Wide encoding support already ships: strict decode/encode for 36 encodings, BOM detection for UTF-8/16/32, and original-encoding + BOM preservation on save. Files without a BOM use the user-selected default and fail closed instead of guessing — by design, to never silently corrupt text.
Scope: auto-detect only with explainable confidence plus strict round-trip validation; low confidence keeps today's explicit failure / user-selected encoding.
Done when: confusable samples (UTF-8 vs Windows-1252 vs GBK/Big5/Shift-JIS) are covered; saving after detection preserves the original encoding; any guess that cannot round-trip losslessly is rejected.
Wide encoding support already ships: strict decode/encode for 36 encodings, BOM detection for UTF-8/16/32, and original-encoding + BOM preservation on save. Files without a BOM use the user-selected default and fail closed instead of guessing — by design, to never silently corrupt text.
Scope: auto-detect only with explainable confidence plus strict round-trip validation; low confidence keeps today's explicit failure / user-selected encoding.
Done when: confusable samples (UTF-8 vs Windows-1252 vs GBK/Big5/Shift-JIS) are covered; saving after detection preserves the original encoding; any guess that cannot round-trip losslessly is rejected.