Support Kanji encoding mode in QR codes

[productdevbook]

Context

Frequently requested across multiple QR code libraries:

• qr-code-styling#283 — Unicode encoding problems (5 comments)
• qr-code-styling#141 — Encoding options for Byte mode (5 comments)
• qr-code-styling#96 — Language support (4 comments)
• qr-code-styling#208 — UTF-8/Chinese/Persian data cannot be scanned (7 comments)
• qr-code-styling#303 — Byte mode with UTF-8 and SJIS encodings (2 comments)

Description

The QR code spec (ISO/IEC 18004) defines a Kanji encoding mode that efficiently encodes Shift JIS characters using 13 bits per character, significantly more compact than byte mode for Japanese text. This is one of the four core encoding modes alongside numeric, alphanumeric, and byte.

Current Status in etiket

The types and mode indicators exist (kanji is in EncodingMode, MODE_INDICATOR.kanji = 0b1000), but the encoder throws an error:

case "kanji":
  throw new Error("Kanji encoding mode is not yet supported. Use byte mode for non-ASCII text.");

Byte mode works for UTF-8 text but produces larger QR codes for Japanese content.

Proposed Solution

1. Implement Shift JIS encoding for Kanji mode in src/encoders/qr/mode.ts
2. Add Kanji character detection to selectMode() in src/encoders/qr/version.ts
3. Update the segment optimizer to consider Kanji mode switching
4. Add ECI (Extended Channel Interpretation) mode support for specifying character sets beyond the default (ISO 8859-1)

[productdevbook]

Resolved in commit 3a40945 — Kanji encoding with 13-bit Shift JIS. encodeKanjiData(), unicodeToShiftJIS(). Works via encodeQR(text, { mode: 'kanji' }). 7 tests.