Fix LZO decompression

[Schamper]

Fixes fox-it/dissect.squashfs#11

[codecov]

Codecov Report

Merging #28 (8351190) into main (1e6dcd2) will increase coverage by 0.03%.
The diff coverage is 86.00%.

@@            Coverage Diff             @@
##             main      #28      +/-   ##
==========================================
+ Coverage   84.97%   85.00%   +0.03%     
==========================================
  Files          16       16              
  Lines        1118     1107      -11     
==========================================
- Hits          950      941       -9     
+ Misses        168      166       -2     

Flag	Coverage Δ
unittests	85.00% <86.00%> (+0.03%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
dissect/util/compression/lzo.py	82.85% <86.00%> (+0.14%)	⬆️

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[Miauwkeru]

Wouldn't something like this add a bit more readability to the whole thing?

LZO_VERSION_1 = 0x10
LZO_VERSION_2 = 0x11


def _determine_lzo_version(src: io.BytesIO, dst: bytearray) -> int:
    """Determines the LZO version of the src.

    Returns:
        The next value in LZO Stream
    """
    val = src.read(1)[0]
    if val == LZO_VERSION_1:
        raise ValueError("LZOv1")
    elif val > LZO_VERSION_2:
        # LZO is a stream
        dst += src.read(val - LZO_VERSION_2)
        val = src.read(1)[0]

        if val < LZO_VERSION_1:
            raise ValueError("Invalid LZO stream")
    return val

[Schamper]

This turns out to be not quite correct. See also https://docs.kernel.org/staging/lzo.html and https://github.com/torvalds/linux/blob/master/lib/lzo/lzo1x_decompress_safe.c.

[pyrco]

Maybe good to add these references.
I looked a bit at making the code more readable, but that is far from trivial. Maybe comment the if statements a bit about which byte encoding is indicated.

[Schamper]

Added them as references

[pyrco]

see comment

[Miauwkeru]

Would it be an idea to add some comments about which instruction encoding is used?
Including the first byte encodings

[Miauwkeru]

wouldn't this need an additional state variable?
As, according to the documentation you provided, state is already assigned at the first byte?

[Schamper]

state looks unused there: https://github.com/FFmpeg/FFmpeg/blob/master/libavutil/lzo.c#L157

[Schamper]

Would it be an idea to add some comments about which instruction encoding is used?
Including the first byte encodings

You seem to understand the algorithm better than me by now so feel free to make some suggestions 😄