Changeset - e63bcce18fef
[Not reviewed]
docs/conf.py
Show inline comments
 
# -*- coding: utf-8 -*-
 
#
 
# Kallithea documentation build configuration file, created by
 
# sphinx-quickstart on Sun Oct 10 16:46:37 2010.
 
#
 
# This file is execfile()d with the current directory set to its containing dir.
 
#
 
# Note that not all possible configuration values are present in this
 
# autogenerated file.
 
#
 
# All configuration values have a default; values that are commented out
 
# serve to show the default.
 

	
 
import os
 
import sys
 

	
 
from kallithea import __version__
 

	
 

	
 
# If extensions (or modules to document with autodoc) are in another directory,
 
# add these directories to sys.path here. If the directory is relative to the
 
# documentation root, use os.path.abspath to make it absolute, like shown here.
 
sys.path.insert(0, os.path.abspath('..'))
 

	
 
# -- General configuration -----------------------------------------------------
 

	
 
# If your documentation needs a minimal Sphinx version, state it here.
 
#needs_sphinx = '1.0'
 

	
 
# Add any Sphinx extension module names here, as strings. They can be extensions
 
# coming with Sphinx (named 'sphinx.ext.*') or your custom ones.
 
extensions = ['sphinx.ext.autodoc', 'sphinx.ext.doctest',
 
              'sphinx.ext.intersphinx', 'sphinx.ext.todo',
 
              'sphinx.ext.viewcode']
 

	
 
# Add any paths that contain templates here, relative to this directory.
 
templates_path = ['_templates']
 

	
 
# The suffix of source filenames.
 
source_suffix = '.rst'
 

	
 
# The encoding of source files.
 
#source_encoding = 'utf-8-sig'
 

	
 
# The master toctree document.
 
master_doc = 'index'
 

	
 
# General information about the project.
 
project = u'Kallithea'
 
copyright = u'2010-2020 by various authors, licensed as GPLv3.'
 
project = 'Kallithea'
 
copyright = '2010-2020 by various authors, licensed as GPLv3.'
 

	
 
# The version info for the project you're documenting, acts as replacement for
 
# |version| and |release|, also used in various other places throughout the
 
# built documents.
 
#
 
# The short X.Y version.
 
root = os.path.dirname(os.path.dirname(__file__))
 
sys.path.append(root)
 
version = __version__
 
# The full version, including alpha/beta/rc tags.
 
release = __version__
 

	
 
# The language for content autogenerated by Sphinx. Refer to documentation
 
# for a list of supported languages.
 
#language = None
 

	
 
# There are two options for replacing |today|: either, you set today to some
 
# non-false value, then it is used:
 
#today = ''
 
# Else, today_fmt is used as the format for a strftime call.
 
#today_fmt = '%B %d, %Y'
 

	
 
# List of patterns, relative to source directory, that match files and
 
# directories to ignore when looking for source files.
 
exclude_patterns = ['_build']
 

	
 
# The reST default role (used for this markup: `text`) to use for all documents.
 
#default_role = None
 

	
 
# If true, '()' will be appended to :func: etc. cross-reference text.
 
#add_function_parentheses = True
 

	
 
# If true, the current module name will be prepended to all description
 
# unit titles (such as .. function::).
 
#add_module_names = True
 

	
 
# If true, sectionauthor and moduleauthor directives will be shown in the
 
# output. They are ignored by default.
 
#show_authors = False
 

	
 
# The name of the Pygments (syntax highlighting) style to use.
 
pygments_style = 'sphinx'
 
highlight_language = 'none'
 

	
 
# A list of ignored prefixes for module index sorting.
 
#modindex_common_prefix = []
 

	
 

	
 
# -- Options for HTML output ---------------------------------------------------
 

	
 
# The theme to use for HTML and HTML Help pages.  See the documentation for
 
# a list of builtin themes.
 
html_theme = 'nature'
 

	
 
# Theme options are theme-specific and customize the look and feel of a theme
 
# further.  For a list of options available for each theme, see the
 
# documentation.
 
#html_theme_options = {}
 

	
 
# Add any paths that contain custom themes here, relative to this directory.
 
html_theme_path = ['theme']
 

	
 
# The name for this set of Sphinx documents.  If None, it defaults to
 
# "<project> v<release> documentation".
 
#html_title = None
 

	
 
# A shorter title for the navigation bar.  Default is the same as html_title.
 
#html_short_title = None
 

	
 
# The name of an image file (relative to this directory) to place at the top
 
# of the sidebar.
 
#html_logo = None
 

	
 
# The name of an image file (within the static path) to use as favicon of the
 
# docs.  This file should be a Windows icon file (.ico) being 16x16 or 32x32
 
# pixels large.
 
#html_favicon = None
 

	
 
# Add any paths that contain custom static files (such as style sheets) here,
 
# relative to this directory. They are copied after the builtin static files,
 
# so a file named "default.css" will overwrite the builtin "default.css".
 
#html_static_path = ['_static']
 

	
 
# If not '', a 'Last updated on:' timestamp is inserted at every page bottom,
 
# using the given strftime format.
 
#html_last_updated_fmt = '%b %d, %Y'
 

	
 
# If true, SmartyPants will be used to convert quotes and dashes to
 
# typographically correct entities.
 
#html_use_smartypants = True
 

	
 
# Custom sidebar templates, maps document names to template names.
 
#html_sidebars = {}
 

	
 
# Additional templates that should be rendered to pages, maps page names to
 
# template names.
 
#html_additional_pages = {}
 

	
 
# If false, no module index is generated.
 
#html_domain_indices = True
 

	
 
# If false, no index is generated.
 
#html_use_index = True
 

	
 
# If true, the index is split into individual pages for each letter.
 
#html_split_index = False
 

	
 
# If true, links to the reST sources are added to the pages.
 
#html_show_sourcelink = True
 

	
 
# If true, "Created using Sphinx" is shown in the HTML footer. Default is True.
 
#html_show_sphinx = True
 

	
 
# If true, "(C) Copyright ..." is shown in the HTML footer. Default is True.
 
#html_show_copyright = True
 

	
 
# If true, an OpenSearch description file will be output, and all pages will
 
# contain a <link> tag referring to it.  The value of this option must be the
 
# base URL from which the finished HTML is served.
 
#html_use_opensearch = ''
 

	
 
# This is the file name suffix for HTML files (e.g. ".xhtml").
 
#html_file_suffix = None
 

	
 
# Output file base name for HTML help builder.
 
htmlhelp_basename = 'Kallithea-docs'
 

	
 

	
 
# -- Options for LaTeX output --------------------------------------------------
 

	
 
# The paper size ('letter' or 'a4').
 
#latex_paper_size = 'letter'
 

	
 
# The font size ('10pt', '11pt' or '12pt').
 
#latex_font_size = '10pt'
 

	
 
# Grouping the document tree into LaTeX files. List of tuples
 
# (source start file, target name, title, author, documentclass [howto/manual]).
 
latex_documents = [
 
  ('index', 'Kallithea.tex', u'Kallithea Documentation',
 
   u'Kallithea Developers', 'manual'),
 
  ('index', 'Kallithea.tex', 'Kallithea Documentation',
 
   'Kallithea Developers', 'manual'),
 
]
 

	
 
# The name of an image file (relative to this directory) to place at the top of
 
# the title page.
 
#latex_logo = None
 

	
 
# For "manual" documents, if this is true, then toplevel headings are parts,
 
# not chapters.
 
#latex_use_parts = False
 

	
 
# If true, show page references after internal links.
 
#latex_show_pagerefs = False
 

	
 
# If true, show URL addresses after external links.
 
#latex_show_urls = False
 

	
 
# Additional stuff for the LaTeX preamble.
 
#latex_preamble = ''
 

	
 
# Documents to append as an appendix to all manuals.
 
#latex_appendices = []
 

	
 
# If false, no module index is generated.
 
#latex_domain_indices = True
 

	
 

	
 
# -- Options for manual page output --------------------------------------------
 

	
 
# One entry per manual page. List of tuples
 
# (source start file, name, description, authors, manual section).
 
man_pages = [
 
    ('index', 'kallithea', u'Kallithea Documentation',
 
     [u'Kallithea Developers'], 1)
 
    ('index', 'kallithea', 'Kallithea Documentation',
 
     ['Kallithea Developers'], 1)
 
]
 

	
 

	
 
# Example configuration for intersphinx: refer to the Python standard library.
 
intersphinx_mapping = {'http://docs.python.org/': None}
kallithea/controllers/api/api.py
Show inline comments
 
@@ -52,1022 +52,1022 @@ from kallithea.model.user import UserMod
 
from kallithea.model.user_group import UserGroupModel
 

	
 

