offload: ASC MHL v2.0 conformance — urn:ASC:MHL:v2.0 namespace, creatorinfo+processinfo, path@size attr, xxh64@action@hashdate; validates against official ascmitc XSD via xmllint

This commit is contained in:
Forge Dev 2026-07-11 04:36:41 +00:00
parent 05693dfece
commit 70e0290eda
2 changed files with 92 additions and 36 deletions

View file

@ -1,8 +1,10 @@
import Foundation import Foundation
/// Media Hash List (ASC MHL-style) manifest: per-file path/size/xxh64. /// ASC Media Hash List v2.0 (urn:ASC:MHL:v2.0) manifest.
/// Structure follows MHL v2 hashlist shape; full ASC MHL chain/directory /// Structure per official XSD (github.com/ascmitc/mhl, xsd/ASCMHL.xsd):
/// hashes can layer on later without breaking this format. /// hashlist > creatorinfo (creationdate, hostname, tool) + processinfo (process)
/// + hashes > hash > path[@size] + <xxh64 action hashdate>.
/// Findings documented in docs/research/red-rcp2-protocol.md (MHL section).
public enum MHL { public enum MHL {
public struct Failure: Equatable, Sendable { public struct Failure: Equatable, Sendable {
@ -33,17 +35,35 @@ public enum MHL {
.replacingOccurrences(of: "&amp;", with: "&") .replacingOccurrences(of: "&amp;", with: "&")
} }
/// Build manifest XML for files (relative paths) under root. private static func isoNow() -> String {
public static func generate(root: String, creator: String, files: [String]) throws -> String { ISO8601DateFormatter().string(from: Date())
}
private static func hostname() -> String {
ProcessInfo.processInfo.hostName
}
/// Build ASC MHL v2.0 manifest XML for files (relative paths) under root.
/// process: "transfer" for offloads, "in-place" for verification-only seals.
public static func generate(
root: String,
creator: String,
files: [String],
process: String = "transfer"
) throws -> String {
let fm = FileManager.default let fm = FileManager.default
let dateFormatter = ISO8601DateFormatter() let now = isoNow()
var out = """ var out = """
<?xml version="1.0" encoding="UTF-8"?> <?xml version="1.0" encoding="UTF-8"?>
<hashlist version="2.0"> <hashlist version="2.0" xmlns="urn:ASC:MHL:v2.0">
<creatorinfo> <creatorinfo>
<creationdate>\(dateFormatter.string(from: Date()))</creationdate> <creationdate>\(now)</creationdate>
<tool>\(escape(creator))</tool> <hostname>\(escape(hostname()))</hostname>
<tool version="0.1">\(escape(creator))</tool>
</creatorinfo> </creatorinfo>
<processinfo>
<process>\(process)</process>
</processinfo>
<hashes> <hashes>
""" """
@ -56,9 +76,8 @@ public enum MHL {
let hash = try FileHasher.xxh64(path: full) let hash = try FileHasher.xxh64(path: full)
out += """ out += """
<hash> <hash>
<path>\(escape(rel))</path> <path size="\(size)">\(escape(rel))</path>
<size>\(size)</size> <xxh64 action="original" hashdate="\(now)">\(hash)</xxh64>
<xxh64>\(hash)</xxh64>
</hash> </hash>
""" """
@ -68,8 +87,14 @@ public enum MHL {
} }
/// Write manifest to file. /// Write manifest to file.
public static func write(root: String, creator: String, files: [String], to path: String) throws { public static func write(
let xml = try generate(root: root, creator: creator, files: files) root: String,
creator: String,
files: [String],
to path: String,
process: String = "transfer"
) throws {
let xml = try generate(root: root, creator: creator, files: files, process: process)
try xml.write(toFile: path, atomically: true, encoding: .utf8) try xml.write(toFile: path, atomically: true, encoding: .utf8)
} }
@ -84,11 +109,13 @@ public enum MHL {
failures.append(Failure(path: e.path, reason: .missing)) failures.append(Failure(path: e.path, reason: .missing))
continue continue
} }
if let expectedSize = e.size {
let size = (try fm.attributesOfItem(atPath: full)[.size] as? UInt64) ?? 0 let size = (try fm.attributesOfItem(atPath: full)[.size] as? UInt64) ?? 0
