From dfbdc97cd38e962d2c8ea2233737a45071e2f359 Mon Sep 17 00:00:00 2001
From: "kiloconnect[bot]" <240665456+kiloconnect[bot]@users.noreply.github.com>
Date: Wed, 22 Apr 2026 13:04:30 +0000
Subject: [PATCH] refactor(cli): simplify isBinaryFile check to UTF-16 BOM only

Per review: CJK and legacy single-byte encodings already passed the old
control-char heuristic (their bytes are all >= 0x80, never NUL), so the
full Encoding.detect pass in isBinaryFile was solving a non-problem for
them. The only realistic regression in the old heuristic is UTF-16 with
BOM, where the second byte of every ASCII character is 0x00 and fires
the NUL-byte early-return.

Replace the detect call with a 2-byte BOM check via a new
Encoding.hasUtf16Bom helper. Also reword the move/add inline markers in
patch/index.ts to note that Encoding.write handles mkdir.
---
 packages/opencode/src/kilocode/encoding.ts | 6 ++++++
 packages/opencode/src/patch/index.ts       | 2 +-
 packages/opencode/src/tool/read.ts         | 6 ++----
 3 files changed, 9 insertions(+), 5 deletions(-)

diff --git a/packages/opencode/src/kilocode/encoding.ts b/packages/opencode/src/kilocode/encoding.ts
index 9345dd56a9..f846da9871 100644
--- a/packages/opencode/src/kilocode/encoding.ts
+++ b/packages/opencode/src/kilocode/encoding.ts
@@ -38,6 +38,12 @@ export namespace Encoding {
     return bytes.length >= 3 && bytes[0] === 0xef && bytes[1] === 0xbb && bytes[2] === 0xbf
   }
 
+  /** True if `bytes[0..limit]` starts with a UTF-16 LE or BE byte-order mark. */
+  export function hasUtf16Bom(bytes: Buffer, limit = bytes.length): boolean {
+    if (limit < 2) return false
+    return (bytes[0] === 0xff && bytes[1] === 0xfe) || (bytes[0] === 0xfe && bytes[1] === 0xff)
+  }
+
   /** Remap jschardet labels to iconv-lite compatible names. */
   function normalize(name: string): string {
     const lower = name.toLowerCase().replace(/[^a-z0-9]/g, "")
diff --git a/packages/opencode/src/patch/index.ts b/packages/opencode/src/patch/index.ts
index 323ce403fc..20c46de136 100644
--- a/packages/opencode/src/patch/index.ts
+++ b/packages/opencode/src/patch/index.ts
@@ -549,7 +549,7 @@ export namespace Patch {
 
           if (hunk.move_path) {
             // Handle file move
-            await Encoding.write(hunk.move_path, fileUpdate.content, fileUpdate.encoding) // kilocode_change
+            await Encoding.write(hunk.move_path, fileUpdate.content, fileUpdate.encoding) // kilocode_change - encoding-aware write (mkdirs)
             await fs.unlink(hunk.path)
             modified.push(hunk.move_path)
             log.info(`Moved file: ${hunk.path} -> ${hunk.move_path}`)
diff --git a/packages/opencode/src/tool/read.ts b/packages/opencode/src/tool/read.ts
index 5d596fd4ec..2dfb1f6587 100644
--- a/packages/opencode/src/tool/read.ts
+++ b/packages/opencode/src/tool/read.ts
@@ -330,10 +330,8 @@ export async function isBinaryFile(filepath: string, fileSize: number): Promise<
     const result = await fh.read(bytes, 0, sampleSize, 0)
     if (result.bytesRead === 0) return false
 
-    // kilocode_change start - treat detected non-UTF-8 text (CJK, UTF-16 with BOM) as text, not binary
-    const sample = bytes.subarray(0, result.bytesRead)
-    const enc = Encoding.detect(sample)
-    if (enc !== "utf-8") return false
+    // kilocode_change start - UTF-16 BOM: NUL bytes are legitimate, skip the NUL/control-char heuristic
+    if (Encoding.hasUtf16Bom(bytes, result.bytesRead)) return false
     // kilocode_change end
 
     let nonPrintableCount = 0