	
 
log = logging.getLogger(__name__)
 

	
 

	
 
def store_update(updates, attr, name):
 
    """
 
    Stores param in updates dict if it's not instance of Optional
 
    allows easy updates of passed in params
 
    """
 
    if not isinstance(attr, Optional):
 
        updates[name] = attr
 

	
 

	
 
def get_user_or_error(userid):
 
    """
 
    Get user by id or name or return JsonRPCError if not found
 

	
 
    :param userid:
 
    """
 
    user = UserModel().get_user(userid)
 
    if user is None:
 
        raise JSONRPCError("user `%s` does not exist" % (userid,))
 
    return user
 

	
 

	
 
def get_repo_or_error(repoid):
 
    """
 
    Get repo by id or name or return JsonRPCError if not found
 

	
 
    :param repoid:
 
    """
 
    repo = RepoModel().get_repo(repoid)
 
    if repo is None:
 
        raise JSONRPCError('repository `%s` does not exist' % (repoid,))
 
    return repo
 

	
 

	
 
def get_repo_group_or_error(repogroupid):
 
    """
 
    Get repo group by id or name or return JsonRPCError if not found
 

	
 
    :param repogroupid:
 
    """
 
    repo_group = RepoGroup.guess_instance(repogroupid)
 
    if repo_group is None:
 
        raise JSONRPCError(
 
            'repository group `%s` does not exist' % (repogroupid,))
 
    return repo_group
 

	
 

	
 
def get_user_group_or_error(usergroupid):
 
    """
 
    Get user group by id or name or return JsonRPCError if not found
 

	
 
    :param usergroupid:
 
    """
 
    user_group = UserGroupModel().get_group(usergroupid)
 
    if user_group is None:
 
        raise JSONRPCError('user group `%s` does not exist' % (usergroupid,))
 
    return user_group
 

	
 

	
 
def get_perm_or_error(permid, prefix=None):
 
    """
 
    Get permission by id or name or return JsonRPCError if not found
 

	
 
    :param permid:
 
    """
 
    perm = Permission.get_by_key(permid)
 
    if perm is None:
 
        raise JSONRPCError('permission `%s` does not exist' % (permid,))
 
    if prefix:
 
        if not perm.permission_name.startswith(prefix):
 
            raise JSONRPCError('permission `%s` is invalid, '
 
                               'should start with %s' % (permid, prefix))
 
    return perm
 

	
 

	
 
def get_gist_or_error(gistid):
 
    """
 
    Get gist by id or gist_access_id or return JsonRPCError if not found
 

	
 
    :param gistid:
 
    """
 
    gist = GistModel().get_gist(gistid)
 
    if gist is None:
 
        raise JSONRPCError('gist `%s` does not exist' % (gistid,))
 
    return gist
 

	
 

	
 
class ApiController(JSONRPCController):
 
    """
 
    API Controller
 

	
 
    The authenticated user can be found as request.authuser.
 

	
 
    Example function::
 

	
 
        def func(arg1, arg2,...):
 
            pass
 

	
 
    Each function should also **raise** JSONRPCError for any
 
    errors that happens.
 
    """
 

	
 
    @HasPermissionAnyDecorator('hg.admin')
 
    def test(self, args):
 
        return args
 

	
 
    @HasPermissionAnyDecorator('hg.admin')
 
    def pull(self, repoid, clone_uri=Optional(None)):
 
        """
 
        Triggers a pull from remote location on given repo. Can be used to
 
        automatically keep remote repos up to date. This command can be executed
 
        only using api_key belonging to user with admin rights
 

	
 
        :param repoid: repository name or repository id
 
        :type repoid: str or int
 
        :param clone_uri: repository URI to pull from (optional)
 
        :type clone_uri: str
 

	
 
        OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : {
 
            "msg": "Pulled from `<repository name>`"
 
            "repository": "<repository name>"
 
          }
 
          error :  null
 

	
 
        ERROR OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : null
 
          error :  {
 
            "Unable to pull changes from `<reponame>`"
 
          }
 

	
 
        """
 

	
 
        repo = get_repo_or_error(repoid)
 

	
 
        try:
 
            ScmModel().pull_changes(repo.repo_name,
 
                                    request.authuser.username,
 
                                    request.ip_addr,
 
                                    clone_uri=Optional.extract(clone_uri))
 
            return dict(
 
                msg='Pulled from `%s`' % repo.repo_name,
 
                repository=repo.repo_name
 
            )
 
        except Exception:
 
            log.error(traceback.format_exc())
 
            raise JSONRPCError(
 
                'Unable to pull changes from `%s`' % repo.repo_name
 
            )
 

	
 
    @HasPermissionAnyDecorator('hg.admin')
 
    def rescan_repos(self, remove_obsolete=Optional(False)):
 
        """
 
        Triggers rescan repositories action. If remove_obsolete is set
 
        than also delete repos that are in database but not in the filesystem.
 
        aka "clean zombies". This command can be executed only using api_key
 
        belonging to user with admin rights.
 

	
 
        :param remove_obsolete: deletes repositories from
 
            database that are not found on the filesystem
 
        :type remove_obsolete: Optional(bool)
 

	
 
        OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : {
 
            'added': [<added repository name>,...]
 
            'removed': [<removed repository name>,...]
 
          }
 
          error :  null
 

	
 
        ERROR OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : null
 
          error :  {
 
            'Error occurred during rescan repositories action'
 
          }
 

	
 
        """
 

	
 
        try:
 
            rm_obsolete = Optional.extract(remove_obsolete)
 
            added, removed = repo2db_mapper(ScmModel().repo_scan(),
 
                                            remove_obsolete=rm_obsolete)
 
            return {'added': added, 'removed': removed}
 
        except Exception:
 
            log.error(traceback.format_exc())
 
            raise JSONRPCError(
 
                'Error occurred during rescan repositories action'
 
            )
 

	
 
    def invalidate_cache(self, repoid):
 
        """
 
        Invalidate cache for repository.
 
        This command can be executed only using api_key belonging to user with admin
 
        rights or regular user that have write or admin or write access to repository.
 

	
 
        :param repoid: repository name or repository id
 
        :type repoid: str or int
 

	
 
        OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : {
 
            'msg': Cache for repository `<repository name>` was invalidated,
 
            'repository': <repository name>
 
          }
 
          error :  null
 

	
 
        ERROR OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : null
 
          error :  {
 
            'Error occurred during cache invalidation action'
 
          }
 

	
 
        """
 
        repo = get_repo_or_error(repoid)
 
        if not HasPermissionAny('hg.admin')():
 
            if not HasRepoPermissionLevel('write')(repo.repo_name):
 
                raise JSONRPCError('repository `%s` does not exist' % (repoid,))
 

	
 
        try:
 
            ScmModel().mark_for_invalidation(repo.repo_name)
 
            return dict(
 
                msg='Cache for repository `%s` was invalidated' % (repoid,),
 
                repository=repo.repo_name
 
            )
 
        except Exception:
 
            log.error(traceback.format_exc())
 
            raise JSONRPCError(
 
                'Error occurred during cache invalidation action'
 
            )
 

	
 
    @HasPermissionAnyDecorator('hg.admin')
 
    def get_ip(self, userid=Optional(OAttr('apiuser'))):
 
        """
 
        Shows IP address as seen from Kallithea server, together with all
 
        defined IP addresses for given user. If userid is not passed data is
 
        returned for user who's calling this function.
 
        This command can be executed only using api_key belonging to user with
 
        admin rights.
 

	
 
        :param userid: username to show ips for
 
        :type userid: Optional(str or int)
 

	
 
        OUTPUT::
 

	
 
            id : <id_given_in_input>
 
            result : {
 
                         "server_ip_addr": "<ip_from_clien>",
 
                         "user_ips": [
 
                                        {
 
                                           "ip_addr": "<ip_with_mask>",
 
                                           "ip_range": ["<start_ip>", "<end_ip>"],
 
                                        },
 
                                        ...
 
                                     ]
 
            }
 

	
 
        """
 
        if isinstance(userid, Optional):
 
            userid = request.authuser.user_id
 
        user = get_user_or_error(userid)
 
        ips = UserIpMap.query().filter(UserIpMap.user == user).all()
 
        return dict(
 
            server_ip_addr=request.ip_addr,
 
            user_ips=ips
 
        )
 

	
 
    # alias for old
 
    show_ip = get_ip
 

	
 
    @HasPermissionAnyDecorator('hg.admin')
 
    def get_server_info(self):
 
        """
 
        return server info, including Kallithea version and installed packages
 

	
 

	
 
        OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : {
 
            'modules': [<module name>,...]
 
