From e06155bcad239f4bb8b25f075e3c54d4a168b942 Mon Sep 17 00:00:00 2001
From: Dennis Bartels <hello@dennisbartels.de>
Date: Thu, 29 Oct 2026 17:33:14 +0100
Subject: [PATCH] fix(Mp4): locate moov from the mdat header instead of
 buffering the whole mdat

A non-optimized MP4 was only recognized once the complete mdat atom had
been downloaded, so a 128 MB hi-res ALAC file buffered for many seconds
before playback could start. The mdat header already holds its size,
which is enough to request the trailing moov atom.
---
 .../Audio Source/Mp4/Mp4Restructure.swift     |  8 ++-
 .../Streaming/Mp4RestructureTests.swift       | 62 +++++++++++++++++++
 2 files changed, 69 insertions(+), 1 deletion(-)
 create mode 100644 AudioStreamingTests/Streaming/Mp4RestructureTests.swift

diff --git a/AudioStreaming/Streaming/Audio Source/Mp4/Mp4Restructure.swift b/AudioStreaming/Streaming/Audio Source/Mp4/Mp4Restructure.swift
index 0b7df80..eaab9f2 100644
--- a/AudioStreaming/Streaming/Audio Source/Mp4/Mp4Restructure.swift	
+++ b/AudioStreaming/Streaming/Audio Source/Mp4/Mp4Restructure.swift	
@@ -211,10 +211,16 @@ final class Mp4Restructure {
                     foundMoov = true
                 case Atoms.mdat:
                     foundMdat = true
+                    // The mdat header carries its size, so the moov offset is known without buffering the whole mdat.
+                    if ftyp != nil, !foundMoov {
+                        atoms.append(MP4Atom(type: atomType, size: atomSize, offset: atomOffset))
+                        Logger.debug("🕵️ detected a non-optimized mp4 (header-only observation)", category: .generic)
+                        return .needsRestructure(moovOffset: atomOffset + atomSize)
+                    }
                 default:
                     break
                 }
-                
+
                 if ftyp != nil {
                     if foundMoov && !foundMdat {
                         Logger.debug("🕵️ detected an optimized mp4 (header-only observation)", category: .generic)
diff --git a/AudioStreamingTests/Streaming/Mp4RestructureTests.swift b/AudioStreamingTests/Streaming/Mp4RestructureTests.swift
new file mode 100644
index 0000000..d3a5e16
--- /dev/null
+++ b/AudioStreamingTests/Streaming/Mp4RestructureTests.swift
@@ -0,0 +1,62 @@
+import XCTest
+@testable import AudioStreaming
+
+final class Mp4RestructureTests: XCTestCase {
+    private func atom(_ type: String, payload: Data) -> Data {
+        var data = Data()
+        data.append(contentsOf: withUnsafeBytes(of: UInt32(8 + payload.count).bigEndian, Array.init))
+        data.append(type.data(using: .ascii)!)
+        data.append(payload)
+        return data
+    }
+
+    private func uint32(_ value: UInt32) -> Data {
+        Data(withUnsafeBytes(of: value.bigEndian, Array.init))
+    }
+
+    private var ftyp: Data { atom("ftyp", payload: Data("M4A \0\0\u{2}\0M4A isom".utf8)) }
+
+    private func moov(chunkOffsets: [UInt32]) -> Data {
+        var stco = Data([0, 0, 0, 0]) + uint32(UInt32(chunkOffsets.count))
+        chunkOffsets.forEach { stco.append(uint32($0)) }
+        return atom("moov", payload: atom("trak", payload: atom("stco", payload: stco)))
+    }
+
+    func testNonOptimizedIsDetectedFromMdatHeaderAlone() throws {
+        let mdatSize = 128_486_896
+        var head = ftyp + atom("free", payload: Data())
+        let mdatOffset = head.count
+        head.append(uint32(UInt32(mdatSize)))
+        head.append("mdat".data(using: .ascii)!)
+        head.append(Data(repeating: 0, count: 4096))
+
+        let result = try Mp4Restructure().checkIsOptimized(data: head)
+
+        XCTAssertEqual(result, .needsRestructure(moovOffset: mdatOffset + mdatSize))
+    }
+
+    func testOptimizedIsDetectedFromMoovHeaderAlone() throws {
+        var head = ftyp
+        head.append(uint32(900_000))
+        head.append("moov".data(using: .ascii)!)
+
+        XCTAssertEqual(try Mp4Restructure().checkIsOptimized(data: head), .optimized)
+    }
+
+    func testRestructureAfterHeaderOnlyDetectionMovesMoovBeforeMdat() throws {
+        let ftyp = self.ftyp
+        let mdat = atom("mdat", payload: Data(repeating: 0xAB, count: 64))
+        let chunkOffset = UInt32(ftyp.count + 8)
+        let moov = moov(chunkOffsets: [chunkOffset])
+        let restructure = Mp4Restructure()
+
+        let check = try restructure.checkIsOptimized(data: ftyp + mdat.prefix(16))
+        XCTAssertEqual(check, .needsRestructure(moovOffset: ftyp.count + mdat.count))
+
+        let (initialData, mdatOffset) = try restructure.restructureMoov(data: moov)
+
+        XCTAssertEqual(mdatOffset, ftyp.count)
+        XCTAssertEqual(initialData, ftyp + self.moov(chunkOffsets: [chunkOffset + UInt32(moov.count)]))
+        XCTAssertEqual(restructure.seekAdjusted(offset: 1000), 1000 - moov.count)
+    }
+}