zpaq: using local .0 to speed extract from remote .1 file

[barrycarter]

I did 'zpaq add "archive.?" files' to create archive.0 and archive.1,
and then remote offsited archive.1, but kept archive.0 local. The
remote machine is sshfs mounted locally.

I now want to 'zpaq extract /mnt/sshfs/archive.1 filename', but the
sshfs link is slow so I want to read as few bytes as possible from the
remote machine.

My question: can I use the local archive.0 to reduce the number of
remote archive.1 bytes I read?

I know archive.0 contains some information about archive.1 (eg, the
table of contents), but is it enough to speed up extracting a file
from archive.1?

[Matt Mahoney]

The extract command doesn't read archive.0 because all of the information can be found in the index of archive.1. To read the index, it reads the 104 byte header, then seeks to the end to read the index information. It does this for each separate part (archive.2, etc). Then it reads only the data blocks that it needs to extract the requested files.

I guess the question is how big is archive.0 and how much time would it save by reading it locally?