ansible/library/source_control/git

#!/usr/bin/python
# -*- coding: utf-8 -*-

# (c) 2012, Michael DeHaan <michael.dehaan@gmail.com>
#
# This file is part of Ansible
#
# Ansible is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# Ansible is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with Ansible.  If not, see <http://www.gnu.org/licenses/>.

DOCUMENTATION = '''
---
module: git
author: Michael DeHaan
version_added: "0.0.1"
short_description: Deploy software (or files) from git checkouts
description:
    - Manage I(git) checkouts of repositories to deploy files or software.
options:
    repo:
        required: true
        aliases: [ name ]
        description:
            - git, SSH, or HTTP protocol address of the git repository.
    dest:
        required: true
        description:
            - Absolute path of where the repository should be checked out to.
    version:
        required: false
        default: "HEAD"
        description:
            - What version of the repository to check out.  This can be the
              full 40-character I(SHA-1) hash, the literal string C(HEAD), a
              branch name, or a tag name.
    accept_hostkey:
        required: false
        default: false
        version_added: "1.5"
        description:
            - Add the hostkey for the repo url if not already added.
              If ssh_args contains "-o StrictHostKeyChecking=no", this
              parameter is ignored.
    ssh_opts:
        required: false
        default: None
        version_added: "1.5"
        description:
            - Creates a wrapper script and exports the path as GIT_SSH
              which git then automatically uses to override ssh arguments.
              An example value could be "-o StrictHostKeyChecking=no"
    key_file:
        requird: false
        default: None
        version_added: "1.5"
        description:
            - Uses the same wrapper method as ssh_opts to pass 
              "-i <key_file>" to the ssh arguments used by git
    reference:
        required: false
        default: null
        version_added: "1.4"
        description:
            - Reference repository (see "git clone --reference ...")
    remote:
        required: false
        default: "origin"
        description:
            - Name of the remote.
    force:
        required: false
        default: "yes"
        choices: [ "yes", "no" ]
        version_added: "0.7"
        description:
            - If C(yes), any modified files in the working
              repository will be discarded.  Prior to 0.7, this was always
              'yes' and could not be disabled.
    depth:
        required: false
        default: null
        version_added: "1.2"
        description:
            - Create a shallow clone with a history truncated to the specified
              number or revisions. The minimum possible value is C(1), otherwise
              ignored.
    update:
        required: false
        default: "yes"
        choices: [ "yes", "no" ]
        version_added: "1.2"
        description:
            - If C(yes), repository will be updated using the supplied
              remote.  Otherwise the repo will be left untouched.
              Prior to 1.2, this was always 'yes' and could not be disabled.
    executable:
        required: false
        default: null
        version_added: "1.4"
        description:
            - Path to git executable to use. If not supplied,
              the normal mechanism for resolving binary paths will be used.
    bare:
        required: false
        default: "no"
        choices: [ "yes", "no" ]
        version_added: "1.4"
        description:
            - if C(yes), repository will be created as a bare repo, otherwise
              it will be a standard repo with a workspace.
notes:
    - "If the task seems to be hanging, first verify remote host is in C(known_hosts).
      SSH will prompt user to authorize the first contact with a remote host.  To avoid this prompt, 
      one solution is to add the remote host public key in C(/etc/ssh/ssh_known_hosts) before calling 
      the git module, with the following command: ssh-keyscan remote_host.com >> /etc/ssh/ssh_known_hosts."
'''

EXAMPLES = '''
# Example git checkout from Ansible Playbooks
- git: repo=git://foosball.example.org/path/to/repo.git
       dest=/srv/checkout
       version=release-0.22

# Example read-write git checkout from github
- git: repo=ssh://git@github.com/mylogin/hello.git dest=/home/mylogin/hello

# Example just ensuring the repo checkout exists
- git: repo=git://foosball.example.org/path/to/repo.git dest=/srv/checkout update=no
'''

import re
import tempfile

def write_ssh_wrapper():
    fh = tempfile.NamedTemporaryFile(delete=False)
    wrapper_path = fh.name
    template = """#!/bin/sh
if [ -z "$GIT_SSH_OPTS" ]; then
    BASEOPTS=""
else
    BASEOPTS=$GIT_SSH_OPTS
fi

if [ -z "$GIT_KEY" ]; then
    ssh $BASEOPTS "$@"
else
    ssh -i "$GIT_KEY" $BASEOPTS "$@"
fi
"""
    fh.write(template)
    fh.close()
    st = os.stat(wrapper_path)
    os.chmod(wrapper_path, st.st_mode | stat.S_IEXEC)
    return wrapper_path

