .flatten()
                 .collect()
+        } else {
+            // run sequential
+            let num_crates = crates.len();
+            crates
+                .into_iter()
+                .map(|krate| krate.download_and_extract())
+                .map(|krate| krate.run_clippy_lints(&cargo_clippy_path, &AtomicUsize::new(0), 1, num_crates))
+                .flatten()
+                .collect()
+        }
     };
 
     // generate some stats