            'py_version': <python version>,
 
            'platform': <platform type>,
 
            'kallithea_version': <kallithea version>
 
          }
 
          error :  null
 
        """
 
        return Setting.get_server_info()
 

	
 
    def get_user(self, userid=Optional(OAttr('apiuser'))):
 
        """
 
        Gets a user by username or user_id, Returns empty result if user is
 
        not found. If userid param is skipped it is set to id of user who is
 
        calling this method. This command can be executed only using api_key
 
        belonging to user with admin rights, or regular users that cannot
 
        specify different userid than theirs
 

	
 
        :param userid: user to get data for
 
        :type userid: Optional(str or int)
 

	
 
        OUTPUT::
 

	
 
            id : <id_given_in_input>
 
            result: None if user does not exist or
 
                    {
 
                        "user_id" :     "<user_id>",
 
                        "api_key" :     "<api_key>",
 
                        "api_keys":     "[<list of all API keys including additional ones>]"
 
                        "username" :    "<username>",
 
                        "firstname":    "<firstname>",
 
                        "lastname" :    "<lastname>",
 
                        "email" :       "<email>",
 
                        "emails":       "[<list of all emails including additional ones>]",
 
                        "ip_addresses": "[<ip_address_for_user>,...]",
 
                        "active" :      "<bool: user active>",
 
                        "admin" :       "<bool: user is admin>",
 
                        "extern_name" : "<extern_name>",
 
                        "extern_type" : "<extern type>
 
                        "last_login":   "<last_login>",
 
                        "permissions": {
 
                            "global": ["hg.create.repository",
 
                                       "repository.read",
 
                                       "hg.register.manual_activate"],
 
                            "repositories": {"repo1": "repository.none"},
 
                            "repositories_groups": {"Group1": "group.read"}
 
                         },
 
                    }
 

	
 
            error:  null
 

	
 
        """
 
        if not HasPermissionAny('hg.admin')():
 
            # make sure normal user does not pass someone else userid,
 
            # he is not allowed to do that
 
            if not isinstance(userid, Optional) and userid != request.authuser.user_id:
 
                raise JSONRPCError(
 
                    'userid is not the same as your user'
 
                )
 

	
 
        if isinstance(userid, Optional):
 
            userid = request.authuser.user_id
 

	
 
        user = get_user_or_error(userid)
 
        data = user.get_api_data()
 
        data['permissions'] = AuthUser(user_id=user.user_id).permissions
 
        return data
 

	
 
    @HasPermissionAnyDecorator('hg.admin')
 
    def get_users(self):
 
        """
 
        Lists all existing users. This command can be executed only using api_key
 
        belonging to user with admin rights.
 

	
 

	
 
        OUTPUT::
 

	
 
            id : <id_given_in_input>
 
            result: [<user_object>, ...]
 
            error:  null
 
        """
 

	
 
        return [
 
            user.get_api_data()
 
            for user in User.query()
 
                .order_by(User.username)
 
                .filter_by(is_default_user=False)
 
        ]
 

	
 
    @HasPermissionAnyDecorator('hg.admin')
 
    def create_user(self, username, email, password=Optional(''),
 
                    firstname=Optional(u''), lastname=Optional(u''),
 
                    firstname=Optional(''), lastname=Optional(''),
 
                    active=Optional(True), admin=Optional(False),
 
                    extern_type=Optional(User.DEFAULT_AUTH_TYPE),
 
                    extern_name=Optional('')):
 
        """
 
        Creates new user. Returns new user object. This command can
 
        be executed only using api_key belonging to user with admin rights.
 

	
 
        :param username: new username
 
        :type username: str or int
 
        :param email: email
 
        :type email: str
 
        :param password: password
 
        :type password: Optional(str)
 
        :param firstname: firstname
 
        :type firstname: Optional(str)
 
        :param lastname: lastname
 
        :type lastname: Optional(str)
 
        :param active: active
 
        :type active: Optional(bool)
 
        :param admin: admin
 
        :type admin: Optional(bool)
 
        :param extern_name: name of extern
 
        :type extern_name: Optional(str)
 
        :param extern_type: extern_type
 
        :type extern_type: Optional(str)
 

	
 

	
 
        OUTPUT::
 

	
 
            id : <id_given_in_input>
 
            result: {
 
                      "msg" : "created new user `<username>`",
 
                      "user": <user_obj>
 
                    }
 
            error:  null
 

	
 
        ERROR OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : null
 
          error :  {
 
            "user `<username>` already exist"
 
            or
 
            "email `<email>` already exist"
 
            or
 
            "failed to create user `<username>`"
 
          }
 

	
 
        """
 

	
 
        if User.get_by_username(username):
 
            raise JSONRPCError("user `%s` already exist" % (username,))
 

	
 
        if User.get_by_email(email):
 
            raise JSONRPCError("email `%s` already exist" % (email,))
 

	
 
        try:
 
            user = UserModel().create_or_update(
 
                username=Optional.extract(username),
 
                password=Optional.extract(password),
 
                email=Optional.extract(email),
 
                firstname=Optional.extract(firstname),
 
                lastname=Optional.extract(lastname),
 
                active=Optional.extract(active),
 
                admin=Optional.extract(admin),
 
                extern_type=Optional.extract(extern_type),
 
                extern_name=Optional.extract(extern_name)
 
            )
 
            Session().commit()
 
            return dict(
 
                msg='created new user `%s`' % username,
 
                user=user.get_api_data()
 
            )
 
        except Exception:
 
            log.error(traceback.format_exc())
 
            raise JSONRPCError('failed to create user `%s`' % (username,))
 

	
 
    @HasPermissionAnyDecorator('hg.admin')
 
    def update_user(self, userid, username=Optional(None),
 
                    email=Optional(None), password=Optional(None),
 
                    firstname=Optional(None), lastname=Optional(None),
 
                    active=Optional(None), admin=Optional(None),
 
                    extern_type=Optional(None), extern_name=Optional(None)):
 
        """
 
        updates given user if such user exists. This command can
 
        be executed only using api_key belonging to user with admin rights.
 

	
 
        :param userid: userid to update
 
        :type userid: str or int
 
        :param username: new username
 
        :type username: str or int
 
        :param email: email
 
        :type email: str
 
        :param password: password
 
        :type password: Optional(str)
 
        :param firstname: firstname
 
        :type firstname: Optional(str)
 
        :param lastname: lastname
 
        :type lastname: Optional(str)
 
        :param active: active
 
        :type active: Optional(bool)
 
        :param admin: admin
 
        :type admin: Optional(bool)
 
        :param extern_name:
 
        :type extern_name: Optional(str)
 
        :param extern_type:
 
        :type extern_type: Optional(str)
 

	
 

	
 
        OUTPUT::
 

	
 
            id : <id_given_in_input>
 
            result: {
 
                      "msg" : "updated user ID:<userid> <username>",
 
                      "user": <user_object>,
 
                    }
 
            error:  null
 

	
 
        ERROR OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : null
 
          error :  {
 
            "failed to update user `<username>`"
 
          }
 

	
 
        """
 

	
 
        user = get_user_or_error(userid)
 

	
 
        # only non optional arguments will be stored in updates
 
        updates = {}
 

	
 
        try:
 

	
 
            store_update(updates, username, 'username')
 
            store_update(updates, password, 'password')
 
            store_update(updates, email, 'email')
 
            store_update(updates, firstname, 'name')
 
            store_update(updates, lastname, 'lastname')
 
            store_update(updates, active, 'active')
 
            store_update(updates, admin, 'admin')
 
            store_update(updates, extern_name, 'extern_name')
 
            store_update(updates, extern_type, 'extern_type')
 

	
 
            user = UserModel().update_user(user, **updates)
 
            Session().commit()
 
            return dict(
 
                msg='updated user ID:%s %s' % (user.user_id, user.username),
 
                user=user.get_api_data()
 
            )
 
        except DefaultUserException:
 
            log.error(traceback.format_exc())
 
            raise JSONRPCError('editing default user is forbidden')
 
        except Exception:
 
            log.error(traceback.format_exc())
 
            raise JSONRPCError('failed to update user `%s`' % (userid,))
 

	
 
    @HasPermissionAnyDecorator('hg.admin')
 
    def delete_user(self, userid):
 
        """
 
        deletes given user if such user exists. This command can
 
        be executed only using api_key belonging to user with admin rights.
 

