[codex] stage npm packages concurrently (#27853)

In the release job from

https://github.com/openai/codex/actions/runs/27391514823

staging the nine npm release tarballs serially took 104 seconds.

Each package build writes to a separate staging directory, output path,
and npm cache. Run them through the script's existing thread pool,
bounded by the available CPU count. Delete each staging tree as its
build finishes so concurrency does not retain all copies until the end.

On ubuntu-24.04 in

https://github.com/openai/codex/actions/runs/27397232050

two serial trials took 103 and 101 seconds, while concurrent trials
both took 41 seconds. Comparing every extracted file from the first
serial and concurrent sets found no differences. This removes about one
minute from every release.

scripts/stage_npm_packages.py

@@ -503,8 +503,7 @@ def main() -> int:
     artifacts_temp_root: Path | None = None
     remove_artifacts_temp_root = False
     resolved_head_sha: str | None = None
-
-    final_messages = []
+    staging_jobs: list[tuple[Path, list[str], str]] = []
 
     try:
         if native_component_sets:
@@ -570,22 +569,33 @@ def main() -> int:
             if vendor_src is not None:
                 cmd.extend(["--vendor-src", str(vendor_src)])
 
-            try:
-                run_command(cmd)
-            finally:
-                if not args.keep_staging_dirs:
-                    shutil.rmtree(staging_dir, ignore_errors=True)
+            staging_jobs.append(
+                (staging_dir, cmd, f"Staged {package} at {pack_output}")
+            )
 
-            final_messages.append(f"Staged {package} at {pack_output}")
+        max_workers = min(len(staging_jobs), os.cpu_count() or 1)
+        with ThreadPoolExecutor(max_workers=max_workers) as executor:
+            futures = {
+                executor.submit(run_command, cmd): staging_dir
+                for staging_dir, cmd, _message in staging_jobs
+            }
+            for future in as_completed(futures):
+                try:
+                    future.result()
+                finally:
+                    if not args.keep_staging_dirs:
+                        shutil.rmtree(futures[future], ignore_errors=True)
     finally:
         if not args.keep_staging_dirs:
+            for staging_dir, _cmd, _message in staging_jobs:
+                shutil.rmtree(staging_dir, ignore_errors=True)
             for vendor_temp_root in vendor_temp_roots:
                 shutil.rmtree(vendor_temp_root, ignore_errors=True)
         if remove_artifacts_temp_root and artifacts_temp_root is not None:
             shutil.rmtree(artifacts_temp_root, ignore_errors=True)
 
-    for msg in final_messages:
-        print(msg, flush=True)
+    for _staging_dir, _cmd, message in staging_jobs:
+        print(message, flush=True)
 
     return 0