rardecode: corrupt ppm data

[shaybix]

Get this error with some of the files approximately a dozen or so out of 130+ rar files.

This is the go version i am using currently:

go version go1.6.2 darwin/amd64

and below is the code responsible for decoding the rar files using rardecode:

   rarfile, err := os.Open(filepath.Join("downloads", f))
    if err != nil {
        return err
    }

    rdr, err := rardecode.NewReader(rarfile, "")
    if err != nil {
        return err
    }

    nf, err := rdr.Next()
    if err != nil {
        return err
    }

    ps := make([]byte, nf.UnPackedSize)
    _, err = rdr.Read(ps)
    if err != nil {
        return err
    }
    newbok, err := os.Create(filepath.Join("bok", bokFile))
    defer newbok.Close()
    if err != nil {
        return err
    }

    _, err = io.CopyBuffer(newbok, rdr, ps)
    if err != nil {
         return err
    }

While this worked for the vast majority of the files, however it doesn't for dozen or so files. Vast majority getting the above mentioned error, and a couple get another but I'll leave that for another issue.

Am i doing it wrong, or is there a better way?

Thank you.

[nwaples]

No, it looks to be an error in my code somewhere.
I was able to produce the same error by compressing some large randomly generated text files. Its still possible yours is caused by a different error in the code, but Ill have a look and see if I can fix this first.

[shaybix]

Thanks a lot.

I'll do some more testing and see if i can come up with anything, but first need to wrap my head around rar specs.

[nwaples]

There's not a lot of documentation. Ive mostly had to rely on the free unrar c++ sourcecode as the standard. And its not pretty.

When I implemented the ppm text decompression I skipped the memory allocator for it. It was hard enough understanding it all (I still don't really understand the algorithm). It worked ok for the small files that I was testing, as I assumed unlimited memory. But it breaks for the larger files when the decode reaches the memory limits.
I think I have more of an idea how it works now, so Im having a go at hacking something together to give the correct behaviour. Its just not as easy as I'd hoped.

[nwaples]

The latest patch may fix the corrupt ppm data problems.

[shaybix]

I'll be testing it this week and will let you know how that went.

[shaybix]

The latest path resolved this issue. Thanks mate.