Hi everybody,
I need to have a compressed file that I can random-access in chunks from an http server, using range requests (something like sozip but with faster decompression times); I am currently doing framing snappy compression splitting the original files in fixed chunks and producing a second file containing an index for the chunks; this works ok for my use case but... I have two files to move around.
I was thinking on using the padding chunk type to store the index directly inside the compressed file I generate, is this a bad idea? The spec says that padding should be all zero but, regardless, client must just skip it.