def set_git_ssh(ssh_wrapper, key_file, ssh_opts):

    if os.environ.get("GIT_SSH"):
        del os.environ["GIT_SSH"]
    os.environ["GIT_SSH"] = ssh_wrapper

    if os.environ.get("GIT_KEY"):
        del os.environ["GIT_KEY"]

    if key_file:
        os.environ["GIT_KEY"] = key_file    

    if os.environ.get("GIT_SSH_OPTS"):
        del os.environ["GIT_SSH_OPTS"]

    if ssh_opts:
        os.environ["GIT_SSH_OPTS"] = ssh_opts

def get_version(git_path, dest, ref="HEAD"):
    ''' samples the version of the git repo '''
    os.chdir(dest)
    cmd = "%s rev-parse %s" % (git_path, ref)
    sha = os.popen(cmd).read().rstrip("\n")
    return sha

def clone(git_path, module, repo, dest, remote, depth, version, bare, reference):
    ''' makes a new git repo if it does not already exist '''
    dest_dirname = os.path.dirname(dest)
    try:
        os.makedirs(dest_dirname)
    except:
        pass
    os.chdir(dest_dirname)
    cmd = [ git_path, 'clone' ]
    if bare:
        cmd.append('--bare')
    else:
        cmd.extend([ '--origin', remote, '--recursive' ])
        if is_remote_branch(git_path, module, dest, repo, version) \
        or is_remote_tag(git_path, module, dest, repo, version):
            cmd.extend([ '--branch', version ])
    if depth:
        cmd.extend([ '--depth', str(depth) ])
    if reference:
        cmd.extend([ '--reference', str(reference) ])
    cmd.extend([ repo, dest ])
    module.run_command(cmd, check_rc=True)
    if bare:
        os.chdir(dest)
        if remote != 'origin':
            module.run_command([git_path, 'remote', 'add', remote, repo], check_rc=True)
  
def has_local_mods(git_path, dest, bare):
    if bare:
        return False
    os.chdir(dest)
    cmd = "%s status -s" % (git_path,)
    lines = os.popen(cmd).read().splitlines()
    lines = filter(lambda c: not re.search('^\\?\\?.*$', c), lines)
    return len(lines) > 0

def reset(git_path, module, dest):
    '''
    Resets the index and working tree to HEAD.
    Discards any changes to tracked files in working
    tree since that commit.
    '''
    os.chdir(dest)
    cmd = "%s reset --hard HEAD" % (git_path,)
    return module.run_command(cmd, check_rc=True)

def get_remote_head(git_path, module, dest, version, remote, bare):
    cloning = False
    if remote == module.params['repo']:
        cloning = True
    else:
        os.chdir(dest)
    if version == 'HEAD':
        if cloning:
            # cloning the repo, just get the remote's HEAD version
            cmd = '%s ls-remote %s -h HEAD' % (git_path, remote)
        else:
            head_branch = get_head_branch(git_path, module, dest, remote, bare)
            cmd = '%s ls-remote %s -h refs/heads/%s' % (git_path, remote, head_branch)
    elif is_remote_branch(git_path, module, dest, remote, version):
        cmd = '%s ls-remote %s -h refs/heads/%s' % (git_path, remote, version)
    elif is_remote_tag(git_path, module, dest, remote, version):
        cmd = '%s ls-remote %s -t refs/tags/%s' % (git_path, remote, version)
    else:
        # appears to be a sha1.  return as-is since it appears
        # cannot check for a specific sha1 on remote
        return version
    (rc, out, err) = module.run_command(cmd, check_rc=True )
    if len(out) < 1:
        module.fail_json(msg="Could not determine remote revision for %s" % version)
    rev = out.split()[0]
    return rev

def is_remote_tag(git_path, module, dest, remote, version):
    cmd = '%s ls-remote %s -t refs/tags/%s' % (git_path, remote, version)
    (rc, out, err) = module.run_command(cmd, check_rc=True)
    if version in out:
        return True
    else:
        return False

def get_branches(git_path, module, dest):
    os.chdir(dest)
    branches = []
    cmd = '%s branch -a' % (git_path,)
    (rc, out, err) = module.run_command(cmd)
    if rc != 0:
        module.fail_json(msg="Could not determine branch data - received %s" % out)
    for line in out.split('\n'):
        branches.append(line.strip())
    return branches

def is_remote_branch(git_path, module, dest, remote, version):
    cmd = '%s ls-remote %s -h refs/heads/%s' % (git_path, remote, version)
    (rc, out, err) = module.run_command(cmd, check_rc=True)
    if version in out:
        return True
    else:
        return False

