3 ms·
I am working on a file archive tool, written in go, to add random file access to .tar.gz of arbitrary size while maintaining compatibility as a valid .tar.gz.
by light_cone 7y ago
I am working on a file archive tool, written in go, to add random file access to .tar.gz of arbitrary size while maintaining compatibility as a valid .tar.gz.
The idea is to embed in gzip metadata the tar file index, and compress the gzip using small chunks of 1MB, whose offsets are also indexed in the metadata.
Since the indexes are written in metadata fields, it is invisible for normal linear decompression.
- bouk 7y agoNeat idea. Would this make it possible to use a tar gz as a squashfs archive, basically?
- light_cone 7y agoThanks! I did not know squashfs, but yes it seems the same idea... Thank you for your input, I think it might be of greater value if I could index enough data to provide a FUSE filesystem instead of a simple archive utility. That's a great idea, I'll look into it.