Python TextIOWrapper: choose whether input line endings are normalized

Operation contract

Two independent wrappers decode the same owned UTF-8 bytes. The universal-newline reader converts CRLF and LF to LF. The second reader keeps the original endings, so a byte-sensitive parser can distinguish them. Both wrappers close their owned buffers.

Failure boundary

Neither mode changes the original bytes in the source buffer. For signed or hashed wire data, authenticate bytes before text translation. A real file reader should also declare encoding, error policy, and a byte limit; the in-memory fixture omits those outer constraints.

Working program

import io

wire_bytes = b"R-47\r\nR-73\n"
with io.TextIOWrapper(io.BytesIO(wire_bytes), encoding="utf-8", newline=None) as normalized:
    print("normalized", repr(normalized.read()))
with io.TextIOWrapper(io.BytesIO(wire_bytes), encoding="utf-8", newline="") as preserved:
    print("preserved", repr(preserved.read()))

Output

normalized 'R-47\nR-73\n'
preserved 'R-47\r\nR-73\n'

Costs and limits

Decoding scans the input and allocates a text representation. The two independent wrappers in this fixture each hold a short source; do not duplicate large buffers merely to compare modes in production.

Common Mistakes

• Universal-newline mode changes the text returned by read.
• Text normalization must not precede verification of exact signed bytes.
• Closing a TextIOWrapper also closes its underlying buffer unless ownership is transferred deliberately.

Connected lessons

• Python strings and bytes: reject decoding errors before parsing records
• Python streaming SHA-256: hash bytes without a second full copy
• Python CSV imports: parse quoted fields before validating rows

Test this contract.