-----BEGIN PGP SIGNED MESSAGE-----
Il 13/12/2012 18:58, John McKown ha scritto:
> I guess I wasn't clear.
> 1) on work machine, do "git mv" on 55 files. This is fairly fast.
> 2) on work machine, do "git push" to the repository.
> This pushes the data to the repository on my home machine via ssh.
> This went quickly
> 3) on work machine, use ssh to bring up a Linux terminal session at home.
> 4) on home machine, do a "git pull" to pull from the repository on my
> home machine. This went slowly.
> From watching what was happening by using a second terminal session on
> my home machine, I noticed that all 55 of the existing files were
> deleted. Then all 55 were uncompressed/restored from the repository (on
> the home machine). And these 55 files total up to 13 gig. Each is about
> 2-3 meg, except for one which is 1.4 gig. Home machine is a Core i7
> running 2.66Ghz.
Git, as explained in the documentation, does not track files renaming.
However it "knows" that in your case they are the same: this is the
reason why the push was fast; your home repo already have the objects
with the same SHA.
pull is slow because git has to unpack them; and yes, probably git could
be a bit more smart and check if an object with the same name/file
mode/SHA already exists in the working directory, bit I'm not sure.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
-----END PGP SIGNATURE-----