人生在世，看得透，又看得远者prevail everywhere. v.gd/YLHADG v.gd/cRqvfD v.gd/bKvfFH v.gd/sRm3tA: How to use git over an HTTP proxy, with socat

Tuesday, 5 February 2013

How to use git over an HTTP proxy, with socat

Corporate firewalls and proxy typically block both of these (and often a lot more); here's how to get around them.
If you're tracking a public repo, you will need to use the "git" protocol, because the "http" protocol is not very efficient, and/or requires some special handling on the server side. If you're pushing code to a public repo, you definitely need to use the "ssh" protocol.

a word about socat

I will be using "socat", an absolute corker of a program that does so many things it's incredible! Other people use corkscrew, ssh-https-tunnel, etc., which are all specialised for just one purpose. I prefer socat, and once you spend the 2-3 years :-) needed to read the man page, you will see why!
The basic idea is that you will somehow invoke socat, which will negotiate with the HTTP(S) proxy server using the CONNECT method to get you a clean pipe to the server on the far side.
However, do note that socat does have one disadvantage: the passwords to your proxy server are visible in to local users running ps -ef or something. I don't care since I don't have anyone else logging into my desktop, and the ability to use a program I already have anyway (socat) is more important.

proxying the git protocol

When I want to download a public repo, I just type

proxied_git clone ...repo...
proxied_git pull

and so on, instead of

git clone ...repo...
git pull

Here's the how and why of it.
To proxy the git protocol, you need to export an environment variable called GIT_PROXY_COMMAND, which contains the command that is to be invoked. I have a shell function in my .bashrc that looks like this:

proxied_git () 
( 
    export GIT_PROXY_COMMAND=/tmp/gitproxy;

    cat  > $GIT_PROXY_COMMAND <<EOF
#!/bin/bash
/usr/bin/socat - PROXY:172.25.149.2:\$1:\$2,proxyport=3128
EOF
    chmod +x $GIT_PROXY_COMMAND;

    git "$@"
)

Possible variations are:

you could give /tmp/gitproxy a more permanent name and remove the middle pararaph completely. I don't do this because that's too small a file to bother with; it just seems cleaner this way)
you could permanently set the environment variable if all your git repos are remote (very unlikely)

One thing you cannot do is to roll the entire socat command into the environment variable. Git passes the host and port as two arguments to the proxy command, but socat expects them in the syntax you see above, so you will need to wrap it in a script as I have done. I guess you could argue that this is a point in favour of corkscrew etc. ;-)

proxying the ssh protocol

The git protocol is handled directly by git (duh!), but if you use the ssh protocol, it invokes ssh explicitly (again, duh!).
Ssh already has this sort of stuff built-in, so you simply add a few lines to your ~/.ssh/config

host gh
    user git
    hostname github.com
    port 22
    proxycommand socat - PROXY:your.proxy.ip:%h:%p,proxyport=3128,proxyauth=user:pwd

Now you can just say (for example):

git clone gh:sitaramc/git-notes.git

ssh proxy using corkscrew instead of socat

download and install corkscrew (http://www.agroman.net/corkscrew/)
create a file (eg., ~/.ssh/myauth) and put your http proxy username and password as "username:password" in it and save it.
safeguard the file
```
chmod 600 ~/.ssh/myauth
```

open ~/.ssh/config and add the following entry, adding an explicit path to corkscrew if needed.

host gh
    user git
    hostname github.com
    port 22
    proxycommand corkscrew your.proxy.ip 3128 %h %p ~/.ssh/myauth

extra coolness for github

Noting that many corporate firewalls block access to the CONNECT method on ports other than 443, the good folks at github have an ssh server listening on 443 if you use the host "ssh.github.com", so you can replace the hostname and the port in the above ssh config stanza as appropriate, and you're all set。

from http://sitaramc.github.com/tips/git-over-proxy.html

Total Pageviews