Surface SyntaxError from input transformation instead of crashing the…

[JimWallace]

… kernel

PyodideKernel.run() called transform_cell() outside any try/except, so a SyntaxError raised during input transformation — most commonly TabError on Python 3.12+, which the tokenizer raises eagerly — escaped run() unhandled and surfaced as a raw kernel traceback instead of a normal error cell.

Guard the transform like InteractiveShell.run_cell does and route failures through showtraceback() (which dispatches SyntaxError to showsyntaxerror), reusing the existing _last_traceback reporting.

https://claude.ai/code/session_017qs8Cz1XgFG4oBPqNPqLom

[github-actions]

👈 Try it on ReadTheDocs

[JimWallace]

Closes #292

Summary

On Python 3.12+, a SyntaxError that CPython's tokenizer raises during tokenization — most commonly TabError: inconsistent use of tabs and spaces in indentation — escaped PyodideKernel.run() unhandled, so the user saw a raw internal traceback through pyodide_kernel/litetransform.py → IPython/core/inputtransformer2.py:make_tokens_by_line instead of a normal error cell.

Root cause

PyodideKernel.run() called the input transforms outside any try/except. On Python 3.12+ the C tokenizer raises TabError/IndentationError (a SyntaxError subclass) eagerly during tokenization (python/cpython#105238). IPython's generate_tokens_catch_errors only does except tokenize.TokenError, so it cannot catch a SyntaxError; both make_tokens_by_line and InteractiveShell.transform_cell propagate it. IPython's own InteractiveShell.run_cell guards transform_cell in a try/except; run() reimplemented the cell loop and omitted that guard. Not fixable by upgrading IPython (confirmed against 9.14.0).

Fix

Wrap the two transform_cell calls in a try/except that routes failures through showtraceback() (which dispatches SyntaxError → showsyntaxerror), nesting the existing _load_packages_from_imports block under else:. Reuses the method's existing _last_traceback → results["status"] = "error" reporting.