fix(csv): emit columns for extensions with multiple FUNCTION_INFO fields

[ArmaanjeetSandhu]

Problem

csv_output() in lizard_ext/csvoutput.py only appended extension columns when an extension declared exactly one FUNCTION_INFO field:

if hasattr(extension, 'FUNCTION_INFO') and \
        len(list(extension.FUNCTION_INFO)) == 1:

Any extension with more than one field was skipped entirely, so lizard --csv -Eio or lizard --csv -Eduplicated_param_list produced the same CSV as a run with no extension at all. The metrics appeared correctly in the default table but vanished in CSV output.

Fix

Replace the len == 1 guard with a plain loop over every FUNCTION_INFO key, accumulating extension_variables and extension_captions in declared order. The downstream header and row-building code is unchanged.

Single-field extensions (-Eexitcount, -Egotocount, etc.) and no-extension runs produce byte-for-byte identical output to before.

Tests

Two regression tests added to test/testOutputCSV.py:

• test_csv_header_with_multi_field_extension: verifies the header contains all declared captions
• test_csv_row_with_multi_field_extension: verifies both field values appear in the data row

The tests use neutral field names (metric_one, metric_two) rather than real extension attribute names to avoid collisions with the FunctionInfo properties that extension modules install at import time.