if size != e.size { if size != expectedSize {
failures.append(Failure(path: e.path, reason: .sizeMismatch)) failures.append(Failure(path: e.path, reason: .sizeMismatch))
continue continue
} }
}
let hash = try FileHasher.xxh64(path: full) let hash = try FileHasher.xxh64(path: full)
if hash != e.xxh64 { if hash != e.xxh64 {
failures.append(Failure(path: e.path, reason: .hashMismatch)) failures.append(Failure(path: e.path, reason: .hashMismatch))
@ -99,11 +126,12 @@ public enum MHL {
struct Entry { struct Entry {
let path: String let path: String
let size: UInt64 let size: UInt64?
let xxh64: String let xxh64: String
} }
/// Scanner-based parse of <hash> blocks. /// Scanner-based parse of <hash> blocks. Accepts v2.0 attribute-style size
/// and (legacy, our v1) element-style <size>.
static func parse(_ xml: String) throws -> [Entry] { static func parse(_ xml: String) throws -> [Entry] {
guard xml.contains("<hashlist") else { guard xml.contains("<hashlist") else {
throw OffloadError.readFailed("not an MHL manifest") throw OffloadError.readFailed("not an MHL manifest")
@ -113,18 +141,37 @@ public enum MHL {
while let blockStart = xml.range(of: "<hash>", range: searchRange), while let blockStart = xml.range(of: "<hash>", range: searchRange),
let blockEnd = xml.range(of: "</hash>", range: blockStart.upperBound..<xml.endIndex) { let blockEnd = xml.range(of: "</hash>", range: blockStart.upperBound..<xml.endIndex) {
let block = String(xml[blockStart.upperBound..<blockEnd.lowerBound]) let block = String(xml[blockStart.upperBound..<blockEnd.lowerBound])
func field(_ tag: String) -> String? {
guard let open = block.range(of: "<\(tag)>"), // Path element with optional attributes: <path size="N" ...>rel</path>
let close = block.range(of: "</\(tag)>"), guard let pathOpen = block.range(of: "<path"),
open.upperBound <= close.lowerBound else { return nil } let pathTagEnd = block.range(of: ">", range: pathOpen.upperBound..<block.endIndex),
return String(block[open.upperBound..<close.lowerBound]) let pathClose = block.range(of: "</path>", range: pathTagEnd.upperBound..<block.endIndex) else {
throw OffloadError.readFailed("malformed <hash> entry: no path")
} }
guard let path = field("path"), let pathAttrs = String(block[pathOpen.upperBound..<pathTagEnd.lowerBound])
let sizeStr = field("size"), let size = UInt64(sizeStr), let path = unescape(String(block[pathTagEnd.upperBound..<pathClose.lowerBound]))
let hash = field("xxh64") else {
throw OffloadError.readFailed("malformed <hash> entry") // size attribute (v2.0) or legacy <size> element.
var size: UInt64?
if let m = pathAttrs.range(of: "size=\"") {
let rest = pathAttrs[m.upperBound...]
if let q = rest.firstIndex(of: "\"") {
size = UInt64(rest[..<q])
} }
entries.append(Entry(path: unescape(path), size: size, xxh64: hash)) } else if let sizeOpen = block.range(of: "<size>"),
let sizeClose = block.range(of: "</size>") {
size = UInt64(block[sizeOpen.upperBound..<sizeClose.lowerBound].trimmingCharacters(in: .whitespacesAndNewlines))
}
// xxh64 element with optional attributes.
guard let hOpen = block.range(of: "<xxh64"),
let hTagEnd = block.range(of: ">", range: hOpen.upperBound..<block.endIndex),
let hClose = block.range(of: "</xxh64>", range: hTagEnd.upperBound..<block.endIndex) else {
throw OffloadError.readFailed("malformed <hash> entry: no xxh64")
}
let hash = String(block[hTagEnd.upperBound..<hClose.lowerBound]).trimmingCharacters(in: .whitespacesAndNewlines)
entries.append(Entry(path: path, size: size, xxh64: hash))
searchRange = blockEnd.upperBound..<xml.endIndex searchRange = blockEnd.upperBound..<xml.endIndex
} }
guard !entries.isEmpty else { guard !entries.isEmpty else {