def is_local_branch(git_path, module, dest, branch):
    branches = get_branches(git_path, module, dest)
    lbranch = '%s' % branch
    if lbranch in branches:
        return True
    elif '* %s' % branch in branches:
        return True
    else:
        return False

def is_not_a_branch(git_path, module, dest):
    branches = get_branches(git_path, module, dest)
    for b in branches:
        if b.startswith('* ') and 'no branch' in b:
            return True
    return False

def get_head_branch(git_path, module, dest, remote, bare=False):
    '''
    Determine what branch HEAD is associated with.  This is partly
    taken from lib/ansible/utils/__init__.py.  It finds the correct
    path to .git/HEAD and reads from that file the branch that HEAD is
    associated with.  In the case of a detached HEAD, this will look
    up the branch in .git/refs/remotes/<remote>/HEAD.
    '''
    if bare:
        repo_path = dest
    else:
        repo_path = os.path.join(dest, '.git')
    # Check if the .git is a file. If it is a file, it means that we are in a submodule structure.
    if os.path.isfile(repo_path):
        try:
            gitdir = yaml.safe_load(open(repo_path)).get('gitdir')
            # There is a posibility the .git file to have an absolute path.
            if os.path.isabs(gitdir):
                repo_path = gitdir
            else:
                repo_path = os.path.join(repo_path.split('.git')[0], gitdir)
        except (IOError, AttributeError):
            return ''
    # Read .git/HEAD for the name of the branch.
    # If we're in a detached HEAD state, look up the branch associated with
    # the remote HEAD in .git/refs/remotes/<remote>/HEAD
    f = open(os.path.join(repo_path, "HEAD"))
    if is_not_a_branch(git_path, module, dest):
        f.close()
        f = open(os.path.join(repo_path, 'refs', 'remotes', remote, 'HEAD'))
    branch = f.readline().split('/')[-1].rstrip("\n")
    f.close()
    return branch

def fetch(git_path, module, repo, dest, version, remote, bare):
    ''' updates repo from remote sources '''
    os.chdir(dest)
    if bare:
        (rc, out1, err1) = module.run_command([git_path, 'fetch', remote, '+refs/heads/*:refs/heads/*'])
    else:
        (rc, out1, err1) = module.run_command("%s fetch %s" % (git_path, remote))
    if rc != 0:
        module.fail_json(msg="Failed to download remote objects and refs")

    if bare:
        (rc, out2, err2) = module.run_command([git_path, 'fetch', remote, '+refs/tags/*:refs/tags/*'])
    else:
        (rc, out2, err2) = module.run_command("%s fetch --tags %s" % (git_path, remote))
    if rc != 0:
        module.fail_json(msg="Failed to download remote objects and refs")
    (rc, out3, err3) = submodule_update(git_path, module, dest)
    return (rc, out1 + out2 + out3, err1 + err2 + err3)

def submodule_update(git_path, module, dest):
    ''' init and update any submodules '''
    os.chdir(dest)
    # skip submodule commands if .gitmodules is not present
    if not os.path.exists(os.path.join(dest, '.gitmodules')):
        return (0, '', '')
    cmd = [ git_path, 'submodule', 'sync' ]
    (rc, out, err) = module.run_command(cmd, check_rc=True)
    cmd = [ git_path, 'submodule', 'update', '--init', '--recursive' ]
    (rc, out, err) = module.run_command(cmd)
    if rc != 0:
        module.fail_json(msg="Failed to init/update submodules")
    return (rc, out, err)

def switch_version(git_path, module, dest, remote, version):
    ''' once pulled, switch to a particular SHA, tag, or branch '''
    os.chdir(dest)
    cmd = ''
    if version != 'HEAD':
        if is_remote_branch(git_path, module, dest, remote, version):
            if not is_local_branch(git_path, module, dest, version):
                cmd = "%s checkout --track -b %s %s/%s" % (git_path, version, remote, version)
            else:
                (rc, out, err) = module.run_command("%s checkout --force %s" % (git_path, version))
                if rc != 0:
                    module.fail_json(msg="Failed to checkout branch %s" % version)
                cmd = "%s reset --hard %s/%s" % (git_path, remote, version)
        else:
            cmd = "%s checkout --force %s" % (git_path, version)
    else:
        branch = get_head_branch(git_path, module, dest, remote)
        (rc, out, err) = module.run_command("%s checkout --force %s" % (git_path, branch))
        if rc != 0:
            module.fail_json(msg="Failed to checkout branch %s" % branch)
        cmd = "%s reset --hard %s" % (git_path, remote)
    (rc, out1, err1) = module.run_command(cmd)
    if rc != 0:
        if version != 'HEAD':
            module.fail_json(msg="Failed to checkout %s" % (version))
        else:
            module.fail_json(msg="Failed to checkout branch %s" % (branch))
    (rc, out2, err2) = submodule_update(git_path, module, dest)
    return (rc, out1 + out2, err1 + err2)

