1.3.2
FIX
What changed
Parser is stream-based with a 4MB read buffer.
Failed rows isolate with a usable line number and reason instead of aborting.
Encoding detection covers UTF-8, UTF-16, and Windows-1251.
Supported encodings
UTF-8 — default, detected via BOM and heuristic.
UTF-16 LE/BE — detected via BOM, surrogate-pair aware.
Windows-1251 — heuristic detection for Cyrillic legacy exports.
Fixed
Memory crash on CSV imports over 50MB.
Generic error message on failed rows.
Fixed a memory crash when importing CSVs over 50MB.
What changed
Parser is now stream-based with a 4MB read buffer.
Failed rows isolate with line number and reason instead of aborting.
Encoding detection covers UTF-8, UTF-16, and Windows-1251.

