[m-messer]

Summary

A quantity whose value stands for several values had the unit factor applied to the list as a whole. sympy_forms built ({value})*({factor}), so {1, 2} km became ((1, 2))*(1000), and parse_quantity raised QuantityParseError ("A bracketed list such as (a, b) can't be used in arithmetic"). This contradicted the documented standard_values ("several for a value such as {1, 2}").

Affected inputs, each with a unit:

• {1, 2} km
• 1,2~\mathrm{km} (a LaTeX comma list)
• plus_minus 2 km
• \pm 2~\mathrm{km}

When create_expression_set finds more than one value, the factor is now applied to each value. A single value is converted exactly as before.

Found while porting compareExpressions' main #295 onto the refactor (test_latex_response_parsed_as_list on compareExpressions #280).

Testing

• New test_unit_factor_applies_to_each_value_of_a_list in physical_quantities/tests/test_regressions.py with four cases. All four fail without the fix.
• scripts/test.sh: slr_parsing 116, criteria 105, expression_parsing 519 and physical_quantities 542, all passing.
• ruff check, ruff format --check and mypy are clean.
• tools/parity/check.py shows 0 unexpected changes and 0 mismatched. A first attempt multiplied the parsed values for single values too, which changed forms such as 3*micro*1 to 3*micro. Single values therefore keep the text route.

Follow-up (compareExpressions)

The app's LaTeX quantity preview (app/preview/physical_quantity.py) reads the value with parse_expression, so it still raises for a list. It needs to show either both values or the last one (main #295 kept the last one, giving 2 metre). That change goes with the pin bump that picks this up.

🤖 Generated with Claude Code