Is a system for managing and tracking changes to files
Your *nix flavoured machine is nothing but a collection of files
Have you ever lost work, because of machine failure before?
Distributed Version Control
Distributed
Everyone has the entire history on their machines
Allows for work off-line
Allows for protection from single points of failture
Git consists of
Repositories: a collection of files and their history
History: Each change made in the history has a unique identifier (commits)
Branches: carbon copies of of these files
Repositories
A folder on your machine
Like PISAE,PISOA, or PIAPP
There will be one for Data Science
Please ...
Make lots of little small commits
Limit the changes in a commit to a single bug fix/new feature
Don't push broken code
Cloning a repository
In git speak one "clones" a repository
This creates a local copy of a repository
git clone git://github.com/mischat/Test.git
Now you should have a folder on your machine called Test
History - Changing a file
Once you have cloned the repository, make a change to the README file
Once you have made a change run git status
This tells you what changes have been made to the repo
melgo:Test mmt$ git status
# On branch master
# Changes not staged for commit:
# (use "git add ..." to update what will be committed)
# (use "git checkout -- ..." to discard changes in working directory)
#
# modified: README
#
no changes added to commit (use "git add" and/or "git commit -a")
melgo:Test mmt$ git add README
melgo:Test mmt$ git status
# On branch master
# Changes to be committed:
# (use "git reset HEAD ..." to unstage)
#
# modified: README
#
Useful shortcut, but BE CAREFUL
git add .
History - Create a "commit" Step 2 - commit
git commit
# Please enter the commit message for your changes. Lines starting
# with '#' will be ignored, and an empty message aborts the commit.
# On branch master
# Changes to be committed:
# (use "git reset HEAD ..." to unstage)
#
# modified: README
#
Another shortcut git commit -m "Was asked to say hello to PeerIndex"
Set your prefered editor using the environment variable GIT_EDITOR
Now you have committed the code to your own repository
melgo:Test mmt$ git log
commit 20176197627ef2a9f5f9ee304b04f218c6b8d8a0
Author: Mischa Tuffield
Date: Fri Feb 17 13:55:45 2012 +0000
Was asked to say hello to PeerIndex
commit efe088f34adc029da74dd72e9e0625d64251dab4
Author: Mischa Tuffield
Date: Fri Feb 17 13:46:21 2012 +0000
Initial Commit to the Repo
You can ask for the history of a given file
git log lame.txt
History - show changes in a given commit
Show what a commit changed
git show 20176197627ef2a9f5f9ee304b04f218c6b8d8a0 [optionalfilename]
melgo:Test mmt$ git show 20176197627ef2a9f5f9ee304b04f218c6b8d8a0
commit 20176197627ef2a9f5f9ee304b04f218c6b8d8a0
Author: Mischa Tuffield
Date: Fri Feb 17 13:55:45 2012 +0000
Was asked to say hello to PeerIndex
diff --git a/README b/README
index ebf366d..9e658d9 100644
--- a/README
+++ b/README
@@ -1 +1 @@
-Hello World :)
+Hello PeerIndex :)
History - Exactly who changes what line when
git blame filename
6a998f24 (Alister Bulman 2011-09-13 13:56:35 +0000 33) role :app, "api.viewsflow.com", "api3.viewsflow.com"
^b688b4a ([email protected] 2011-06-06 10:29:44 +0000 34)
b6c800d5 ([email protected] 2011-07-01 17:24:17 +0000 35) # adding other known process servers as they come
^b688b4a ([email protected] 2011-06-06 10:29:44 +0000 36) # this is so we can # cap deploy, > with process !deploy
16ba0bdc ([email protected] 2011-06-23 14:51:48 +0000 37) # process1 aka process(4/9)
16ba0bdc ([email protected] 2011-06-23 14:51:48 +0000 38) # process2 aka process(2/8)
6a998f24 (Alister Bulman 2011-09-13 13:56:35 +0000 39) #role :process, "process.viewsflow.com", "process2.viewsflow.com"
b6c800d5 ([email protected] 2011-07-01 17:24:17 +0000 40)
^b688b4a ([email protected] 2011-06-06 10:29:44 +0000 41) # 46.51.130.109 worker, larry
^b688b4a ([email protected] 2011-06-06 10:29:44 +0000 42) # 79.125.64.99 worker, scruffy
^b688b4a ([email protected] 2011-06-06 10:29:44 +0000 43) # 46.137.54.254 worker, panucci
b6c800d5 ([email protected] 2011-07-01 17:24:17 +0000 44) # 79.125.88.69 worker, calculon
b6c800d5 ([email protected] 2011-07-01 17:24:17 +0000 45) # 46.137.133.93 worker, crushinator
History - Refreshing your repo to the latest checout
Please NEVER use git pull
And try to use git pull --rebase or git up
This doesn't allow you to
Using --rebase, you might not want to commit
Make use of git stash and git pop
As you wont be able to pull unless you have NO uncommitted changes
Revert
This unapplies a commit made
git revert [SHA1 of a commit]
Merging
Demo :)
Branching
Branches are Lines of development/changes
If anything, a branch is a directed acyclic graph of development rather than a line.
Create branches for new features, for significant changes
Branches are your friends - use them
Allows one to conceptually split of a thread of work
So now A, B, C, D, E, F, G and H are on "stable", while A, B, D, F and I are on "new-idea".
Resetting work in a branch
last version from another repository
|
v
M---N-----O----P---Q ("master")
Resetting work in a branch 2
git branch dubious-experiment
M---N-----O----P---Q ("master" and "dubious-experiment")
Resetting work in a branch 3
git checkout master
# Be careful with this next command: make sure "git status" is
# clean, you're definitely on "master" and the
# "dubious-experiment" branch has the commits you were working
# on first...
git reset --hard [SHA1sum of commit N]
("master")
M---N-------------O----P---Q ("dubious-experiment")
This reverts the master branch ignoring the dubious experiment
If you find that you would like to select commits from a different branch, and merge into a different branch without having to merge the whole lot, you can cherry-pick git commits individually.