	
 
        :param userid: user to delete
 
        :type userid: str or int
 

	
 
        OUTPUT::
 

	
 
            id : <id_given_in_input>
 
            result: {
 
                      "msg" : "deleted user ID:<userid> <username>",
 
                      "user": null
 
                    }
 
            error:  null
 

	
 
        ERROR OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : null
 
          error :  {
 
            "failed to delete user ID:<userid> <username>"
 
          }
 

	
 
        """
 
        user = get_user_or_error(userid)
 

	
 
        try:
 
            UserModel().delete(userid)
 
            Session().commit()
 
            return dict(
 
                msg='deleted user ID:%s %s' % (user.user_id, user.username),
 
                user=None
 
            )
 
        except Exception:
 

	
 
            log.error(traceback.format_exc())
 
            raise JSONRPCError('failed to delete user ID:%s %s'
 
                               % (user.user_id, user.username))
 

	
 
    # permission check inside
 
    def get_user_group(self, usergroupid):
 
        """
 
        Gets an existing user group. This command can be executed only using api_key
 
        belonging to user with admin rights or user who has at least
 
        read access to user group.
 

	
 
        :param usergroupid: id of user_group to edit
 
        :type usergroupid: str or int
 

	
 
        OUTPUT::
 

	
 
            id : <id_given_in_input>
 
            result : None if group not exist
 
                     {
 
                       "users_group_id" : "<id>",
 
                       "group_name" :     "<groupname>",
 
                       "active":          "<bool>",
 
                       "members" :  [<user_obj>,...]
 
                     }
 
            error : null
 

	
 
        """
 
        user_group = get_user_group_or_error(usergroupid)
 
        if not HasPermissionAny('hg.admin')():
 
            if not HasUserGroupPermissionLevel('read')(user_group.users_group_name):
 
                raise JSONRPCError('user group `%s` does not exist' % (usergroupid,))
 

	
 
        data = user_group.get_api_data()
 
        return data
 

	
 
    # permission check inside
 
    def get_user_groups(self):
 
        """
 
        Lists all existing user groups. This command can be executed only using
 
        api_key belonging to user with admin rights or user who has at least
 
        read access to user group.
 

	
 

	
 
        OUTPUT::
 

	
 
            id : <id_given_in_input>
 
            result : [<user_group_obj>,...]
 
            error : null
 
        """
 

	
 
        return [
 
            user_group.get_api_data()
 
            for user_group in UserGroupList(UserGroup.query().all(), perm_level='read')
 
        ]
 

	
 
    @HasPermissionAnyDecorator('hg.admin', 'hg.usergroup.create.true')
 
    def create_user_group(self, group_name, description=Optional(u''),
 
    def create_user_group(self, group_name, description=Optional(''),
 
                          owner=Optional(OAttr('apiuser')), active=Optional(True)):
 
        """
 
        Creates new user group. This command can be executed only using api_key
 
        belonging to user with admin rights or an user who has create user group
 
        permission
 

	
 
        :param group_name: name of new user group
 
        :type group_name: str
 
        :param description: group description
 
        :type description: str
 
        :param owner: owner of group. If not passed apiuser is the owner
 
        :type owner: Optional(str or int)
 
        :param active: group is active
 
        :type active: Optional(bool)
 

	
 
        OUTPUT::
 

	
 
            id : <id_given_in_input>
 
            result: {
 
                      "msg": "created new user group `<groupname>`",
 
                      "user_group": <user_group_object>
 
                    }
 
            error:  null
 

	
 
        ERROR OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : null
 
          error :  {
 
            "user group `<group name>` already exist"
 
            or
 
            "failed to create group `<group name>`"
 
          }
 

	
 
        """
 

	
 
        if UserGroupModel().get_by_name(group_name):
 
            raise JSONRPCError("user group `%s` already exist" % (group_name,))
 

	
 
        try:
 
            if isinstance(owner, Optional):
 
                owner = request.authuser.user_id
 

	
 
            owner = get_user_or_error(owner)
 
            active = Optional.extract(active)
 
            description = Optional.extract(description)
 
            ug = UserGroupModel().create(name=group_name, description=description,
 
                                         owner=owner, active=active)
 
            Session().commit()
 
            return dict(
 
                msg='created new user group `%s`' % group_name,
 
                user_group=ug.get_api_data()
 
            )
 
        except Exception:
 
            log.error(traceback.format_exc())
 
            raise JSONRPCError('failed to create group `%s`' % (group_name,))
 

	
 
    # permission check inside
 
    def update_user_group(self, usergroupid, group_name=Optional(''),
 
                          description=Optional(''), owner=Optional(None),
 
                          active=Optional(True)):
 
        """
 
        Updates given usergroup.  This command can be executed only using api_key
 
        belonging to user with admin rights or an admin of given user group
 

	
 
        :param usergroupid: id of user group to update
 
        :type usergroupid: str or int
 
        :param group_name: name of new user group
 
        :type group_name: str
 
        :param description: group description
 
        :type description: str
 
        :param owner: owner of group.
 
        :type owner: Optional(str or int)
 
        :param active: group is active
 
        :type active: Optional(bool)
 

	
 
        OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : {
 
            "msg": 'updated user group ID:<user group id> <user group name>',
 
            "user_group": <user_group_object>
 
          }
 
          error :  null
 

	
 
        ERROR OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : null
 
          error :  {
 
            "failed to update user group `<user group name>`"
 
          }
 

	
 
        """
 
        user_group = get_user_group_or_error(usergroupid)
 
        if not HasPermissionAny('hg.admin')():
 
            if not HasUserGroupPermissionLevel('admin')(user_group.users_group_name):
 
                raise JSONRPCError('user group `%s` does not exist' % (usergroupid,))
 

	
 
        if not isinstance(owner, Optional):
 
            owner = get_user_or_error(owner)
 

	
 
        updates = {}
 
        store_update(updates, group_name, 'users_group_name')
 
        store_update(updates, description, 'user_group_description')
 
        store_update(updates, owner, 'owner')
 
        store_update(updates, active, 'users_group_active')
 
        try:
 
            UserGroupModel().update(user_group, updates)
 
            Session().commit()
 
            return dict(
 
                msg='updated user group ID:%s %s' % (user_group.users_group_id,
 
                                                     user_group.users_group_name),
 
                user_group=user_group.get_api_data()
 
            )
 
        except Exception:
 
            log.error(traceback.format_exc())
 
            raise JSONRPCError('failed to update user group `%s`' % (usergroupid,))
 

	
 
    # permission check inside
 
    def delete_user_group(self, usergroupid):
 
        """
 
        Delete given user group by user group id or name.
 
        This command can be executed only using api_key
 
        belonging to user with admin rights or an admin of given user group
 

	
 
        :param usergroupid:
 
        :type usergroupid: int
 

	
 
        OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : {
 
            "msg": "deleted user group ID:<user_group_id> <user_group_name>"
 
          }
 
          error :  null
 

	
 
        ERROR OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : null
 
          error :  {
 
            "failed to delete user group ID:<user_group_id> <user_group_name>"
 
            or
 
            "RepoGroup assigned to <repo_groups_list>"
 
          }
 

	
 
        """
 
        user_group = get_user_group_or_error(usergroupid)
 
        if not HasPermissionAny('hg.admin')():
 
            if not HasUserGroupPermissionLevel('admin')(user_group.users_group_name):
 
                raise JSONRPCError('user group `%s` does not exist' % (usergroupid,))
 

	
 
        try:
 
            UserGroupModel().delete(user_group)
 
            Session().commit()
 
            return dict(
 
                msg='deleted user group ID:%s %s' %
 
                    (user_group.users_group_id, user_group.users_group_name),
 
                user_group=None
 
            )
 
        except UserGroupsAssignedException as e:
 
            log.error(traceback.format_exc())
 
            raise JSONRPCError(str(e))
 
        except Exception:
 
            log.error(traceback.format_exc())
 
            raise JSONRPCError('failed to delete user group ID:%s %s' %
 
                               (user_group.users_group_id,
 
                                user_group.users_group_name)
 
                               )
 

	
 
    # permission check inside
 
    def add_user_to_user_group(self, usergroupid, userid):
 
        """
 
        Adds a user to a user group. If user exists in that group success will be
 
        `false`. This command can be executed only using api_key
 
        belonging to user with admin rights  or an admin of given user group
 

	
 
        :param usergroupid:
 
        :type usergroupid: int
 
        :param userid:
 
        :type userid: int
 

	
 
        OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : {
 
              "success": True|False # depends on if member is in group
 
              "msg": "added member `<username>` to user group `<groupname>` |
 
                      User is already in that group"
 

	
 
          }
 
          error :  null
 

	
 
        ERROR OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : null
 
          error :  {
 
            "failed to add member to user group `<user_group_name>`"
 
          }
 

	
 
