If I understand correctly [*1*] what you did during this "Pull" was to get the 
data from your home bare repo to your work machine for git to decode, and then 
for git to feed the results back to your home working repo, so taking a full 
ssh round trip (ssh encoded both ways).

That is, you maybe thought that the ssh excuted git on your home machine rather 
than simply giving you file access.

"Unlike a VNC server, SSH (alone) doesn't let you drive your computer remotely"

I could well be wrong as I only use ssh for my github connection.... I was just 
trying to learn the lessons. 
I've split your original into chunks below so I could pick out nuances.


----- Original Message ----- 
  From: John McKown 
  To: git-users@googlegroups.com 
  Sent: Thursday, December 13, 2012 3:10 PM
  Subject: [git-users] interesting? observation about "git mv" propagation

  This may not be a shock to most, but it was a real surprise to me. 

  I have two machines. One at work. One at home. 

  I am using git to keep a project in sync between the two. I have a working 
directory on both machines. 

  On my home machine, I have a separate "repository" which is on my home 
machine. So I do the changes in the "working directory" (either at work or at 
home) and "git push" and "git pull" to/from the repository to maintain 
synchronization. It works well. 

  I decided that I wanted to do some renames on my work machine. I had 55 text 
files which I decided I should have a ".txt" suffix on. So I did a "git mv" on 
each, adding the ".txt" to the end of the file name. 

  I did a "git commit" after each "git mv" detailing the renaming, similar 
to:"rename x to x.txt". I then did a "git push" to my repository (which did an 
SSH push to my home machine). Worked fairly fast, although the "git commit" was 
a bit slow. 

  Anyway, I then did an "ssh" onto my home machine. Did a "cd" to the working 
directory and did a "git pull". Oh my word. It seemed to take forever. 

  I did a second ssh to get another terminal session going. I guess that I 
should have known, but what actually happened was that the "git pull" deleted 
all the old versions of the files from my working directory, then expanded the 
compressed versions back with the new name. I guess I was expecting a simple 
rename. Not a delete / restore. 

  OK, people "in the know" may have expected this. But it was a complete shock 
to me. And I thought it might be of interest to other newbies.


Reply via email to