ePOM - Fundamentals of Research Software Development - Code Version Control

CODE VERSION CONTROL
GIUSEPPE MASETTI
ESCI 872 – APPLIED TOOLS FOR OCEAN MAPPING – FUNDAMENTALS OF RESEARCH SOFTWARE DEVELOPMENT
Durham, NH – November 19 & 21, 2019
V1

WHAT IS CODE VERSION CONTROL?
A mechanism to manage changes to code over time.
• Tracking of per-file creation/modifications/deletion
• Ability to switch to past versions to fix bugs
• Facilitating concurrent work on the same project
Help to protect code from both catastrophes
and casual introduction of human error!

VERSION CONTROL SYSTEM (VCS)
Also known as:
Source Code Management (SCM)
Revision Control System (RCS)
CVS, Subversion, Mercurial, Bazaar, Git, etc.
Essential part of the every-day workflow
of a modern research software development team!

GIT .
• Command-line tool that tracks changes in
files and eases collaboration
• Created in 2005 by Linus Torvalds
• Free and open-source → GNU GPL v2
• Goals:
• Handling large project with speed/efficiency
• Data integrity
• Support for non-linear, distributed development
GFDL. Permission of Martin Streicher, Editor-in-
Chief, LINUXMAG.com. CC BY-SA 3.0

SVN VS. GIT
Git
Repo
Git
Repo
Git
Repo
SVN Central
Repo
Working
Copy
Working
Copy
Working
Copy
Centralized VCS Distributed VCS

2019 RhodeCode, https://rhodecode.com/insights/version-control-systems-2016

Website → https://git-scm.com/

OPEN A WINDOWS SHELL,
THEN RETRIEVE THE GIT’S VERSION.

GIT SETUP
• Configure Git with your name and email
• git config --global user.email "your@email.com"
• git config --global user.name "your name"
• Check the current configuration
• git config --list
• The config commands only need to be done once

IN A WINDOWS SHELL
CONFIGURE GIT
WITH YOUR NAME AND EMAIL.

USING GIT
• Creating a new project
• git init → convert a directory into a Git repository (a .git directory is created)
• git add . → a snapshot of the directory content in a temporary staging area
• git commit –m "First commit" → permanently store the snapshot
• The init commands only need to be done once per project.

IN A WINDOWS SHELL
CREATE A GIT PROJECT, THEN
CREATE AND COMMIT A README FILE.

USING GIT
• Making code changes
• git add file0.py file1.py → stage new/updated files
• git diff → to list un-staged changes
• git diff --cached → to see what is ready to be committed
• git status → a brief summary of the repository situation
• git commit –m "A meaningful message" → finally commit your changes

USING GIT
• Making code changes
• git diff → to list un-staged changes
• git add file0.py file1.py → stage new/updated files
• git diff --cached → to see what is ready to be committed
• git status → a brief summary of the repository situation
• git commit –m "A meaningful message" → finally commit your changes
• When to commit code?
• Do it frequently
• Whenever you have reached a milestone/step in your task

FIRST CREATE AND COMMIT TWO SCRIPT FILES,
THEN EDIT ONE OF THEM AND
COMMIT THE CHANGES.

LOCAL CLOUD
Working
Directory
Staging
Area
Local
Repository
Remote
Repository
git add
git commit
git push
git fetch
git checkout
git merge

GIT HOSTING SERVICES
• Several sites offer services for hosting Git repositories
• Popular ones that are free for open-source projects:

GITHUB
• A social network for code share/collaboration
• It was acquired by Microsoft in 2018 for $7.5 billion.

GITHUB
• A social network for code share/collaboration
• It was acquired by Microsoft in 2018 for $7.5 billion.
• To push to the GitHub repository:
• Setup a remote (just once):
• git remote add origin https://github.com/giumas/myproject.git
• Push committed code (multiple times):
• git push origin master
• Reload project webpage

PUSH THE PREVIOUSLY CREATED PROJECT CODE
TO A GIT HOSTING SERVICE.

EPOM SERVER → JUPYTER TERMINAL

EPOM SERVER → CLONING GIT REPO
git clone https://github.com/giumas/myproject.git

EPOM SERVER → GIT REPO HISTORY
git log

ON THE EPOM SERVER
CLONE YOUR GIT REPOSITORY,
THEN EXECUTE ONE OF THE SCRIPTS.

LOCAL MACHINE GITHUB
Git
Repository
Git
Repository
git push
EPOM SERVER
git clone
Git
Repository
git push
git pull
git push
git pull
git push

COMMIT CHANGES ON YOUR LOCAL MACHINE,
THEN RETRIEVE THEM ON THE EPOM SERVER.

https://jrebel.com/rebellabs/git-commands-and-best-practices-cheat-sheet/

GIT GUI CLIENT → TORTOISEGIT
A windows shell extension to work with Git repositories.

• Shell integration
• Icon overlays
• Easy access to Git commands

TORTOISEGIT → ICON OVERLAYS
• Normal → all changes committed
• Modified → changes are present
• Staged → changes are ready to be committed
• Deleted → file/folder scheduled to be deleted
• Added → file/folder scheduled to be added
• Ignored → file/folder ignored by Git
• Conflict → there is a conflict between code versions
https://tortoisegit.org/docs/tortoisegit/tgit-dug-wcstatus.html#tgit-dug-wcstatus-1

TORTOISEGIT → SWITCH TO PAST COMMITS

USING TORTOISEGIT
SWITCH TO A PAST COMMIT,
THEN EXECUTE THE RESULTING SCRIPT.

QUESTIONS?
Contact me at: gmasetti@ccom.unh.edu

ePOM - Fundamentals of Research Software Development - Code Version Control

More Related Content

What's hot

Similar to ePOM - Fundamentals of Research Software Development - Code Version Control

More from Giuseppe Masetti

Recently uploaded

ePOM - Fundamentals of Research Software Development - Code Version Control