fix: use 0-based byte offset in ps1 encoding test failure message

[Copilot]

The failure message in test_ps1_file_is_ascii_only reported "byte offset" but stored i + 1 (1-based), creating an off-by-one inconsistency when debugging encoding failures.

Change

• tests/test_ps1_encoding.py: Replace i + 1 with i so the recorded position is 0-based, matching the conventional meaning of "byte offset".

# Before
non_ascii = [(i + 1, byte) for i, byte in enumerate(content) if byte > 127]
# "first at byte offset 1" for the very first byte — misleading

# After
non_ascii = [(i, byte) for i, byte in enumerate(content) if byte > 127]
# "first at byte offset 0" for the very first byte — correct