Hello. I would like to access files in tar.gz and tar.bz2 packages without untarring the files to disk. This done as a loop disk device would make the access transparent.
Has that been done yet? Some design notes: The disk files are read in blocks. The file-block query will be <file> <read offset> <file-block size> That can be done now by tar xzOf package.tar.gz <filepath> | tail --bytes +<read offset> | head --bytes <file-block size> but that takes too much time. Can gzip data be uncompressed starting from the middle of file? Unknown. Can bzip2 data be uncompressed starting from the middle of file? Yes. If "yes", then we may create an file/block index for faster access. But now "yes" answers does not help because tar cannot find the file headers. Also, in the case of tar.bz2 packages, bzip2recover extracts all the blocks, not a given range of blocks. Two programs should help: -zcat/bczat with ability to start from the middle of file -tar with ability to start from the middle of file If the files inside the tar package are large we also need a way to seek in the tar package. Juhana
