Tencent/matrix · error · DexDataException

Endian constant has unexpected value

Error message

Endian constant has unexpected value ${Integer.toHexString(mHeaderItem.endianTag)}

What it means

After magic validation, DexData.parseHeaderItem reads the endian tag, which must equal ENDIAN_CONSTANT (0x12345678, little-endian) or REVERSE_ENDIAN_CONSTANT (0x78563412, big-endian). Any other value means the header is corrupt or the file isn't really a dex despite passing the magic check; it throws DexDataException.

Solutions

Rebuild or re-obtain the dex; a bad endian tag almost always means corruption.
Check endianness handling in whatever wrote/modified the dex — byte-swapped headers indicate a broken repacking tool.
Verify with `dexdump` or `d8` that the file opens in standard tooling; if not, the file is bad, not this library.
Compare file size and checksum/signature fields against the APK's copy to detect truncation.

Example fix

// before
DexData dexData = new DexData(raf);
dexData.load(); // throws on bad endian tag
// after
try {
    dexData.load();
} catch (DexDataException e) {
    MatrixLog.w(TAG, "corrupt dex header, skipping: " + inputPath);
}

Defensive patterns

Strategy: try-catch

Validate before calling

// verify dex integrity via header before load
byte[] header = readBytes(f, 0, 112);
int endianTag = ByteBuffer.wrap(header, 40, 4).order(ByteOrder.LITTLE_ENDIAN).getInt();
boolean ok = endianTag == 0x78563412L || endianTag == 0x12345678L;

Try / catch

try {
    dexData.load();
} catch (DexDataException e) {
    Log.w(TAG, "corrupt dex header (bad endian tag), skipping " + inputPath);
}

Prevention

Check the dex file size matches the header's fileSize field to detect truncation.
Avoid post-processing tools that rewrite dex headers without fixing endian/checksum fields.
Validate with standard tooling (dexdump) when a dex repeatedly fails.
Re-transfer/rebuild files that fail — bad endian tag means corruption, not a config issue.

When it happens

Trigger: parseHeaderItem (via load) reads an endianTag other than the two defined constants — bit-rotated/truncated dex, a file that coincidentally starts with dex magic but has damaged header, or wrong-offset reads caused by prior parsing bugs.

Common situations: Corrupted dexes from interrupted builds or bad file transfer; dexes modified by post-processing tools (repackaging, optimization) that broke the header; exotic big-endian-produced artifacts mis-handled by tooling.

Understand the failure class

Background: Checksum mismatch errors: "checksum verification failed", "digest mismatch", "expected vs actual checksum" — what they mean and how to fix them — this error's family across 41 libraries.

Related errors

AI-assisted analysis of Tencent/matrix@3b8293bd65 (2026-09-08). Data as JSON: /api/errors/7fc982ed28d078c3. Report an issue: GitHub.

Appendix: source

Thrown at matrix/matrix-android/matrix-commons/src/main/java/com/android/dexdeps/DexData.java:108

                    + "this is a DEX file?");
            throw new DexDataException();
        }

        /*
         * Read the endian tag, so we properly swap things as we read
         * them from here on.
         */
        seek(8 + 4 + 20 + 4 + 4);
        mHeaderItem.endianTag = readInt();
        if (mHeaderItem.endianTag == HeaderItem.ENDIAN_CONSTANT) {
            /* do nothing */
        } else if (mHeaderItem.endianTag == HeaderItem.REVERSE_ENDIAN_CONSTANT) {
            /* file is big-endian (!), reverse future reads */
            isBigEndian = true;
        } else {
            System.err.println("Endian constant has unexpected value "
                    + Integer.toHexString(mHeaderItem.endianTag));
            throw new DexDataException();
        }

        seek(8 + 4 + 20);  // magic, checksum, signature
        mHeaderItem.fileSize = readInt();
        mHeaderItem.headerSize = readInt();
        /*mHeaderItem.endianTag =*/ readInt();
        /*mHeaderItem.linkSize =*/ readInt();
        /*mHeaderItem.linkOff =*/ readInt();
        /*mHeaderItem.mapOff =*/ readInt();
        mHeaderItem.stringIdsSize = readInt();
        mHeaderItem.stringIdsOff = readInt();
        mHeaderItem.typeIdsSize = readInt();
        mHeaderItem.typeIdsOff = readInt();
        mHeaderItem.protoIdsSize = readInt();
        mHeaderItem.protoIdsOff = readInt();
        mHeaderItem.fieldIdsSize = readInt();
        mHeaderItem.fieldIdsOff = readInt();
        mHeaderItem.methodIdsSize = readInt();

View on GitHub (pinned to 3b8293bd65)