[Bug] get wrong stream position while get "FileHeader" stream in ".hwp" file

[sboh1214]

Hello, I'm using OLEKit with hwp file which use in Korea.
It has stream "FileHeader" with fixed size 256 bytes.
However, OLEKit throws uncaught error Thread 1: EXC_BAD_INSTRUCTION (code=EXC_I386_INVOP, subcode=0x0)
when I try to get data of stream by readDataToEnd()

let positiveURL = URL(fileURLWithPath: #file)
        .deletingLastPathComponent()
        .appendingPathComponent("SampleHwp")
        .appendingPathComponent("blank.hwp")
let ole = try OLEFile(positiveURL.path)
let fileHeaderStream = ole.root.children.first(where: { $0.name == "FileHeader"})!
let reader = try ole.stream(fileHeaderStream)
let data = reader.readDataToEnd()

You can also find this code snippet in my repo sboh1214/HwpKit

In hex, stream "FileHeader" should read as ascii (about in position 16112)

48 57 50 20 44 6F 63 75 6D 65 6E 74 20 46 69 6C 65 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 04 03 00 05
HWP Document File ................ 4 3 0 5 (version codes of hwp file) --> Omitted below

Overall, thanks for your library!

[MaxDesiatov]

Thanks for checking out this library! If I understand correctly, in HwpKit in HwpFileHeader initializer you're trying to read the signature, but there isn't enough data in the corresponding DataReader. Here on the screenshot byteOffset is already 255, but you're trying to read 32 more bytes, and the total length of data is 256 bytes.

Thanks for pointing me to this issue, DataReader should check for the length by itself. I'll add preconditions to DataReader to make this error obvious.

[MaxDesiatov]

Also, the error is specifically in this snippet in HwpFileHeader:

        data = dataReader.readDataToEnd()
        signature = String(data: dataReader.readData(ofLength: 32), encoding: .utf16LittleEndian) ?? "Error"

You've used readDataToEnd() once, which set the byteOffset value (the reading "cursor") to the end of the stream. There's nothing left to read, so you can't use readData again on this stream unless you use seek to rewind the reading position to somewhere else where data is available. Or maybe you need to create a different stream from other entry altogether to get the data you're looking for, I don't know the details of your format to say for sure.

Does that resolve your issue?

[sboh1214]

Thank you for your kind answer and for investigating my error!
However, there is one more question as I wrote above.

I open this file with ole-py, which is another ole/com tool written in python.
You can see code in this repo.
https://github.com/sboh1214/ole-py/blob/master/test.py

As a result, the "FileHeader" stream of .hwp file starts with the signature "HWP Document File" encoded with ASCII.

('FileHeader', 'DocInfo', '\x05HwpSummaryInformation', 'PrvImage', 'PrvText', 'Scripts/JScriptVersion', 'Scripts/DefaultJScript', 'DocOptions/_LinkDoc', 'BinData/BIN0001.png', 'BodyText/Section0')
========================================
HWP Document File

However, when I use OLEKit, I have a result below.
https://github.com/sboh1214/HwpKit/blob/646014bb13910badc6f4dbc6cc88589d249495dd/Tests/HwpKitTests/HwpKitTests.swift#L35

Optional("´¥\0­\u{14}zøÄf�"\u{12}ÍA�tgu�9¹\u{19}�>�\u{1E}\u{1D}W\u{1A}\rñ£")

Can you help me what is wrong? Sorry for using the infamous file format :)

[MaxDesiatov]

Thanks for the update, this looks like a bug to me, I'm investigating now...