# ===========================================

def main():
    module = AnsibleModule(
        argument_spec = dict(
            dest=dict(required=True),
            repo=dict(required=True, aliases=['name']),
            version=dict(default='HEAD'),
            remote=dict(default='origin'),
            reference=dict(default=None),
            force=dict(default='yes', type='bool'),
            depth=dict(default=None, type='int'),
            update=dict(default='yes', type='bool'),
            accept_hostkey=dict(default='no', type='bool'),
            key_file=dict(default=None, required=False),
            ssh_opts=dict(default=None, required=False),
            executable=dict(default=None),
            bare=dict(default='no', type='bool'),
        ),
        supports_check_mode=True
    )

    dest      = os.path.abspath(os.path.expanduser(module.params['dest']))
    repo      = module.params['repo']
    version   = module.params['version']
    remote    = module.params['remote']
    force     = module.params['force']
    depth     = module.params['depth']
    update    = module.params['update']
    bare      = module.params['bare']
    reference = module.params['reference']
    git_path  = module.params['executable'] or module.get_bin_path('git', True)
    
    key_file   = module.params['key_file']
    ssh_opts   = module.params['ssh_opts']

    # create a wrapper script and export
    # GIT_SSH=<path> as an environment variable
    # for git to use the wrapper script
    ssh_wrapper = None
    if key_file or ssh_opts:
        ssh_wrapper = write_ssh_wrapper()
        set_git_ssh(ssh_wrapper, key_file, ssh_opts)

    # add the git repo's hostkey 
    if module.params['ssh_opts'] is not None:
        if not "-o StrictHostKeyChecking=no" in module.params['ssh_opts']:
            add_git_host_key(module, repo, accept_hostkey=module.params['accept_hostkey'])
    else:            
        add_git_host_key(module, repo, accept_hostkey=module.params['accept_hostkey'])

    if bare:
        gitconfig = os.path.join(dest, 'config')
    else:
        gitconfig = os.path.join(dest, '.git', 'config')

    rc, out, err, status = (0, None, None, None)

    # if there is no git configuration, do a clone operation
    # else pull and switch the version
    before = None
    local_mods = False
    if not os.path.exists(gitconfig):
        if module.check_mode:
            remote_head = get_remote_head(git_path, module, dest, version, repo, bare)
            module.exit_json(changed=True, before=before, after=remote_head)
        clone(git_path, module, repo, dest, remote, depth, version, bare, reference)
    elif not update:
        # Just return having found a repo already in the dest path
        # this does no checking that the repo is the actual repo
        # requested.
        before = get_version(git_path, dest)
        module.exit_json(changed=False, before=before, after=before)
    else:
        # else do a pull
        local_mods = has_local_mods(git_path, dest, bare)
        before = get_version(git_path, dest)
        if local_mods:
            # failure should happen regardless of check mode
            if not force:
                module.fail_json(msg="Local modifications exist in repository (force=no).")
            # if force and in non-check mode, do a reset
            if not module.check_mode:
                reset(git_path, module, dest)
        # exit if already at desired sha version
        remote_head = get_remote_head(git_path, module, dest, version, remote, bare)
        if before == remote_head:
            if local_mods:
                module.exit_json(changed=True, before=before, after=remote_head,
                    msg="Local modifications exist")
            else:
                module.exit_json(changed=False, before=before, after=remote_head)
        if module.check_mode:
            module.exit_json(changed=True, before=before, after=remote_head)
        fetch(git_path, module, repo, dest, version, remote, bare)

    # switch to version specified regardless of whether
    # we cloned or pulled
    if not bare:
        switch_version(git_path, module, dest, remote, version)

    # determine if we changed anything
    after = get_version(git_path, dest)
    changed = False

    if before != after or local_mods:
        changed = True

    # cleanup the wrapper script
    if ssh_wrapper:
        os.remove(ssh_wrapper) 

    module.exit_json(changed=changed, before=before, after=after)

# import module snippets
from ansible.module_utils.basic import *
from ansible.module_utils.known_hosts import *

main()