Refactor settings controller worker for FreeRTOS#144
Conversation
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
Codex Review: Here are some suggestions.
Reply with @codex fix comments to fix any unresolved comments.
About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you open a pull request for review, mark a draft as ready, or comment "@codex review". If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex fix this CI failure" or "@codex address that feedback".
|
|
||
| void SettingsController::Impl::enqueue_task(std::function<void()> task) | ||
| { | ||
| #if defined(ESP_PLATFORM) | ||
| if (!running_.load() || task_queue_ == nullptr) | ||
| { | ||
| return; | ||
| } | ||
|
|
||
| auto* heap_task = new (std::nothrow) std::function<void()>(std::move(task)); | ||
| if (heap_task == nullptr) | ||
| { | ||
| APP_LOG_WARN(kTag, "Failed to allocate task"); | ||
| return; | ||
| } | ||
|
|
||
| if (xQueueSend(task_queue_, &heap_task, 0) != pdPASS) | ||
| { | ||
| APP_LOG_WARN(kTag, "Task queue full"); | ||
| delete heap_task; | ||
| return; |
There was a problem hiding this comment.
[P1] Avoid dropping queued settings tasks when queue fills
The new FreeRTOS worker uses a queue with a fixed length of 8 and xQueueSend(..., 0) rejects enqueues when the queue is full. When a burst of UI actions enqueues more than 8 lambdas (for example moving the brightness slider or toggling multiple options quickly), later tasks are silently discarded after logging "Task queue full". Previously the std::queue path buffered every request, so the final user action was eventually executed. With this change, a user’s last action can be lost and the device ends up in a state that doesn’t match the UI. Consider either blocking until space is available or draining/overwriting older tasks so that the most recent command is processed.
Useful? React with 👍 / 👎.
1 participant
Summary
Testing
idf.py build(fails: command not found in container)https://chatgpt.com/codex/tasks/task_e_68ce683f67688324abbffe5d219332c1