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

# This program 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.

#

# This program 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 this program.  If not, see <http://www.gnu.org/licenses/>.

"""

kallithea.lib.hooks

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

Hooks run by Kallithea

This file was forked by the Kallithea project in July 2014.

Original author and date, and relevant copyright and licensing information is below:

:created_on: Aug 6, 2010

:author: marcink

:copyright: (c) 2013 RhodeCode GmbH, and others.

:license: GPLv3, see LICENSE.md for more details.

"""

import os

import time

import mercurial.scmutil

from kallithea.lib import helpers as h

from kallithea.lib.exceptions import UserCreationError

from kallithea.lib.utils import action_logger, make_ui, setup_cache_regions

from kallithea.lib.utils2 import ascii_str, get_hook_environment, safe_str, safe_unicode

from kallithea.lib.vcs.backends.base import EmptyChangeset

from kallithea.model.db import Repository, User

def _get_scm_size(alias, root_path):

    if not alias.startswith('.'):

        alias += '.'

    size_scm, size_root = 0, 0

    for path, dirs, files in os.walk(safe_str(root_path)):

        if path.find(alias) != -1:

            for f in files:

                try:

                    size_scm += os.path.getsize(os.path.join(path, f))

                except OSError:

                    pass

        else:

            for f in files:

                try:

                    size_root += os.path.getsize(os.path.join(path, f))

                except OSError:

                    pass

    size_scm_f = h.format_byte_size(size_scm)

    size_root_f = h.format_byte_size(size_root)

    size_total_f = h.format_byte_size(size_root + size_scm)

    return size_scm_f, size_root_f, size_total_f

def repo_size(ui, repo, hooktype=None, **kwargs):

    size_hg_f, size_root_f, size_total_f = _get_scm_size('.hg', repo.root)

    last_cs = repo[len(repo) - 1]

    msg = ('Repository size .hg: %s Checkout: %s Total: %s\n'

           'Last revision is now r%s:%s\n') % (

        size_hg_f, size_root_f, size_total_f, last_cs.rev(), ascii_str(last_cs.hex())[:12]

    )

    ui.status(msg)

def log_pull_action(ui, repo, **kwargs):

    """Logs user last pull action

    Called as Mercurial hook outgoing.pull_logger or from Kallithea before invoking Git.

    Does *not* use the action from the hook environment but is always 'pull'.

    """

    ex = get_hook_environment()

    user = User.get_by_username(ex.username)

    action = 'pull'

    action_logger(user, action, ex.repository, ex.ip, commit=True)

    # extension hook call

    from kallithea import EXTENSIONS

    callback = getattr(EXTENSIONS, 'PULL_HOOK', None)

    if callable(callback):

        kw = {}

        kw.update(ex)

        callback(**kw)

    return 0

def log_push_action(ui, repo, node, node_last, **kwargs):

    """

    Entry point for Mercurial hook changegroup.push_logger.

    The pushed changesets is given by the revset 'node:node_last'.

    Note: This hook is not only logging, but also the side effect invalidating

    cahes! The function should perhaps be renamed.

    """

    revs = [ascii_str(repo[r].hex()) for r in mercurial.scmutil.revrange(repo, [b'%s:%s' % (node, node_last)])]

    process_pushed_raw_ids(revs)

    return 0

def process_pushed_raw_ids(revs):

    """

    Register that changes have been added to the repo - log the action *and* invalidate caches.

    Called from  Mercurial changegroup.push_logger calling hook log_push_action,

    or from the Git post-receive hook calling handle_git_post_receive ...

    or from scm _handle_push.

    """

    ex = get_hook_environment()

    action = '%s:%s' % (ex.action, ','.join(revs))

    action_logger(ex.username, action, ex.repository, ex.ip, commit=True)

    from kallithea.model.scm import ScmModel

    ScmModel().mark_for_invalidation(ex.repository)

    # extension hook call

    from kallithea import EXTENSIONS

    callback = getattr(EXTENSIONS, 'PUSH_HOOK', None)

    if callable(callback):

        kw = {'pushed_revs': revs}

        kw.update(ex)

        callback(**kw)

def log_create_repository(repository_dict, created_by, **kwargs):

    """

    Post create repository Hook.

