Migration from SVN to GIT

Readme.md

Migration from SVN to Git

Migration is not always easy. However, git-svn is a quite useful tool to do the job.

To do the job right, this post from stackoverflow is tremendously helpful in my entire process.

I wrote the m.py to automate the entire process.

m.py

#!/usr/bin/env python
"""
m.py: Migrate svn to git. 
Beware that this version ignore the initial direction creation commit.
"""
__author__ = "Zhenyang Hua"
__license__ = "MIT"
__version__ = "0.0.1"
__maintainer__ = "Zhenyang Hua"
__email__ = "edison.hua@gmail.com"
__status__ = "Development"

from subprocess import (PIPE, Popen)
import os
import shutil

"""
Settings
"""
# Don't miss the trailing slash
svnUsername = "svn username"
svnRemoteRepo = "url to svn remote repo"
gitRemoteRepo = "url to git remote repo"

"""
Program starts here
"""
root = os.path.dirname(os.path.abspath(__file__))
svnUrlParts = svnRemoteRepo.split('/')
svnLocalRepo = "./" + svnUrlParts[len(svnUrlParts) - 2]
gitUrlParts = gitRemoteRepo.split('/')
gitLocalRepo = "./" + gitUrlParts[len(gitUrlParts) - 2]
starting_revision = ""

def invoke(command):
	return Popen(command, stdout=PIPE, shell=True).stdout.read()

def getSvnRepo():
	co = "svn co " + svnRemoteRepo
	invoke(co)

def getAuthors():
	os.chdir(svnLocalRepo)
	lines = invoke("svn log -q").split('\n')
	authorList = []
	revisionList = []
	for line in lines:
		entry = line.split(' | ')
		if (len(entry) > 1):
			name = entry[1]
			author = name + " = " + name + ' <' + name + '@woolpert.com>'
			authorList.append(author)
			revision = entry[0]
			revisionList.append(revision)
	authorList = sorted(list(set(authorList)));
	with open("../author_list.txt", "w") as authorListTxt:
		for author in authorList:
			authorListTxt.write(author)
	global starting_revision
	starting_revision = revisionList[len(revisionList) - 2][1:]

def checkAuthors():
	editingWarning = "> Please make sure you have updated the author_list.txt, enter Y to confirm: "
	response = raw_input(editingWarning)
	if response.upper() == "Y":
		print "\n> author_list.txt was confirmed.\n"
	else:
		checkAuthors()

def cloneSvnRepo():
	os.chdir(root)
	cloneRepo = "git svn clone -r " + starting_revision + " --stdlayout --username=" + svnUsername + " --authors-file=author_list.txt --no-minimize-url " + svnRemoteRepo + " " + gitLocalRepo
	invoke(cloneRepo)
	os.chdir(gitLocalRepo)
	rebase = "git svn rebase"
	invoke(rebase)

def pushToGit():
	os.chdir(root)
	os.chdir(gitLocalRepo)
	addRemote = "git remote add origin " + gitRemoteRepo
	invoke(addRemote)
	pushRepo = "git push origin master"
	invoke(pushRepo)

def onerror(func, path, exc_info):
	"""
	Error handler for ``shutil.rmtree``.

	If the error is due to an access error (read only file)
	it attempts to add write permission and then retries.

	If the error is for another reason it re-raises the error.

	Usage : ``shutil.rmtree(path, onerror=onerror)``
	"""
	import stat
	if not os.access(path, os.W_OK):
		# Is the error an access error ?
		os.chmod(path, stat.S_IWUSR)
		func(path)
	else:
		raise

def cleanup():
	os.chdir(root)
	if os.path.isdir(svnLocalRepo):
		shutil.rmtree(svnLocalRepo, onerror=onerror)
	if os.path.isdir(gitLocalRepo):
		shutil.rmtree(gitLocalRepo, onerror=onerror)
	if os.path.exists("./author_list.txt"):
		os.remove("./author_list.txt")

def main():
	print "\n----Start----\n"

	print "1. Checking out the svn repository...\n"
	getSvnRepo()

	print "2. Retrieving a list of all Subversion committers...\n"
	getAuthors()

	print "3. Confirming username and email of the committers\n"
	checkAuthors()

	print "4. Cloning the Subversion repository using git-svn...\n"
	cloneSvnRepo()

	print "5. Pushing the local repository to the remote...\n"
	pushToGit()

	print "6. Cleaning up...\n"
	cleanup()

	print "----End----\n"

if __name__ == "__main__":
	main()