        """
 
        user = get_user_or_error(userid)
 
        user_group = get_user_group_or_error(usergroupid)
 
        if not HasPermissionAny('hg.admin')():
 
            if not HasUserGroupPermissionLevel('admin')(user_group.users_group_name):
 
                raise JSONRPCError('user group `%s` does not exist' % (usergroupid,))
 

	
 
        try:
 
            ugm = UserGroupModel().add_user_to_group(user_group, user)
 
            success = True if ugm is not True else False
 
            msg = 'added member `%s` to user group `%s`' % (
 
                user.username, user_group.users_group_name
 
            )
 
            msg = msg if success else 'User is already in that group'
 
            Session().commit()
 

	
 
            return dict(
 
                success=success,
 
                msg=msg
 
            )
 
        except Exception:
 
            log.error(traceback.format_exc())
 
            raise JSONRPCError(
 
                'failed to add member to user group `%s`' % (
 
                    user_group.users_group_name,
 
                )
 
            )
 

	
 
    # permission check inside
 
    def remove_user_from_user_group(self, usergroupid, userid):
 
        """
 
        Removes a user from a user group. If user is not in given group success will
 
        be `false`. This command can be executed only
 
        using api_key belonging to user with admin rights or an admin of given user group
 

	
 
        :param usergroupid:
 
        :param userid:
 

	
 

	
 
        OUTPUT::
 

	
 
            id : <id_given_in_input>
 
            result: {
 
                      "success":  True|False,  # depends on if member is in group
 
                      "msg": "removed member <username> from user group <groupname> |
 
                              User wasn't in group"
 
                    }
 
            error:  null
 

	
 
        """
 
        user = get_user_or_error(userid)
 
        user_group = get_user_group_or_error(usergroupid)
 
        if not HasPermissionAny('hg.admin')():
 
            if not HasUserGroupPermissionLevel('admin')(user_group.users_group_name):
 
                raise JSONRPCError('user group `%s` does not exist' % (usergroupid,))
 

	
 
        try:
 
            success = UserGroupModel().remove_user_from_group(user_group, user)
 
            msg = 'removed member `%s` from user group `%s`' % (
 
                user.username, user_group.users_group_name
 
            )
 
            msg = msg if success else "User wasn't in group"
 
            Session().commit()
 
            return dict(success=success, msg=msg)
 
        except Exception:
 
            log.error(traceback.format_exc())
 
            raise JSONRPCError(
 
                'failed to remove member from user group `%s`' % (
 
                    user_group.users_group_name,
 
                )
 
            )
 

	
 
    # permission check inside
 
    def get_repo(self, repoid,
 
                 with_revision_names=Optional(False),
 
                 with_pullrequests=Optional(False)):
 
        """
 
        Gets an existing repository by it's name or repository_id. Members will return
 
        either users_group or user associated to that repository. This command can be
 
        executed only using api_key belonging to user with admin
 
        rights or regular user that have at least read access to repository.
 

	
 
        :param repoid: repository name or repository id
 
        :type repoid: str or int
 

	
 
        OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : {
 
            {
 
                "repo_id" :          "<repo_id>",
 
                "repo_name" :        "<reponame>"
 
                "repo_type" :        "<repo_type>",
 
                "clone_uri" :        "<clone_uri>",
 
                "enable_downloads":  "<bool>",
 
                "enable_statistics": "<bool>",
 
                "private":           "<bool>",
 
                "created_on" :       "<date_time_created>",
 
                "description" :      "<description>",
 
                "landing_rev":       "<landing_rev>",
 
                "last_changeset":    {
 
                                       "author":   "<full_author>",
 
                                       "date":     "<date_time_of_commit>",
 
                                       "message":  "<commit_message>",
 
                                       "raw_id":   "<raw_id>",
 
                                       "revision": "<numeric_revision>",
 
                                       "short_id": "<short_id>"
 
                                     }
 
                "owner":             "<repo_owner>",
 
                "fork_of":           "<name_of_fork_parent>",
 
                "members" :     [
 
                                  {
 
                                    "name":     "<username>",
 
                                    "type" :    "user",
 
                                    "permission" : "repository.(read|write|admin)"
 
                                  },
 
                                  …
 
                                  {
 
                                    "name":     "<usergroup name>",
 
                                    "type" :    "user_group",
 
                                    "permission" : "usergroup.(read|write|admin)"
 
                                  },
 
                                  …
 
                                ]
 
                 "followers":   [<user_obj>, ...],
 
                 <if with_revision_names == True>
 
                 "tags": {
 
                            "<tagname>": "<raw_id>",
 
                            ...
 
                         },
 
                 "branches": {
 
                            "<branchname>": "<raw_id>",
 
                            ...
 
                         },
 
                 "bookmarks": {
 
                            "<bookmarkname>": "<raw_id>",
 
                            ...
 
                         },
 
            }
 
          }
 
          error :  null
 

	
 
        """
 
        repo = get_repo_or_error(repoid)
 

	
 
        if not HasPermissionAny('hg.admin')():
 
            if not HasRepoPermissionLevel('read')(repo.repo_name):
 
                raise JSONRPCError('repository `%s` does not exist' % (repoid,))
 

	
 
        members = []
 
        for user in repo.repo_to_perm:
 
            perm = user.permission.permission_name
 
            user = user.user
 
            user_data = {
 
                'name': user.username,
 
                'type': "user",
 
                'permission': perm
 
            }
 
            members.append(user_data)
 

	
 
        for user_group in repo.users_group_to_perm:
 
            perm = user_group.permission.permission_name
 
            user_group = user_group.users_group
 
            user_group_data = {
 
                'name': user_group.users_group_name,
 
                'type': "user_group",
 
                'permission': perm
 
            }
 
            members.append(user_group_data)
 

	
 
        followers = [
 
            uf.user.get_api_data()
 
            for uf in repo.followers
 
        ]
 

	
 
        data = repo.get_api_data(with_revision_names=Optional.extract(with_revision_names),
 
                                 with_pullrequests=Optional.extract(with_pullrequests))
 
        data['members'] = members
 
        data['followers'] = followers
 
        return data
 

	
 
    # permission check inside
 
@@ -1992,432 +1992,432 @@ class ApiController(JSONRPCController):
 
        ERROR OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : null
 
          error :  {
 
            "failed to edit permission for user: `<userid>` in repo group: `<repo_group_name>`"
 
          }
 

	
 
        """
 

	
 
        repo_group = get_repo_group_or_error(repogroupid)
 

	
 
        if not HasPermissionAny('hg.admin')():
 
            if not HasRepoGroupPermissionLevel('admin')(repo_group.group_name):
 
                raise JSONRPCError('repository group `%s` does not exist' % (repogroupid,))
 

	
 
        user = get_user_or_error(userid)
 
        apply_to_children = Optional.extract(apply_to_children)
 

	
 
        try:
 
            RepoGroupModel().delete_permission(repo_group=repo_group,
 
                                               obj=user,
 
                                               obj_type="user",
 
                                               recursive=apply_to_children)
 

	
 
            Session().commit()
 
            return dict(
 
                msg='Revoked perm (recursive:%s) for user: `%s` in repo group: `%s`' % (
 
                    apply_to_children, user.username, repo_group.name
 
                ),
 
                success=True
 
            )
 
        except Exception:
 
            log.error(traceback.format_exc())
 
            raise JSONRPCError(
 
                'failed to edit permission for user: `%s` in repo group: `%s`' % (
 
                    userid, repo_group.name))
 

	
 
    # permission check inside
 
    def grant_user_group_permission_to_repo_group(
 
            self, repogroupid, usergroupid, perm,
 
            apply_to_children=Optional('none')):
 
        """
 
        Grant permission for user group on given repository group, or update
 
        existing one if found. This command can be executed only using
 
        api_key belonging to user with admin rights, or user who has admin
 
        right to given repository group.
 

