chore(release): prepare 0.7.5

CHANGELOG.md

@@ -1,5 +1,9 @@
 # Changelog
 
+## 0.7.5 - 2026-10-06
+
+- Harden the `test-pi-local.mjs` temp-home cleanup against transient `ENOTEMPTY` when RPC children flush session files while exiting: poll with fresh removal attempts instead of relying on `maxRetries`, whose behavior on `ENOTEMPTY` varies across Node versions. Test-only change; the 0.7.4 release run failed on this cleanup after all tests had passed.
+
 ## 0.7.4 - 2026-10-05
 
 - Use the Provider API's OpenAI Responses wire (`/provider/v1/responses`) for models whose `/provider/v1/models` entry advertises `/responses`, keeping OpenAI Chat Completions for chat-only models and Anthropic Messages for Claude. The wire is resolved per model from `supported_endpoints`, falling back to Chat Completions when the catalog omits the field. This also enables prompt caching (`prompt_cache_key` is forwarded on `/responses` but ignored on `/chat/completions`): repeated prefixes bill at cache-read instead of full input price ([#132](https://github.com/patlux/pi-commandcode-provider/issues/132), [#133](https://github.com/patlux/pi-commandcode-provider/pull/133)).

package.json

@@ -1,6 +1,6 @@
 {
   "name": "pi-commandcode-provider",
-  "version": "0.7.4",
+  "version": "0.7.5",
   "description": "pi custom provider for Command Code API (commandcode.ai)",
   "type": "module",
   "keywords": [

tests/test-pi-local.mjs

@@ -1200,7 +1200,20 @@ try {
   console.log("[pi-local] PASS")
 } finally {
   await new Promise((resolve) => server.close(resolve))
-  // RPC children can still flush session files while exiting after SIGTERM.
-  // Retry transient ENOTEMPTY, but keep a persistent cleanup failure fatal.
-  rmSync(tempHome, { recursive: true, force: true, maxRetries: 5, retryDelay: 100 })
+  // RPC children can still flush session files while exiting after SIGTERM,
+  // so a single rmSync can hit transient ENOTEMPTY even with maxRetries
+  // (retry behavior on ENOTEMPTY varies across Node versions). Poll with
+  // fresh attempts instead, but keep a persistent cleanup failure fatal.
+  let lastCleanupError
+  for (let attempt = 0; attempt < 30; attempt += 1) {
+    try {
+      rmSync(tempHome, { recursive: true, force: true })
+      lastCleanupError = undefined
+      break
+    } catch (error) {
+      lastCleanupError = error
+      await new Promise((resolve) => setTimeout(resolve, 500))
+    }
+  }
+  if (lastCleanupError) throw lastCleanupError
 }