# -*- coding: utf-8 -*-

"""

    vcs.backends.git.repository

    ~~~~~~~~~~~~~~~~~~~~~~~~~~~

    Git repository implementation.

    :created_on: Apr 8, 2010

    :copyright: (c) 2010-2011 by Marcin Kuzminski, Lukasz Balcerzak.

"""

import os

import re

import time

import urllib

import urllib2

import logging

import posixpath

import string

try:

    # Python <=2.7

    from pipes import quote

except ImportError:

    # Python 3.3+

    from shlex import quote

from dulwich.objects import Tag

from dulwich.repo import Repo, NotGitRepository

from dulwich.config import ConfigFile

from kallithea.lib.vcs import subprocessio

from kallithea.lib.vcs.backends.base import BaseRepository, CollectionGenerator

from kallithea.lib.vcs.conf import settings

from kallithea.lib.vcs.exceptions import (

    BranchDoesNotExistError, ChangesetDoesNotExistError, EmptyRepositoryError,

    RepositoryError, TagAlreadyExistError, TagDoesNotExistError

)

from kallithea.lib.vcs.utils import safe_unicode, makedate, date_fromtimestamp

from kallithea.lib.vcs.utils.lazy import LazyProperty

from kallithea.lib.vcs.utils.ordered_dict import OrderedDict

from kallithea.lib.vcs.utils.paths import abspath, get_user_home

from kallithea.lib.vcs.utils.hgcompat import (

    hg_url, httpbasicauthhandler, httpdigestauthhandler

)

from .changeset import GitChangeset

from .inmemory import GitInMemoryChangeset

from .workdir import GitWorkdir

SHA_PATTERN = re.compile(r'^[[0-9a-fA-F]{12}|[0-9a-fA-F]{40}]$')

log = logging.getLogger(__name__)

class GitRepository(BaseRepository):

    """

    Git repository backend.

    """

    DEFAULT_BRANCH_NAME = 'master'

    scm = 'git'

    def __init__(self, repo_path, create=False, src_url=None,

                 update_after_clone=False, bare=False):

        self.path = abspath(repo_path)

        repo = self._get_repo(create, src_url, update_after_clone, bare)

        self.bare = repo.bare

    @property

    def _config_files(self):

        return [

            self.bare and abspath(self.path, 'config')

                      or abspath(self.path, '.git', 'config'),

             abspath(get_user_home(), '.gitconfig'),

         ]

    @property

    def _repo(self):

        return Repo(self.path)

    @property

    def head(self):

        try:

            return self._repo.head()

        except KeyError:

            return None

    @property

    def _empty(self):

        """

        Checks if repository is empty ie. without any changesets

        """

        try:

            self.revisions[0]

        except (KeyError, IndexError):

            return True

        return False

    @LazyProperty

    def revisions(self):

        """

        Returns list of revisions' ids, in ascending order.  Being lazy

        attribute allows external tools to inject shas from cache.

        """

        return self._get_all_revisions()

    @classmethod

    def _run_git_command(cls, cmd, **opts):

        """

        Runs given ``cmd`` as git command and returns tuple

        (stdout, stderr).

        :param cmd: git command to be executed

        :param opts: env options to pass into Subprocess command

        """

        if '_bare' in opts:

            _copts = []

            del opts['_bare']

        else:

            _copts = ['-c', 'core.quotepath=false', ]

        safe_call = False

        if '_safe' in opts:

            #no exc on failure

            del opts['_safe']

            safe_call = True

        _str_cmd = False

        if isinstance(cmd, basestring):

            cmd = [cmd]

            _str_cmd = True

        gitenv = os.environ

        # need to clean fix GIT_DIR !

        if 'GIT_DIR' in gitenv:

            del gitenv['GIT_DIR']

        gitenv['GIT_CONFIG_NOGLOBAL'] = '1'

        _git_path = settings.GIT_EXECUTABLE_PATH

        cmd = [_git_path] + _copts + cmd

        if _str_cmd:

            cmd = ' '.join(cmd)

        try:

            _opts = dict(

                env=gitenv,

                shell=True,

            )

            _opts.update(opts)

            p = subprocessio.SubprocessIOChunker(cmd, **_opts)

        except (EnvironmentError, OSError), err:

            tb_err = ("Couldn't run git command (%s).\n"

                      "Original error was:%s\n" % (cmd, err))

            log.error(tb_err)

            if safe_call:

                return '', err

            else:

                raise RepositoryError(tb_err)

        return ''.join(p.output), ''.join(p.error)

    def run_git_command(self, cmd):

        opts = {}

        if os.path.isdir(self.path):

            opts['cwd'] = self.path

        return self._run_git_command(cmd, **opts)

    @classmethod

    def _check_url(cls, url):

        """

        Function will check given url and try to verify if it's a valid

        link. Sometimes it may happened that git will issue basic

        auth request that can cause whole API to hang when used from python

        or other external calls.

