Check args more thoroghly in sense_hat.py and mangle traceback for sense_hat.py errors

[patch0]

Closes https://github.com/RaspberryPiFoundation/digital-maintenance-team/issues/7

This commit specifically fixes the issue triggered by using incorrect arguments in the set_pixel and set_pixels functions.

TypeError: object of type 'int' has no len() on line 1081 of sense_hat.py

I've added an explicit check to see if the pixel object has a __len__ attribute.

I've opened an identical PR on the Sense Hat library, and here and here!

Traceback mangling

I changed the error message when the filename is ./sense_hat.py to point to the next file in the stack, hopefully the user's code.

The issue with this is that it effectively masks errors in the sense_hat library, blaming the user instead. At the moment it only catches specific errors: ValueError, RuntimeError which are specifically mentioned in the sense hat shim.

Package update

This also updates the package version to 0.34.7.

Example python

# Import the libraries
from sense_hat import SenseHat
from time import sleep

# Set up the Sense HAT
sense = SenseHat()

sense.set_pixel(1, 2, 3)

# Or:
sense.set_pixels(1)

# Or:
# sense.set_pixels([1]*64)

Before

After

Tested on https://staging-editor-static.raspberrypi.org/branches/1348_merge/web-component.html

[Copilot]

Pull request overview

This pull request improves error handling for the sense_hat.py shim by making two key changes: (1) implementing safer pixel validation that explicitly checks for the __len__ attribute before calling len(), preventing TypeErrors when non-iterable objects are passed; and (2) manipulating the error traceback to hide internal shim frames and show users the line in their own code that caused the error, making debugging easier.

Changes:

• Enhanced pixel validation in set_pixel() to use hasattr() check before calling len(), fixing "TypeError: object of type 'int' has no len()"
• Added traceback manipulation logic to remove sense_hat.py frames from error messages
• Improved error message to suggest checking for missing x,y coordinates

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.

File	Description
src/components/Editor/Runners/PythonRunner/SkulptRunner/SkulptRunner.jsx	Adds logic to strip sense_hat.py frames from error tracebacks to show user code lines
public/shims/sense_hat/sense_hat_blob.py	Improves pixel validation with hasattr check and enhanced error message

---

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

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.

---

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

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 4 comments.

---

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

[Copilot]

This test sets up a local sense_hat.py project component, so the import resolves to user code (because builtinRead prioritizes local project files) rather than the external Sense HAT shim loaded via fetch(). That means the test isn't actually validating the shim traceback-mangling behavior, and it can also lock in incorrect behavior for projects that legitimately define their own sense_hat.py. Consider mocking the external library fetch (or asserting mangling only when there is no local sense_hat.py component) so the test matches the real shim scenario.

[patch0]

Mocking the fetch is really complicated, and I think this will do for now.

[Copilot]

The traceback mangling triggers for any exception whose top frame filename is "./sense_hat.py", even if the user has created their own local sense_hat.py file. Since builtinRead prefers local project files over external libraries, this can incorrectly attribute real user-code errors in sense_hat.py to main.py. Consider guarding this logic by checking whether a sense_hat.py component exists in the current project (and skipping mangling if it does), or otherwise positively identifying the external shim source before rewriting the traceback frame.

Suggested change

	// If this is an error in the sense_hat.py shim, use the next entry in
	// the traceback as this will be the line in the shim which we don't want
	// to show to users, so that the error message will instead point to the
	// line in the user's code which caused the error.
	if (
	err.traceback.length > 1 &&
	fileName === "./sense_hat.py" &&
	["RuntimeError", "ValueError"].includes(errorType)
	) {
	lineNumber = err.traceback[1].lineno;
	fileName = err.traceback[1].filename;
	}
	// Note: we intentionally avoid rewriting tracebacks originating from
	// sense_hat.py, because a user may define their own sense_hat.py file.
	// Blindly remapping those errors to another frame can misattribute
	// real user-code errors to a different file (e.g. main.py).

[patch0]

This is a good idea, but I think it is vanishingly small that someone would write a sense_hat.py file. It would make testing much more complex, as we'd have to stub the fetch(), which is tricky.