Corrupt input data

[kitsudog]

PARTS_01_SHIRTF.l2db.zip

CompressionHelper.py

  39:  try:
  40:      return dec.decompress(data[5:])
  41:  except:
  42:      return lzma.decompress(data)

it works.

[K0lb3]

Thanks for the report. The fix you proposed wouldn't fix the root of the issue, which is probably a false flag detection. I will look into it in the coming days.

[K0lb3]

I noticed that the data in question is compressed via lzma.FORMAT_ALONE which is the legacy lzma format.

So I checked how the properties are encoded in that format and found the following code

void DecodeProperties(const Byte *properties)
  {
    unsigned d = properties[0];
    if (d >= (9 * 5 * 5))
      throw "Incorrect LZMA properties";
    lc = d % 9;
    d /= 9;
    pb = d / 5;
    lp = d % 5;
    dictSizeInProperties = 0;
    for (int i = 0; i < 4; i++)
      dictSizeInProperties |= (UInt32)properties[i + 1] << (8 * i);
    dictSize = dictSizeInProperties;
    if (dictSize < LZMA_DIC_MIN)
      dictSize = LZMA_DIC_MIN;
  }

which is pretty much the same as the current python code besides the min dict size.

So it's better to simply use lzma.decompress(data, format = lzma.FORMAT_ALONE) .

I'm going to run some tests now to confirm this.

[K0lb3]

Hm, according to my tests it's not exactly the same. I found a good way to identify the compression format tho.

When the ALONE format is used, data[5] equals 255, if the specific compressor is used, it equals 0.

I'm going to run some further tests and so how it goes with that.