        On failures it'll raise urllib2.HTTPError, exception is also thrown

        when the return code is non 200

        """

        # check first if it's not an local url

        if os.path.isdir(url) or url.startswith('file:'):

            return True

        if '+' in url[:url.find('://')]:

            url = url[url.find('+') + 1:]

        handlers = []

        url_obj = hg_url(url)

        test_uri, authinfo = url_obj.authinfo()

        url_obj.passwd = '*****'

        cleaned_uri = str(url_obj)

        if not test_uri.endswith('info/refs'):

            test_uri = test_uri.rstrip('/') + '/info/refs'

        if authinfo:

            #create a password manager

            passmgr = urllib2.HTTPPasswordMgrWithDefaultRealm()

            passmgr.add_password(*authinfo)

            handlers.extend((httpbasicauthhandler(passmgr),

                             httpdigestauthhandler(passmgr)))

        return heads if reverse else dict((y, x) for x, y in heads.iteritems())

    def get_changeset(self, revision=None):

        """

        Returns ``GitChangeset`` object representing commit from git repository

        at the given revision or head (most recent commit) if None given.

        """

        if isinstance(revision, GitChangeset):

            return revision

        revision = self._get_revision(revision)

        changeset = GitChangeset(repository=self, revision=revision)

        return changeset

    def get_changesets(self, start=None, end=None, start_date=None,

           end_date=None, branch_name=None, reverse=False):

        """

        Returns iterator of ``GitChangeset`` objects from start to end (both

        are inclusive), in ascending date order (unless ``reverse`` is set).

        :param start: changeset ID, as str; first returned changeset

        :param end: changeset ID, as str; last returned changeset

        :param start_date: if specified, changesets with commit date less than

          ``start_date`` would be filtered out from returned set

        :param end_date: if specified, changesets with commit date greater than

          ``end_date`` would be filtered out from returned set

        :param branch_name: if specified, changesets not reachable from given

          branch would be filtered out from returned set

        :param reverse: if ``True``, returned generator would be reversed

          (meaning that returned changesets would have descending date order)

        :raise BranchDoesNotExistError: If given ``branch_name`` does not

            exist.

        :raise ChangesetDoesNotExistError: If changeset for given ``start`` or

          ``end`` could not be found.

        """

        if branch_name and branch_name not in self.branches:

            raise BranchDoesNotExistError("Branch '%s' not found" \

                                          % branch_name)

        # actually we should check now if it's not an empty repo to not spaw

        # subprocess commands

        if self._empty:

            raise EmptyRepositoryError("There are no changesets yet")

        # %H at format means (full) commit hash, initial hashes are retrieved

        # in ascending date order

        cmd_template = 'log --date-order --reverse --pretty=format:"%H"'

        cmd_params = {}

        if start_date:

            cmd_template += ' --since "$since"'

            cmd_params['since'] = start_date.strftime('%m/%d/%y %H:%M:%S')

        if end_date:

            cmd_template += ' --until "$until"'

            cmd_params['until'] = end_date.strftime('%m/%d/%y %H:%M:%S')

        if branch_name:

            cmd_template += ' $branch_name'

            cmd_params['branch_name'] = branch_name

        else:

            rev_filter = settings.GIT_REV_FILTER

            cmd_template += ' %s' % (rev_filter)

        cmd = string.Template(cmd_template).safe_substitute(**cmd_params)

        revs = self.run_git_command(cmd)[0].splitlines()

        start_pos = 0

        end_pos = len(revs)

        if start:

            _start = self._get_revision(start)

            try:

                start_pos = revs.index(_start)

            except ValueError:

                pass

        if end is not None:

            _end = self._get_revision(end)

            try:

                end_pos = revs.index(_end)

            except ValueError:

                pass

        if None not in [start, end] and start_pos > end_pos:

            raise RepositoryError('start cannot be after end')

        if end_pos is not None:

            end_pos += 1

        revs = revs[start_pos:end_pos]

        if reverse:

            revs = reversed(revs)

        return CollectionGenerator(self, revs)

    def get_diff(self, rev1, rev2, path=None, ignore_whitespace=False,

                 context=3):

        """

        Returns (git like) *diff*, as plain text. Shows changes introduced by

        ``rev2`` since ``rev1``.