	
 
        :param repogroupid: name or id of repository group
 
        :type repogroupid: str or int
 
        :param usergroupid: id of usergroup
 
        :type usergroupid: str or int
 
        :param perm: (group.(none|read|write|admin))
 
        :type perm: str
 
        :param apply_to_children: 'none', 'repos', 'groups', 'all'
 
        :type apply_to_children: str
 

	
 
        OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : {
 
            "msg" : "Granted perm: `<perm>` (recursive:<apply_to_children>) for user group: `<usersgroupname>` in repo group: `<repo_group_name>`",
 
            "success": true
 

	
 
          }
 
          error :  null
 

	
 
        ERROR OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : null
 
          error :  {
 
            "failed to edit permission for user group: `<usergroup>` in repo group: `<repo_group_name>`"
 
          }
 

	
 
        """
 
        repo_group = get_repo_group_or_error(repogroupid)
 
        perm = get_perm_or_error(perm, prefix='group.')
 
        user_group = get_user_group_or_error(usergroupid)
 
        if not HasPermissionAny('hg.admin')():
 
            if not HasRepoGroupPermissionLevel('admin')(repo_group.group_name):
 
                raise JSONRPCError(
 
                    'repository group `%s` does not exist' % (repogroupid,))
 

	
 
            if not HasUserGroupPermissionLevel('read')(user_group.users_group_name):
 
                raise JSONRPCError(
 
                    'user group `%s` does not exist' % (usergroupid,))
 

	
 
        apply_to_children = Optional.extract(apply_to_children)
 

	
 
        try:
 
            RepoGroupModel().add_permission(repo_group=repo_group,
 
                                            obj=user_group,
 
                                            obj_type="user_group",
 
                                            perm=perm,
 
                                            recursive=apply_to_children)
 
            Session().commit()
 
            return dict(
 
                msg='Granted perm: `%s` (recursive:%s) for user group: `%s` in repo group: `%s`' % (
 
                    perm.permission_name, apply_to_children,
 
                    user_group.users_group_name, repo_group.name
 
                ),
 
                success=True
 
            )
 
        except Exception:
 
            log.error(traceback.format_exc())
 
            raise JSONRPCError(
 
                'failed to edit permission for user group: `%s` in '
 
                'repo group: `%s`' % (
 
                    usergroupid, repo_group.name
 
                )
 
            )
 

	
 
    # permission check inside
 
    def revoke_user_group_permission_from_repo_group(
 
            self, repogroupid, usergroupid,
 
            apply_to_children=Optional('none')):
 
        """
 
        Revoke permission for user group on given repository. This command can be
 
        executed only using api_key belonging to user with admin rights, or
 
        user who has admin right to given repository group.
 

	
 
        :param repogroupid: name or id of repository group
 
        :type repogroupid: str or int
 
        :param usergroupid:
 
        :param apply_to_children: 'none', 'repos', 'groups', 'all'
 
        :type apply_to_children: str
 

	
 
        OUTPUT::
 

	
 
            id : <id_given_in_input>
 
            result: {
 
                      "msg" : "Revoked perm (recursive:<apply_to_children>) for user group: `<usersgroupname>` in repo group: `<repo_group_name>`",
 
                      "success": true
 
                    }
 
            error:  null
 

	
 
        ERROR OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : null
 
          error :  {
 
            "failed to edit permission for user group: `<usergroup>` in repo group: `<repo_group_name>`"
 
          }
 

	
 

	
 
        """
 
        repo_group = get_repo_group_or_error(repogroupid)
 
        user_group = get_user_group_or_error(usergroupid)
 
        if not HasPermissionAny('hg.admin')():
 
            if not HasRepoGroupPermissionLevel('admin')(repo_group.group_name):
 
                raise JSONRPCError(
 
                    'repository group `%s` does not exist' % (repogroupid,))
 

	
 
            if not HasUserGroupPermissionLevel('read')(user_group.users_group_name):
 
                raise JSONRPCError(
 
                    'user group `%s` does not exist' % (usergroupid,))
 

	
 
        apply_to_children = Optional.extract(apply_to_children)
 

	
 
        try:
 
            RepoGroupModel().delete_permission(repo_group=repo_group,
 
                                               obj=user_group,
 
                                               obj_type="user_group",
 
                                               recursive=apply_to_children)
 
            Session().commit()
 
            return dict(
 
                msg='Revoked perm (recursive:%s) for user group: `%s` in repo group: `%s`' % (
 
                    apply_to_children, user_group.users_group_name, repo_group.name
 
                ),
 
                success=True
 
            )
 
        except Exception:
 
            log.error(traceback.format_exc())
 
            raise JSONRPCError(
 
                'failed to edit permission for user group: `%s` in repo group: `%s`' % (
 
                    user_group.users_group_name, repo_group.name
 
                )
 
            )
 

	
 
    def get_gist(self, gistid):
 
        """
 
        Get given gist by id
 

	
 
        :param gistid: id of private or public gist
 
        :type gistid: str
 
        """
 
        gist = get_gist_or_error(gistid)
 
        if not HasPermissionAny('hg.admin')():
 
            if gist.owner_id != request.authuser.user_id:
 
                raise JSONRPCError('gist `%s` does not exist' % (gistid,))
 
        return gist.get_api_data()
 

	
 
    def get_gists(self, userid=Optional(OAttr('apiuser'))):
 
        """
 
        Get all gists for given user. If userid is empty returned gists
 
        are for user who called the api
 

	
 
        :param userid: user to get gists for
 
        :type userid: Optional(str or int)
 
        """
 
        if not HasPermissionAny('hg.admin')():
 
            # make sure normal user does not pass someone else userid,
 
            # he is not allowed to do that
 
            if not isinstance(userid, Optional) and userid != request.authuser.user_id:
 
                raise JSONRPCError(
 
                    'userid is not the same as your user'
 
                )
 

	
 
        if isinstance(userid, Optional):
 
            user_id = request.authuser.user_id
 
        else:
 
            user_id = get_user_or_error(userid).user_id
 

	
 
        return [
 
            gist.get_api_data()
 
            for gist in Gist().query()
 
                .filter_by(is_expired=False)
 
                .filter(Gist.owner_id == user_id)
 
                .order_by(Gist.created_on.desc())
 
        ]
 

	
 
    def create_gist(self, files, owner=Optional(OAttr('apiuser')),
 
                    gist_type=Optional(Gist.GIST_PUBLIC), lifetime=Optional(-1),
 
                    description=Optional('')):
 

	
 
        """
 
        Creates new Gist
 

	
 
        :param files: files to be added to gist
 
            {'filename': {'content':'...', 'lexer': null},
 
             'filename2': {'content':'...', 'lexer': null}}
 
        :type files: dict
 
        :param owner: gist owner, defaults to api method caller
 
        :type owner: Optional(str or int)
 
        :param gist_type: type of gist 'public' or 'private'
 
        :type gist_type: Optional(str)
 
        :param lifetime: time in minutes of gist lifetime
 
        :type lifetime: Optional(int)
 
        :param description: gist description
 
        :type description: Optional(str)
 

	
 
        OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : {
 
            "msg": "created new gist",
 
            "gist": {}
 
          }
 
          error :  null
 

	
 
        ERROR OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : null
 
          error :  {
 
            "failed to create gist"
 
          }
 

	
 
        """
 
        try:
 
            if isinstance(owner, Optional):
 
                owner = request.authuser.user_id
 

	
 
            owner = get_user_or_error(owner)
 
            description = Optional.extract(description)
 
            gist_type = Optional.extract(gist_type)
 
            lifetime = Optional.extract(lifetime)
 

	
 
            gist = GistModel().create(description=description,
 
                                      owner=owner,
 
                                      ip_addr=request.ip_addr,
 
                                      gist_mapping=files,
 
                                      gist_type=gist_type,
 
                                      lifetime=lifetime)
 
            Session().commit()
 
            return dict(
 
                msg='created new gist',
 
                gist=gist.get_api_data()
 
            )
 
        except Exception:
 
            log.error(traceback.format_exc())
 
            raise JSONRPCError('failed to create gist')
 

	
 
    # def update_gist(self, gistid, files, owner=Optional(OAttr('apiuser')),
 
    #                 gist_type=Optional(Gist.GIST_PUBLIC),
 
    #                 gist_lifetime=Optional(-1), gist_description=Optional('')):
 
    #     gist = get_gist_or_error(gistid)
 
    #     updates = {}
 

	
 
    # permission check inside
 
    def delete_gist(self, gistid):
 
        """
 
        Deletes existing gist
 

	
 
        :param gistid: id of gist to delete
 
        :type gistid: str
 

	
 
        OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : {
 
            "deleted gist ID: <gist_id>",
 
            "gist": null
 
          }
 
          error :  null
 

	
 
        ERROR OUTPUT::
 

	
 
          id : <id_given_in_input>
 
          result : null
 
          error :  {
 
            "failed to delete gist ID:<gist_id>"
 
          }
 

	
 
        """
 
        gist = get_gist_or_error(gistid)
 
        if not HasPermissionAny('hg.admin')():
 
            if gist.owner_id != request.authuser.user_id:
 
                raise JSONRPCError('gist `%s` does not exist' % (gistid,))
 

	
 
        try:
 
            GistModel().delete(gist)
 
