From ec4f4ddc588a564b75d69249b4ebbe4d4051c792 Mon Sep 17 00:00:00 2001
From: yasshmittal <yash.mittal@op.iitg.ac.in>
Date: Tue, 13 Oct 2026 18:11:28 +0530
Subject: [PATCH] fix(ghidra): read UTF-8 signed language definitions (#2772)

Buffer#toString kept a leading UTF-8 byte order mark as U+FEFF, so the
strict XML parser rejected a valid .ldefs file as content outside the root
element and aborted the whole language catalog read. XML 1.0 section 4.3.3
defines the mark as an encoding signature; decode with TextDecoder, which
consumes it. Malformed XML is still rejected.

Fixes #2533

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
---
 src/ghidra/GhidraLanguageCatalog.ts       |  4 +++-
 src/ghidra/GhidraLanguageOverride.test.ts | 23 +++++++++++++++++++++++
 2 files changed, 26 insertions(+), 1 deletion(-)

diff --git a/src/ghidra/GhidraLanguageCatalog.ts b/src/ghidra/GhidraLanguageCatalog.ts
index 5da5943dc..6373f187f 100644
--- a/src/ghidra/GhidraLanguageCatalog.ts
+++ b/src/ghidra/GhidraLanguageCatalog.ts
@@ -75,8 +75,10 @@ export const readGhidraLanguageCatalog = async (
       if (!entry.endsWith(".ldefs")) continue;
       const bytes = await readFile(join(directory, entry));
       if (bytes.length > MAX_LDEFS_BYTES) continue;
+      // XML 1.0 section 4.3.3: a leading UTF-8 BOM is an encoding signature,
+      // not document content; TextDecoder consumes it, Buffer#toString does not.
       definitions.push(
-        ...parseGhidraLanguageDefinitions(bytes.toString("utf8")),
+        ...parseGhidraLanguageDefinitions(new TextDecoder().decode(bytes)),
       );
     }
   }
diff --git a/src/ghidra/GhidraLanguageOverride.test.ts b/src/ghidra/GhidraLanguageOverride.test.ts
index 5f90a1454..02c5bf245 100644
--- a/src/ghidra/GhidraLanguageOverride.test.ts
+++ b/src/ghidra/GhidraLanguageOverride.test.ts
@@ -2,6 +2,7 @@ import { mkdir, mkdtemp, rm, writeFile } from "node:fs/promises";
 import { tmpdir } from "node:os";
 import { join } from "node:path";
 import { afterAll, beforeAll, describe, expect, it } from "vitest";
+import { createTestTempDirectory } from "../../tests/fixtures/temporaryDirectory.js";
 import { parseConfig } from "../config/parseConfig.js";
 import type { BinaryTarget } from "../domain/binaryTargetTypes.js";
 import { projectAnalysisError } from "../domain/analysisErrorProjection.js";
@@ -190,6 +191,28 @@ describe("Ghidra language catalog", () => {
     );
   });
 
+  it("reads a UTF-8 signed definition beside an unsigned sibling", async () => {
+    const root = await createTestTempDirectory("rea-ghidra-bom-");
+    const module = (name: string) =>
+      join(root, "Ghidra/Processors", name, "data/languages");
+    await mkdir(module("probe"), { recursive: true });
+    await mkdir(module("x86"), { recursive: true });
+    await writeFile(
+      join(module("probe"), "probe.ldefs"),
+      Buffer.concat([
+        Buffer.from([0xef, 0xbb, 0xbf]),
+        Buffer.from(
+          '<?xml version="1.0" encoding="UTF-8"?>\n<language_definitions><language id="probe:LE:32:default"/></language_definitions>',
+        ),
+      ]),
+    );
+    await writeFile(join(module("x86"), "x86.ldefs"), X86_LDEFS);
+    expect(await readGhidraLanguageCatalog(root)).toEqual([
+      { id: "probe:LE:32:default", compilerSpecIds: [] },
+      ...parseGhidraLanguageDefinitions(X86_LDEFS),
+    ]);
+  });
+
   it("admits only languages and compiler specs the installation defines", () => {
     const catalog = parseGhidraLanguageDefinitions(X86_LDEFS);
     expect(