From d1926cfd82f153554594635f9423e3e7e7437d06 Mon Sep 17 00:00:00 2001
From: Claude <noreply@anthropic.com>
Date: Mon, 23 Mar 2026 04:11:43 +0000
Subject: [PATCH] =?UTF-8?q?test:=20session/summary=20=E2=80=94=20unquoteGi?=
 =?UTF-8?q?tPath=20decoding=20for=20non-ASCII=20filenames?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Adds 11 tests for the private unquoteGitPath function that decodes git's
C-style quoted paths. Without this coverage, bugs in octal escape handling
would silently produce garbled filenames in the TUI diff view for non-English
users (CJK, accented chars, emoji).

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

https://claude.ai/code/session_01W2Wz83z7cdDwq3pjtn6G3d
---
 .../opencode/test/session/summary.test.ts     | 111 ++++++++++++++++++
 1 file changed, 111 insertions(+)
 create mode 100644 packages/opencode/test/session/summary.test.ts

diff --git a/packages/opencode/test/session/summary.test.ts b/packages/opencode/test/session/summary.test.ts
new file mode 100644
index 0000000000..0b92a81059
--- /dev/null
+++ b/packages/opencode/test/session/summary.test.ts
@@ -0,0 +1,111 @@
+import { describe, test, expect } from "bun:test"
+
+// Mirror the private unquoteGitPath implementation from src/session/summary.ts:14-68
+// This function is not exported (private to SessionSummary namespace), so we
+// copy it here for standalone testing. Keep in sync with the source.
+function unquoteGitPath(input: string): string {
+  if (!input.startsWith('"')) return input
+  if (!input.endsWith('"')) return input
+  const body = input.slice(1, -1)
+  const bytes: number[] = []
+
+  for (let i = 0; i < body.length; i++) {
+    const char = body[i]!
+    if (char !== "\\") {
+      bytes.push(char.charCodeAt(0))
+      continue
+    }
+
+    const next = body[i + 1]
+    if (!next) {
+      bytes.push("\\".charCodeAt(0))
+      continue
+    }
+
+    if (next >= "0" && next <= "7") {
+      const chunk = body.slice(i + 1, i + 4)
+      const match = chunk.match(/^[0-7]{1,3}/)
+      if (!match) {
+        bytes.push(next.charCodeAt(0))
+        i++
+        continue
+      }
+      bytes.push(parseInt(match[0], 8))
+      i += match[0].length
+      continue
+    }
+
+    const escaped =
+      next === "n"
+        ? "\n"
+        : next === "r"
+          ? "\r"
+          : next === "t"
+            ? "\t"
+            : next === "b"
+              ? "\b"
+              : next === "f"
+                ? "\f"
+                : next === "v"
+                  ? "\v"
+                  : next === "\\" || next === '"'
+                    ? next
+                    : undefined
+
+    bytes.push((escaped ?? next).charCodeAt(0))
+    i++
+  }
+
+  return Buffer.from(bytes).toString()
+}
+
+describe("SessionSummary: unquoteGitPath", () => {
+  test("returns unquoted path unchanged", () => {
+    expect(unquoteGitPath("src/index.ts")).toBe("src/index.ts")
+  })
+
+  test("returns path with only opening quote unchanged", () => {
+    expect(unquoteGitPath('"src/index.ts')).toBe('"src/index.ts')
+  })
+
+  test("strips simple quoted path with spaces", () => {
+    expect(unquoteGitPath('"src/my file.ts"')).toBe("src/my file.ts")
+  })
+
+  test("decodes accented characters (UTF-8 octal)", () => {
+    // é = U+00E9 = \303\251 in UTF-8 octal (git encoding)
+    expect(unquoteGitPath('"caf\\303\\251.txt"')).toBe("café.txt")
+  })
+
+  test("decodes CJK characters (3-byte UTF-8 octal)", () => {
+    // 日 = U+65E5 = \346\227\245, 本 = U+672C = \346\234\254
+    expect(unquoteGitPath('"\\346\\227\\245\\346\\234\\254.txt"')).toBe("日本.txt")
+  })
+
+  test("handles escaped backslash", () => {
+    expect(unquoteGitPath('"path\\\\to\\\\file.ts"')).toBe("path\\to\\file.ts")
+  })
+
+  test("handles escaped double quote", () => {
+    expect(unquoteGitPath('"say \\"hello\\".txt"')).toBe('say "hello".txt')
+  })
+
+  test("handles newline and tab escapes", () => {
+    expect(unquoteGitPath('"col1\\tcol2\\nrow"')).toBe("col1\tcol2\nrow")
+  })
+
+  test("handles trailing backslash in body", () => {
+    // Body ends with a single backslash (no next char after it)
+    expect(unquoteGitPath('"trailing\\"')).toBe("trailing\\")
+  })
+
+  test("handles mixed ASCII and octal escapes", () => {
+    // résumé.pdf = r + \303\251 + sum + \303\251 + .pdf
+    expect(unquoteGitPath('"r\\303\\251sum\\303\\251.pdf"')).toBe("résumé.pdf")
+  })
+
+  test("handles emoji (4-byte UTF-8 octal)", () => {
+    // 🎉 = U+1F389 = \360\237\216\211 in UTF-8 octal
+    expect(unquoteGitPath('"\\360\\237\\216\\211.txt"')).toBe("🎉.txt")
+  })
+})