            Session().commit()
 
            return dict(
 
                msg='deleted gist ID:%s' % (gist.gist_access_id,),
 
                gist=None
 
            )
 
        except Exception:
 
            log.error(traceback.format_exc())
 
            raise JSONRPCError('failed to delete gist ID:%s'
 
                               % (gist.gist_access_id,))
 

	
 
    # permission check inside
 
    def get_changesets(self, repoid, start=None, end=None, start_date=None,
 
                       end_date=None, branch_name=None, reverse=False, with_file_list=False, max_revisions=None):
 
        repo = get_repo_or_error(repoid)
 
        if not HasRepoPermissionLevel('read')(repo.repo_name):
 
            raise JSONRPCError('Access denied to repo %s' % repo.repo_name)
 

	
 
        format = "%Y-%m-%dT%H:%M:%S"
 
        try:
 
            return [e.__json__(with_file_list) for e in
 
                repo.scm_instance.get_changesets(start,
 
                                                 end,
 
                                                 datetime.strptime(start_date, format) if start_date else None,
 
                                                 datetime.strptime(end_date, format) if end_date else None,
 
                                                 branch_name,
 
                                                 reverse, max_revisions)]
 
        except EmptyRepositoryError as e:
 
            raise JSONRPCError('Repository is empty')
 

	
 
    # permission check inside
 
    def get_changeset(self, repoid, raw_id, with_reviews=Optional(False)):
 
        repo = get_repo_or_error(repoid)
 
        if not HasRepoPermissionLevel('read')(repo.repo_name):
 
            raise JSONRPCError('Access denied to repo %s' % repo.repo_name)
 
        changeset = repo.get_changeset(raw_id)
 
        if isinstance(changeset, EmptyChangeset):
 
            raise JSONRPCError('Changeset %s does not exist' % raw_id)
 

	
 
        info = dict(changeset.as_dict())
 

	
 
        with_reviews = Optional.extract(with_reviews)
 
        if with_reviews:
 
            reviews = ChangesetStatusModel().get_statuses(
 
                                repo.repo_name, raw_id)
 
            info["reviews"] = reviews
 

	
 
        return info
 

	
 
    # permission check inside
 
    def get_pullrequest(self, pullrequest_id):
 
        """
 
        Get given pull request by id
 
        """
 
        pull_request = PullRequest.get(pullrequest_id)
 
        if pull_request is None:
 
            raise JSONRPCError('pull request `%s` does not exist' % (pullrequest_id,))
 
        if not HasRepoPermissionLevel('read')(pull_request.org_repo.repo_name):
 
            raise JSONRPCError('not allowed')
 
        return pull_request.get_api_data()
 

	
 
    # permission check inside
 
    def comment_pullrequest(self, pull_request_id, comment_msg=u'', status=None, close_pr=False):
 
    def comment_pullrequest(self, pull_request_id, comment_msg='', status=None, close_pr=False):
 
        """
 
        Add comment, close and change status of pull request.
 
        """
 
        apiuser = get_user_or_error(request.authuser.user_id)
 
        pull_request = PullRequest.get(pull_request_id)
 
        if pull_request is None:
 
            raise JSONRPCError('pull request `%s` does not exist' % (pull_request_id,))
 
        if (not HasRepoPermissionLevel('read')(pull_request.org_repo.repo_name)):
 
            raise JSONRPCError('No permission to add comment. User needs at least reading permissions'
 
                               ' to the source repository.')
 
        owner = apiuser.user_id == pull_request.owner_id
 
        reviewer = apiuser.user_id in [reviewer.user_id for reviewer in pull_request.reviewers]
 
        if close_pr and not (apiuser.admin or owner):
 
            raise JSONRPCError('No permission to close pull request. User needs to be admin or owner.')
 
        if status and not (apiuser.admin or owner or reviewer):
 
            raise JSONRPCError('No permission to change pull request status. User needs to be admin, owner or reviewer.')
 
        if pull_request.is_closed():
 
            raise JSONRPCError('pull request is already closed')
 

	
 
        comment = ChangesetCommentsModel().create(
 
            text=comment_msg,
 
            repo=pull_request.org_repo.repo_id,
 
            author=apiuser.user_id,
 
            pull_request=pull_request.pull_request_id,
 
            f_path=None,
 
            line_no=None,
 
            status_change=ChangesetStatus.get_status_lbl(status),
 
            closing_pr=close_pr
 
        )
 
        action_logger(apiuser,
 
                      'user_commented_pull_request:%s' % pull_request_id,
 
                      pull_request.org_repo, request.ip_addr)
 
        if status:
 
            ChangesetStatusModel().set_status(
 
                pull_request.org_repo_id,
 
                status,
 
                apiuser.user_id,
 
                comment,
 
                pull_request=pull_request_id
 
            )
 
        if close_pr:
 
            PullRequestModel().close_pull_request(pull_request_id)
 
            action_logger(apiuser,
 
                          'user_closed_pull_request:%s' % pull_request_id,
 
                          pull_request.org_repo, request.ip_addr)
 
        Session().commit()
 
        return True
kallithea/controllers/search.py
Show inline comments
 
# -*- 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.controllers.search
 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 

	
 
Search controller for 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 7, 2010
 
:author: marcink
 
:copyright: (c) 2013 RhodeCode GmbH, and others.
 
:license: GPLv3, see LICENSE.md for more details.
 
"""
 

	
 
import logging
 
import traceback
 

	
 
from tg import config, request
 
from tg import tmpl_context as c
 
from tg.i18n import ugettext as _
 
from whoosh.index import EmptyIndexError, exists_in, open_dir
 
from whoosh.qparser import QueryParser, QueryParserError
 
from whoosh.query import Phrase, Prefix
 

	
 
from kallithea.lib.auth import LoginRequired
 
from kallithea.lib.base import BaseRepoController, render
 
from kallithea.lib.indexers import CHGSET_IDX_NAME, CHGSETS_SCHEMA, IDX_NAME, SCHEMA, WhooshResultWrapper
 
from kallithea.lib.page import Page
 
from kallithea.lib.utils2 import safe_int
 
from kallithea.model.repo import RepoModel
 

	
 

	
 
log = logging.getLogger(__name__)
 

	
 

	
 
class SearchController(BaseRepoController):
 

	
 
    @LoginRequired(allow_default_user=True)
 
    def index(self, repo_name=None):
 
        c.repo_name = repo_name
 
        c.formated_results = []
 
        c.runtime = ''
 
        c.cur_query = request.GET.get('q', None)
 
        c.cur_type = request.GET.get('type', 'content')
 
        c.cur_search = search_type = {'content': 'content',
 
                                      'commit': 'message',
 
                                      'path': 'path',
 
                                      'repository': 'repository'
 
                                      }.get(c.cur_type, 'content')
 

	
 
        index_name = {
 
            'content': IDX_NAME,
 
            'commit': CHGSET_IDX_NAME,
 
            'path': IDX_NAME
 
        }.get(c.cur_type, IDX_NAME)
 

	
 
        schema_defn = {
 
            'content': SCHEMA,
 
            'commit': CHGSETS_SCHEMA,
 
            'path': SCHEMA
 
        }.get(c.cur_type, SCHEMA)
 

	
 
        log.debug('IDX: %s', index_name)
 
        log.debug('SCHEMA: %s', schema_defn)
 

	
 
        if c.cur_query:
 
            cur_query = c.cur_query.lower()
 
            log.debug(cur_query)
 

	
 
        if c.cur_query:
 
            p = safe_int(request.GET.get('page'), 1)
 
            highlight_items = set()
 
            index_dir = config['index_dir']
 
            try:
 
                if not exists_in(index_dir, index_name):
 
                    raise EmptyIndexError
 
                idx = open_dir(index_dir, indexname=index_name)
 
                searcher = idx.searcher()
 

	
 
                qp = QueryParser(search_type, schema=schema_defn)
 
                if c.repo_name:
 
                    # use "repository_rawname:" instead of "repository:"
 
                    # for case-sensitive matching
 
                    cur_query = u'repository_rawname:%s %s' % (c.repo_name, cur_query)
 
                    cur_query = 'repository_rawname:%s %s' % (c.repo_name, cur_query)
 
                try:
 
                    query = qp.parse(cur_query)
 
                    # extract words for highlight
 
                    if isinstance(query, Phrase):
 
                        highlight_items.update(query.words)
 
                    elif isinstance(query, Prefix):
 
                        highlight_items.add(query.text)
 
                    else:
 
                        for i in query.all_terms():
 
                            if i[0] in ['content', 'message']:
 
                                highlight_items.add(i[1])
 

	
 
                    matcher = query.matcher(searcher)
 

	
 
                    log.debug('query: %s', query)
 
                    log.debug('hl terms: %s', highlight_items)
 
                    results = searcher.search(query)
 
                    res_ln = len(results)
 
                    c.runtime = '%s results (%.3f seconds)' % (
 
                        res_ln, results.runtime
 
                    )
 

	
 
                    repo_location = RepoModel().repos_path
 
                    c.formated_results = Page(
 
                        WhooshResultWrapper(search_type, searcher, matcher,
 
                                            highlight_items, repo_location),
 
                        page=p,
 
                        item_count=res_ln,
 
                        items_per_page=10,
 
                        type=c.cur_type,
 
                        q=c.cur_query,
 
                    )
 

	
 
                except QueryParserError:
 
                    c.runtime = _('Invalid search query. Try quoting it.')
 
                searcher.close()
 
            except EmptyIndexError:
 
                log.error("Empty search index - run 'kallithea-cli index-create' regularly")
 
                c.runtime = _('The server has no search index.')
 
            except Exception:
 
                log.error(traceback.format_exc())
 
                c.runtime = _('An error occurred during search operation.')
 