    :param repository: dict dump of repository object

    :param created_by: username who created repository

    available keys of repository_dict:

     'repo_type',

     'description',

     'private',

     'created_on',

     'enable_downloads',

     'repo_id',

     'owner_id',

     'enable_statistics',

     'clone_uri',

     'fork_id',

     'group_id',

     'repo_name'

    """

            else:

                return os.stat(he_path).st_mtime

    @LazyProperty

    def description(self):

        return safe_unicode(self._repo.get_description() or b'unknown')

    @LazyProperty

    def contact(self):

        undefined_contact = u'Unknown'

        return undefined_contact

    @property

    def branches(self):

        if not self.revisions:

            return {}

        sortkey = lambda ctx: ctx[0]

        _branches = [(x[0], ascii_str(x[1][0]))

                     for x in self._parsed_refs.iteritems() if x[1][1] == b'H']

        return OrderedDict(sorted(_branches, key=sortkey, reverse=False))

    @LazyProperty

    def closed_branches(self):

        return {}

    @LazyProperty

    def tags(self):

        return self._get_tags()

    def _get_tags(self):

        if not self.revisions:

            return {}

        sortkey = lambda ctx: ctx[0]

        _tags = [(x[0], ascii_str(x[1][0]))

                 for x in self._parsed_refs.iteritems() if x[1][1] == b'T']

        return OrderedDict(sorted(_tags, key=sortkey, reverse=True))

    def tag(self, name, user, revision=None, message=None, date=None,

            **kwargs):

        """

        Creates and returns a tag for the given ``revision``.

        :param name: name for new tag

        :param user: full username, i.e.: "Joe Doe <joe.doe@example.com>"

        :param revision: changeset id for which new tag would be created

        :param message: message of the tag's commit

        :param date: date of tag's commit

        :raises TagAlreadyExistError: if tag with same name already exists

        """

        if name in self.tags:

            raise TagAlreadyExistError("Tag %s already exists" % name)

        changeset = self.get_changeset(revision)

        message = message or "Added tag %s for commit %s" % (name,

            changeset.raw_id)

        self._repo.refs[b"refs/tags/%s" % name] = changeset._commit.id

        self._parsed_refs = self._get_parsed_refs()

        self.tags = self._get_tags()

        return changeset

    def remove_tag(self, name, user, message=None, date=None):

        """

        Removes tag with the given ``name``.

        :param name: name of the tag to be removed

        :param user: full username, i.e.: "Joe Doe <joe.doe@example.com>"

        :param message: message of the tag's removal commit

        :param date: date of tag's removal commit

        :raises TagDoesNotExistError: if tag with given name does not exists

        """

        if name not in self.tags:

            raise TagDoesNotExistError("Tag %s does not exist" % name)

        # self._repo.refs is a DiskRefsContainer, and .path gives the full absolute path of '.git'

        tagpath = os.path.join(self._repo.refs.path, 'refs', 'tags', name)

        try:

            os.remove(tagpath)

            self._parsed_refs = self._get_parsed_refs()

            self.tags = self._get_tags()

        except OSError as e:

            raise RepositoryError(e.strerror)

    @LazyProperty

    def bookmarks(self):

        """

        Gets bookmarks for this repository

        """

        return {}

    @LazyProperty

    def _parsed_refs(self):

        return self._get_parsed_refs()

    def _get_parsed_refs(self):

        _repo = self._repo

        refs = _repo.get_refs()

        keys = [(b'refs/heads/', b'H'),

                (b'refs/remotes/origin/', b'RH'),

                (b'refs/tags/', b'T')]

        _refs = {}

        for ref, sha in refs.iteritems():

            for k, type_ in keys:

                if ref.startswith(k):

                    _key = ref[len(k):]

                    if type_ == b'T':

                        obj = _repo.get_object(sha)

                        if isinstance(obj, Tag):

                            sha = _repo.get_object(sha).object[1]

                    _refs[_key] = [sha, type_]

                    break

        return _refs

    def _heads(self, reverse=False):

        refs = self._repo.get_refs()

        heads = {}

        for key, val in refs.items():

            for ref_key in [b'refs/heads/', b'refs/remotes/origin/']:

                if key.startswith(ref_key):

                    n = key[len(ref_key):]

                    if n not in [b'HEAD']:

