WipperSnapper for Raspberry Pi, Learn Code Element Updates, and more. What’s new with Adafruit IO and the Adafruit Learning System.

WipperSnapper for Raspberry Pi

From Brent:

WipperSnapper for Raspberry Pi (and Linux SBCs!) is shaping up to be the fastest way yet to connect a Raspberry Pi to Adafruit IO and start building.

We’re putting the finishing touches on WipperSnapper for Raspberry Pi and Linux SBCs, and we’ve reimagined how easy getting started should be. Most Raspberry Pi installers ask a lot of users, editing config files, running command after command, setting file permissions, and troubleshooting.

We’ve built WipperSnapper’s Raspberry Pi installer to take a different approach. The installer is a single line to paste into a terminal that installs WipperSnapper (and all of its dependencies, permissions, etc), provision the device’s credentials, and setup automatic run at boot. A webpage guides you through the whole process. There’s no manual setup and no guesswork.

The prototype site is vibe-coded with Claude, but, a polished, official version of this installer will ship on Adafruit IO at launch.

New Learn Code Element Update

Sheehan just released a big update to the way code is displayed in the Adafruit Learning System. Now long code elements will collapse down automatically so they don’t take up your entire screen. You can still quickly download or copy the code with a single click. There is better code highlighting, and more. She also spent time cleaning up how code elements look inside PDFs:

Hardware In the Loop (HIL) Testing

From Tyeth:

Having been told to not work on Hardware In the Loop (HIL) testing a few weeks ago until the other tasks were done, I obviously began immediately…while doing the other important tasks too of course

Mainly display related work, splash screen and optional status bar support, updating protobufs (the common binary language) and our MQTT testbed (protoMQ), and catching up on the v1 sensors in the v2 WipperSnapper codebase.

To that end we are creating two firmwares for V2, Python and Arduino, so both of those needed sensor + displays support adding. All in all that’s a lot of testing, and in my life-long bid to automate myself out of tasks I decided to pursue the agent assisted HIL test game.

As our new python Wippersnapper V2 client runs on Raspberry Pi (and other linux based) single board computers it seemed the ideal testbed for some mini hardware test platforms, and so the first display test host was born, rpi-displays! Using a simple pi camera v3 (4000×3000 pixels approx) with neopixel ring pointed at 13 microcontrollers (existing V1 boards) all with displays attached, linked via a controllable USB hub to a Raspberry Pi.

Then a bit of ssh testing first to prove the idea and finally a “simple” API to allow GitHub jobs to request test runs too. Finally my AI agents can see all the hardware and interact with it, proving the v1 existing work before testing V2.

Then when fixing a pixel write bug I realised I could use the test farm to prove that the regression was present in v126 and fixed in v130. Pure joy as I saw my first pull request on github receive an automated test result with proof attached (the deployment log covering erasing and flashing the MCU, it’s serial log covering bootup and debug output, and the protoMQ / MQTT log showing broker interactions and protobuf messages). Huzzah!

Since then I’ve been doing WipperSnapper Python and Arduino HIL testing, and had visual proofs in github comments too, including a PR adding a whole new parallel display type, and there’s a collection of reliability improvements gone in to support the more awkward boards that can appear to get stuck (plus focus and exposure support).

The webpage is a bit gnarly, but allows me (or ideally agents) to setup devices and hosts and peripherals (like displays) easily, configure cameras and Regions Of Interest (ROI), and do USB lookup stuff to help debug or setup. One day it might get a usability facelift, but for now it’s functional enough (less so for humans), and I’m more interested in pursuing the testing of physical phenomena to show sensor comparisons (like a box with a controllable hair dryer for temp/humidity sensors)