CSVFixLab
Lab Case · Confirmed Repair

Windows-1252 CSV converted to UTF-8 without guessing Auto mode

The browser can prove the file is not valid UTF-8, but the user must confirm the legacy source encoding before conversion.

Reproducible binary fixture
Issue: encodingAuto outcome: blockedAfter source confirmation: repair allowed

Intended text in the fixture

name,city,note
José,Montréal,Café résumé
Müller,Zürich,Crème brûlée

The downloadable file stores those characters as Windows-1252 bytes rather than UTF-8 bytes.

Download Windows-1252 sampleOpen Encoding Fix

The safety behavior to reproduce

Leave Source encoding on Auto

Upload the fixture. Auto should identify that the byte sequence is not valid UTF-8 without claiming to know the legacy charset.

Confirm repair is disabled

The report should require source-encoding confirmation rather than offering a guessed conversion.

Select Windows-1252

The same file is decoded using the encoding you explicitly selected. The repair can now be enabled if CSV structure is otherwise safe.

Download and re-check

Upload the repaired copy again with Auto. The converted bytes should validate as UTF-8 while the accented text remains intact.

Why this matters

Many legacy encodings overlap for ordinary ASCII bytes. A short CSV can therefore be impossible to classify reliably from bytes alone. A tool that silently labels every non-UTF-8 file as Windows-1252 risks corrupting Korean, Japanese, or other text. CSVFixLab chooses a confirmation step instead.

Decision: conversion is safe only after the actual source encoding is confirmed.