View file

@ -18,14 +18,23 @@ final class MHLTests: XCTestCase {
} }
// MHL XML structure: hashlist root, creatorinfo, hash entries with file/size/xxh64. // MHL XML structure: hashlist root, creatorinfo, hash entries with file/size/xxh64.
// ASC MHL v2.0 XSD shape (urn:ASC:MHL:v2.0):
// hashlist > creatorinfo (creationdate, hostname, tool) + processinfo (process)
// + hashes > hash > path[@size] + xxh64[@action][@hashdate].
func testWriteStructure() throws { func testWriteStructure() throws {
let xml = try MHL.generate(root: root, creator: "Forge 0.1", files: ["CLIPS/C001.mov", "CLIPS/C002.mov"]) let xml = try MHL.generate(root: root, creator: "Forge 0.1", files: ["CLIPS/C001.mov", "CLIPS/C002.mov"])
XCTAssertTrue(xml.hasPrefix("<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n<hashlist version=\"2.0\">")) XCTAssertTrue(xml.hasPrefix("<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n<hashlist version=\"2.0\" xmlns=\"urn:ASC:MHL:v2.0\">"))
XCTAssertTrue(xml.contains("<creatorinfo>")) XCTAssertTrue(xml.contains("<creatorinfo>"))
XCTAssertTrue(xml.contains("<tool>Forge 0.1</tool>")) XCTAssertTrue(xml.contains("<creationdate>"))
XCTAssertTrue(xml.contains("<path>CLIPS/C001.mov</path>")) XCTAssertTrue(xml.contains("<hostname>"))
XCTAssertTrue(xml.contains("<size>17</size>")) XCTAssertTrue(xml.contains("<tool version=\"0.1\">Forge 0.1</tool>"))
XCTAssertTrue(xml.contains("<xxh64>")) // processinfo required by XSD; offload = transfer.
XCTAssertTrue(xml.contains("<processinfo>"))
XCTAssertTrue(xml.contains("<process>transfer</process>"))
// size is a path ATTRIBUTE per XSD, not an element.
XCTAssertTrue(xml.contains("<path size=\"17\">CLIPS/C001.mov</path>"))
XCTAssertTrue(xml.contains("<xxh64 action=\"original\" hashdate="))
XCTAssertFalse(xml.contains("<size>"))
XCTAssertTrue(xml.hasSuffix("</hashlist>\n")) XCTAssertTrue(xml.hasSuffix("</hashlist>\n"))
} }
@ -76,7 +85,7 @@ final class MHLTests: XCTestCase {
func testPathEscaping() throws { func testPathEscaping() throws {
try Data("x".utf8).write(to: URL(fileURLWithPath: root + "/a&b.mov")) try Data("x".utf8).write(to: URL(fileURLWithPath: root + "/a&b.mov"))
let xml = try MHL.generate(root: root, creator: "Forge", files: ["a&b.mov"]) let xml = try MHL.generate(root: root, creator: "Forge", files: ["a&b.mov"])
XCTAssertTrue(xml.contains("<path>a&amp;b.mov</path>")) XCTAssertTrue(xml.contains(">a&amp;b.mov</path>"))
let result = try MHL.verify(manifest: xml, root: root) let result = try MHL.verify(manifest: xml, root: root)
XCTAssertTrue(result.passed) XCTAssertTrue(result.passed)
} }