        :param rev1: Entry point from which diff is shown. Can be

          ``self.EMPTY_CHANGESET`` - in this case, patch showing all

          the changes since empty state of the repository until ``rev2``

        :param rev2: Until which revision changes should be shown.

        :param ignore_whitespace: If set to ``True``, would not show whitespace

          changes. Defaults to ``False``.

        :param context: How many lines before/after changed lines should be

          shown. Defaults to ``3``.

        """

        flags = ['-U%s' % context, '--full-index', '--binary', '-p', '-M', '--abbrev=40']

        if ignore_whitespace:

            flags.append('-w')

        if hasattr(rev1, 'raw_id'):

            rev1 = getattr(rev1, 'raw_id')

        if hasattr(rev2, 'raw_id'):

            rev2 = getattr(rev2, 'raw_id')

        if rev1 == self.EMPTY_CHANGESET:

            rev2 = self.get_changeset(rev2).raw_id

            cmd = ' '.join(['show'] + flags + [rev2])

        else:

            rev1 = self.get_changeset(rev1).raw_id

            rev2 = self.get_changeset(rev2).raw_id

            cmd = ' '.join(['diff'] + flags + [rev1, rev2])

        if path:

            cmd += ' -- "%s"' % path

        stdout, stderr = self.run_git_command(cmd)

        # TODO: don't ignore stderr

        # If we used 'show' command, strip first few lines (until actual diff

        # starts)

        if rev1 == self.EMPTY_CHANGESET:

            parts = stdout.split('\ndiff ', 1)

            if len(parts) > 1:

                stdout = 'diff ' + parts[1]

        return stdout

    @LazyProperty

    def in_memory_changeset(self):

        """

        Returns ``GitInMemoryChangeset`` object for this repository.

        """

        return GitInMemoryChangeset(self)

    def clone(self, url, update_after_clone=True, bare=False):

        """

        Tries to clone changes from external location.

        :param update_after_clone: If set to ``False``, git won't checkout

          working directory

        :param bare: If set to ``True``, repository would be cloned into

          *bare* git repository (no working directory at all).

        """

        url = self._get_url(url)

        cmd = ['clone', '-q']

        if bare:

            cmd.append('--bare')

        elif not update_after_clone:

            cmd.append('--no-checkout')

        cmd += ['--', quote(url), quote(self.path)]

        cmd = ' '.join(cmd)

        # If error occurs run_git_command raises RepositoryError already

        self.run_git_command(cmd)

    def pull(self, url):

        """

        Tries to pull changes from external location.

        """

        url = self._get_url(url)

        cmd = ['pull', "--ff-only", quote(url)]

        cmd = ' '.join(cmd)

        # If error occurs run_git_command raises RepositoryError already

        self.run_git_command(cmd)

    def fetch(self, url):

        """

        Tries to pull changes from external location.

        """

        url = self._get_url(url)

        so, se = self.run_git_command('ls-remote -h %s' % quote(url))

        refs = []

        for line in (x for x in so.splitlines()):

            sha, ref = line.split('\t')

            refs.append(ref)

        refs = ' '.join(('+%s:%s' % (r, r) for r in refs))

        cmd = '''fetch %s -- %s''' % (quote(url), refs)

        self.run_git_command(cmd)

    def _update_server_info(self):

        """

        runs gits update-server-info command in this repo instance

        """

        from dulwich.server import update_server_info

        update_server_info(self._repo)

    @LazyProperty

    def workdir(self):

        """

        Returns ``Workdir`` instance for this repository.

        """

        return GitWorkdir(self)

    def get_config_value(self, section, name, config_file=None):

        """

        Returns configuration value for a given [``section``] and ``name``.

        :param section: Section we want to retrieve value from

        :param name: Name of configuration we want to retrieve

        :param config_file: A path to file which should be used to retrieve

          configuration from (might also be a list of file paths)

        """

        if config_file is None:

            config_file = []

        elif isinstance(config_file, basestring):

            config_file = [config_file]

        def gen_configs():

            for path in config_file + self._config_files:

                try:

                    yield ConfigFile.from_path(path)

                except (IOError, OSError, ValueError):

                    continue

        for config in gen_configs():

            try:

                return config.get(section, name)

            except KeyError:

                continue

        return None

    def get_user_name(self, config_file=None):

        """

        Returns user's name from global configuration file.

        :param config_file: A path to file which should be used to retrieve

          configuration from (might also be a list of file paths)

        """

        return self.get_config_value('user', 'name', config_file)

    def get_user_email(self, config_file=None):

        """

        Returns user's email from global configuration file.

        :param config_file: A path to file which should be used to retrieve

          configuration from (might also be a list of file paths)

        """

        return self.get_config_value('user', 'email', config_file)