                        heads[n] = val

        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, max_revisions=None):

        """

        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 = ['log', '--date-order', '--reverse', '--pretty=format:%H']

        if max_revisions:

            cmd += ['--max-count=%s' % max_revisions]

        if start_date:

            cmd += ['--since', start_date.strftime('%m/%d/%y %H:%M:%S')]

        if end_date:

            cmd += ['--until', end_date.strftime('%m/%d/%y %H:%M:%S')]

        if branch_name:

            cmd.append(branch_name)

        else:

            cmd.append(settings.GIT_REV_FILTER)

        revs = self.run_git_command(cmd).splitlines()

        start_pos = 0

        end_pos = len(revs)

        if start:

            _start = self._get_revision(start)

            try:

                start_pos = revs.index(_start)

            except ValueError:

                pass

        self._nodes_dict = dict((node.path, node) for node in nodes)

        return sorted(nodes)

    @LazyProperty

    def files(self):

        return sorted((node for node in self.nodes if node.is_file()))

    @LazyProperty

    def dirs(self):

        return sorted((node for node in self.nodes if node.is_dir()))

    def __iter__(self):

        for node in self.nodes:

            yield node

    def get_node(self, path):

        """

        Returns node from within this particular ``DirNode``, so it is now

        allowed to fetch, i.e. node located at 'docs/api/index.rst' from node

        'docs'. In order to access deeper nodes one must fetch nodes between

        them first - this would work::

           docs = root.get_node('docs')

           docs.get_node('api').get_node('index.rst')

        :param: path - relative to the current node

        .. note::

           To access lazily (as in example above) node have to be initialized

           with related changeset object - without it node is out of

           context and may know nothing about anything else than nearest

           (located at same level) nodes.

        """

        try:

            path = path.rstrip('/')

            if path == '':

                raise NodeError("Cannot retrieve node without path")

            self.nodes  # access nodes first in order to set _nodes_dict

            paths = path.split('/')

            if len(paths) == 1:

                if not self.is_root():

                    path = '/'.join((self.path, paths[0]))

                else:

                    path = paths[0]

                return self._nodes_dict[path]

            elif len(paths) > 1:

                if self.changeset is None:

                    raise NodeError("Cannot access deeper "

                                    "nodes without changeset")

                else:

                    path1, path2 = paths[0], '/'.join(paths[1:])

                    return self.get_node(path1).get_node(path2)

            else:

                raise KeyError

        except KeyError:

            raise NodeError("Node does not exist at %s" % path)

    @LazyProperty

    def state(self):

        raise NodeError("Cannot access state of DirNode")

    @LazyProperty

    def size(self):

        size = 0

        for root, dirs, files in self.changeset.walk(self.path):

            for f in files:

                size += f.size

        return size

    def __repr__(self):

        return '<%s %r @ %s>' % (self.__class__.__name__, self.path,

                                 getattr(self.changeset, 'short_id', ''))

class RootNode(DirNode):

    """

    DirNode being the root node of the repository.

    """

    def __init__(self, nodes=(), changeset=None):

        super(RootNode, self).__init__(path='', nodes=nodes,

            changeset=changeset)

    def __repr__(self):

        return '<%s>' % self.__class__.__name__

class SubModuleNode(Node):

    """

    represents a SubModule of Git or SubRepo of Mercurial

    """

    is_binary = False

    size = 0

    def __init__(self, name, url, changeset=None, alias=None):

        self.path = name

        self.kind = NodeKind.SUBMODULE

        self.alias = alias

        # we have to use emptyChangeset here since this can point to svn/git/hg

        # submodules we cannot get from repository

        self.changeset = EmptyChangeset(str(changeset), alias=alias)

        self.url = url

    def __repr__(self):

        return '<%s %r @ %s>' % (self.__class__.__name__, self.path,

                                 getattr(self.changeset, 'short_id', ''))

    @LazyProperty

    def name(self):

        """

        Returns name of the node so if its path

        then only last part is returned.

        """

        org = safe_unicode(self.path.rstrip('/').split('/')[-1])

        return u'%s @ %s' % (org, self.changeset.short_id)