Decompress tar.lz tar.bz2 cause exception

[zhuxb711]

I tried to decompress the tar.bz2 or tar.lz that created from the SharpCompress

But unfortunately, I got the exception:
Cannot determine compressed stream type. Supported Archive Formats: Zip, Rar, 7Zip, GZip, Tar

StackTrace:
    at SharpCompress.Archives.ArchiveFactory.FindFactory[T](Stream stream) in /_/src/SharpCompress/Archives/ArchiveFactory.cs:line 164
    at SharpCompress.Archives.ArchiveFactory.Open(Stream stream, ReaderOptions readerOptions) in /_/src/SharpCompress/Archives/ArchiveFactory.cs:line 24

[zhuxb711]

tar & tar.gz works fine

[adamhathcock]

How are you access those formats? FileStreams or another stream?

[zhuxb711]

Use FileStream directly @adamhathcock

[adamhathcock]

I think the problem is that bzip2 and lzip only work for forward only streams and there is no archive implementation for either.

GZip works mainly because it's got a single entry.

What is the rest of the code do you have? Try ReaderFactory instead?

[zhuxb711]

But Archive API & Reader API should be both supported as your FORMAT.md indicated

[zhuxb711]

Actually, I use the ReaderFactory to extract the files, but before that I would use the ArchiveFactory to open the file and extract the TotalUncompressSize. I think even "forward only streams" should be able to calculate the TotalUncompressSize. Any suggestion? An alternative way is to walk the Reader first to get TotalUncompressSize and walk the Reader again to extract the files. Will it cost to much?

using (IArchive Archive = ArchiveFactory.Open(SourceStream, Options))
{
    ulong TotalSize = Convert.ToUInt64(Archive.TotalUncompressSize);
}

[zhuxb711]

Also I need the ArchiveFactory to open the stream first to determine whether the archive is solid or 7zip. Which could not be opened from ReaderFactory. This is how my code works:

using (IArchive Archive = ArchiveFactory.Open(SourceStream, Options))
{
    ulong TotalSize = Convert.ToUInt64(Archive.TotalUncompressSize);

    if (SourceStream.CanSeek && SourceStream.Position > 0)
    {
        SourceStream.Seek(0, SeekOrigin.Begin);
    }

    if (Archive.IsSolid || Archive.Type == ArchiveType.SevenZip)
    {
        using (IReader ArchiveReader = Archive.ExtractAllEntries())
        {
            await ExtractCoreAsync(ArchiveReader, ExtractFolder, TotalSize, ProgressHandler, CancelToken); // TotalUncompressSize would be used in ExtractCoreAsync for progress calculation
        }
    }
    else
    {
        using (IReader ArchiveReader = ReaderFactory.Open(SourceStream, Options))
        {
            await ExtractCoreAsync(ArchiveReader, ExtractFolder, TotalSize, ProgressHandler, CancelToken); // TotalUncompressSize would be used in ExtractCoreAsync for progress calculation
        }
    }
}

[zhuxb711]

@adamhathcock Any suggestion?

[adamhathcock]

You have several issues noted that are separate:

Actually, I use the ReaderFactory to extract the files, but before that I would use the ArchiveFactory to open the file and extract the TotalUncompressSize. I think even "forward only streams" should be able to calculate the TotalUncompressSize. Any suggestion? An alternative way is to walk the Reader first to get TotalUncompressSize and walk the Reader again to extract the files. Will it cost to much?

using (IArchive Archive = ArchiveFactory.Open(SourceStream, Options))
{
ulong TotalSize = Convert.ToUInt64(Archive.TotalUncompressSize);
}

ReaderFactory is forward-only and can't get the total size without walking the stream. ArchiveFactory (seekable streams) for zip uses the dictionary. Different formats work differently so I can't answer this.

Also I need the ArchiveFactory to open the stream first to determine whether the archive is solid or 7zip. Which could not be opened from ReaderFactory. This is how my code works:

using (IArchive Archive = ArchiveFactory.Open(SourceStream, Options))
{
ulong TotalSize = Convert.ToUInt64(Archive.TotalUncompressSize);

if (SourceStream.CanSeek && SourceStream.Position > 0)
{
    SourceStream.Seek(0, SeekOrigin.Begin);
}

if (Archive.IsSolid || Archive.Type == ArchiveType.SevenZip)
{
    using (IReader ArchiveReader = Archive.ExtractAllEntries())
    {
        await ExtractCoreAsync(ArchiveReader, ExtractFolder, TotalSize, ProgressHandler, CancelToken); // TotalUncompressSize would be used in ExtractCoreAsync for progress calculation
    }
}
else
{
    using (IReader ArchiveReader = ReaderFactory.Open(SourceStream, Options))
    {
        await ExtractCoreAsync(ArchiveReader, ExtractFolder, TotalSize, ProgressHandler, CancelToken); // TotalUncompressSize would be used in ExtractCoreAsync for progress calculation
    }
}

}

This was attempted to be addressed here: #1070

[zhuxb711]

Seems PR 1077 did not provide a way to get IReader without checking

Archive.IsSolid || Archive.Type == ArchiveType.SevenZip

[adamhathcock]

The issue says you can do the following without that check:

using var archive = ArchiveFactory.Open(testArchive);
        archive.WriteToDirectory(SCRATCH_FILES_PATH, options);