did not start with 'ZV' signature bytes about compress HOT 5 CLOSED

As far as I know, first couple of bytes should be compatible with the command-line tool.
However I don't know of formal specification so it is difficult to say for sure.
I do know there are some implementations that do not do this.

Which platform is this on?

from compress.

Also: it would be interesting to see what first couple of bytes are. I think other codecs leave out Z and V but then follow with expected type byte (0x00 or 0x01).

from compress.

I am not actually using command line tool; I am compressing from withing my application (linux) using the original reference lzf implementation (http://oldhome.schmorp.de/marc/liblzf.html). Your documentation seems to suggest interoperability with that.

from compress.

@usergoodvery to the best of my knowledge, yes, original codec used this encoding. Command-line utilities use liblzf as well I think. As far as I know, Java algorithm used is indeed direct translation of original C version (at least starting point was); no changes were made to logic.

If you want further help, you would need to see what encoded results with embedded algorithm are.

from compress.

Thanks I'll peek at the first few bytes to get a feel of what's being stashed in there...

from compress.