Preserve circular buffer item semantics

[mjc]

Fix two independent circular-buffer behaviors.

circular_buffer_iterate() previously used an item index as a byte offset. For buffers containing multi-byte records, callbacks after the first could therefore receive pointers inside another record. The index is now multiplied by item_size so each callback receives a complete item. This fixes the experimental recorder plotting path, which iterates over Sample records.

Following review feedback, circular_buffer_pop() now implements a conventional FIFO pop: its unused index argument was removed, it copies the item at the tail, and then advances the tail. There are currently no callers requiring migration.

The defensive argument guards and indexed-item shifting from earlier revisions were removed following review feedback.

[Copilot]

Pull request overview

This PR updates the circular buffer implementation so that operations consistently treat the underlying storage as item-sized records (not bytes), improving correctness for non-item_size == 1 buffers while keeping FIFO behavior unchanged.

Changes:

• Added input validation in circular_buffer_push / circular_buffer_get to ignore invalid buffer/item pointers and zero-sized configurations.
• Fixed indexed removal (circular_buffer_pop) to shift whole items to preserve item semantics.
• Fixed iteration to pass item-aligned pointers to callbacks.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[lukash]

Minor, the commit title "Circular buffer: Remove indexed items" doesn't make much sense, could you improve it? Besides that looks good now.

[mjc]

should be all set now. sorry about the amount of churn, doing my best

[lukash]

No worries, thanks for the patience, I think it's good to maintain a readable history. FWIW I prefer a bit more descriptive commit titles overall (not too much though), but this is good enough 😁