	
 
        # Return a rendered template
 
        return render('/search/search.html')
kallithea/lib/auth_modules/__init__.py
Show inline comments
 
# -*- 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/>.
 
"""
 
Authentication modules
 
"""
 

	
 
import importlib
 
import logging
 
import traceback
 

	
 
from kallithea.lib.auth import AuthUser, PasswordGenerator
 
from kallithea.lib.compat import formatted_json, hybrid_property
 
from kallithea.lib.utils2 import str2bool
 
from kallithea.model.db import Setting, User
 
from kallithea.model.meta import Session
 
from kallithea.model.user import UserModel
 
from kallithea.model.user_group import UserGroupModel
 

	
 

	
 
log = logging.getLogger(__name__)
 

	
 

	
 
class LazyFormencode(object):
 
    def __init__(self, formencode_obj, *args, **kwargs):
 
        self.formencode_obj = formencode_obj
 
        self.args = args
 
        self.kwargs = kwargs
 

	
 
    def __call__(self, *args, **kwargs):
 
        from inspect import isfunction
 
        formencode_obj = self.formencode_obj
 
        if isfunction(formencode_obj):
 
            # case we wrap validators into functions
 
            formencode_obj = self.formencode_obj(*args, **kwargs)
 
        return formencode_obj(*self.args, **self.kwargs)
 

	
 

	
 
class KallitheaAuthPluginBase(object):
 
    auth_func_attrs = {
 
        "username": "unique username",
 
        "firstname": "first name",
 
        "lastname": "last name",
 
        "email": "email address",
 
        "groups": '["list", "of", "groups"]',
 
        "extern_name": "name in external source of record",
 
        "admin": 'True|False defines if user should be Kallithea admin',
 
    }
 

	
 
    @property
 
    def validators(self):
 
        """
 
        Exposes Kallithea validators modules
 
        """
 
        # this is a hack to overcome issues with pylons threadlocals and
 
        # translator object _() not being registered properly.
 
        class LazyCaller(object):
 
            def __init__(self, name):
 
                self.validator_name = name
 

	
 
            def __call__(self, *args, **kwargs):
 
                from kallithea.model import validators as v
 
                obj = getattr(v, self.validator_name)
 
                #log.debug('Initializing lazy formencode object: %s', obj)
 
                return LazyFormencode(obj, *args, **kwargs)
 

	
 
        class ProxyGet(object):
 
            def __getattribute__(self, name):
 
                return LazyCaller(name)
 

	
 
        return ProxyGet()
 

	
 
    @hybrid_property
 
    def name(self):
 
        """
 
        Returns the name of this authentication plugin.
 

	
 
        :returns: string
 
        """
 
        raise NotImplementedError("Not implemented in base class")
 

	
 
    @hybrid_property
 
    def is_container_auth(self):
 
        """
 
        Returns bool if this module uses container auth.
 

	
 
        This property will trigger an automatic call to authenticate on
 
        a visit to the website or during a push/pull.
 

	
 
        :returns: bool
 
        """
 
        return False
 

	
 
    def accepts(self, user, accepts_empty=True):
 
        """
 
        Checks if this authentication module should accept a request for
 
        the current user.
 

	
 
        :param user: user object fetched using plugin's get_user() method.
 
        :param accepts_empty: if True accepts don't allow the user to be empty
 
        :returns: boolean
 
        """
 
        plugin_name = self.name
 
        if not user and not accepts_empty:
 
            log.debug('User is empty not allowed to authenticate')
 
            return False
 

	
 
        if user and user.extern_type and user.extern_type != plugin_name:
 
            log.debug('User %s should authenticate using %s this is %s, skipping',
 
                      user, user.extern_type, plugin_name)
 

	
 
            return False
 
        return True
 

	
 
    def get_user(self, username=None, **kwargs):
 
        """
 
        Helper method for user fetching in plugins, by default it's using
 
        simple fetch by username, but this method can be customized in plugins
 
        eg. container auth plugin to fetch user by environ params
 

	
 
        :param username: username if given to fetch from database
 
        :param kwargs: extra arguments needed for user fetching.
 
        """
 
        user = None
 
        log.debug('Trying to fetch user `%s` from Kallithea database',
 
                  username)
 
        if username:
 
            user = User.get_by_username_or_email(username)
 
            if user is None:
 
                log.debug('Fallback to fetch user in case insensitive mode')
 
                user = User.get_by_username(username, case_insensitive=True)
 
        else:
 
            log.debug('provided username:`%s` is empty skipping...', username)
 
        return user
 

	
 
    def settings(self):
 
        """
 
        Return a list of the form:
 
        [
 
            {
 
                "name": "OPTION_NAME",
 
                "type": "[bool|password|string|int|select]",
 
                ["values": ["opt1", "opt2", ...]]
 
                "validator": "expr"
 
                "description": "A short description of the option" [,
 
                "default": Default Value],
 
                ["formname": "Friendly Name for Forms"]
 
            } [, ...]
 
        ]
 

	
 
        This is used to interrogate the authentication plugin as to what
 
        settings it expects to be present and configured.
 

	
 
        'type' is a shorthand notation for what kind of value this option is.
 
        This is primarily used by the auth web form to control how the option
 
        is configured.
 
                bool : checkbox
 
                password : password input box
 
                string : input box
 
                select : single select dropdown
 

	
 
        'validator' is an lazy instantiated form field validator object, ala
 
        formencode. You need to *call* this object to init the validators.
 
        All calls to Kallithea validators should be used through self.validators
 
        which is a lazy loading proxy of formencode module.
 
        """
 
        raise NotImplementedError("Not implemented in base class")
 

	
 
    def plugin_settings(self):
 
        """
 
        This method is called by the authentication framework, not the .settings()
 
        method. This method adds a few default settings (e.g., "enabled"), so that
 
        plugin authors don't have to maintain a bunch of boilerplate.
 

	
 
        OVERRIDING THIS METHOD WILL CAUSE YOUR PLUGIN TO FAIL.
 
        """
 

	
 
        rcsettings = self.settings()
 
        rcsettings.insert(0, {
 
            "name": "enabled",
 
            "validator": self.validators.StringBoolean(if_missing=False),
 
            "type": "bool",
 
            "description": "Enable or Disable this Authentication Plugin",
 
            "formname": "Enabled"
 
            }
 
        )
 
        return rcsettings
 

	
 
    def auth(self, userobj, username, passwd, settings, **kwargs):
 
        """
 
        Given a user object (which may be None), username, a plaintext password,
 
        and a settings object (containing all the keys needed as listed in settings()),
 
        authenticate this user's login attempt.
 

	
 
        Return None on failure. On success, return a dictionary with keys from
 
        KallitheaAuthPluginBase.auth_func_attrs.
 

	
 
        This is later validated for correctness.
 
        """
 
        raise NotImplementedError("not implemented in base class")
 

	
 
    def _authenticate(self, userobj, username, passwd, settings, **kwargs):
 
        """
 
        Wrapper to call self.auth() that validates call on it
 
        """
 
        user_data = self.auth(userobj, username, passwd, settings, **kwargs)
 
        if user_data is not None:
 
            return self._validate_auth_return(user_data)
 
        return None
 

	
 
    def _validate_auth_return(self, user_data):
 
        if not isinstance(user_data, dict):
 
            raise Exception('returned value from auth must be a dict')
 
        for k in self.auth_func_attrs:
 
            if k not in user_data:
 
                raise Exception('Missing %s attribute from returned data' % k)
 
        return user_data
 

	
 

	
 
class KallitheaExternalAuthPlugin(KallitheaAuthPluginBase):
 
    def use_fake_password(self):
 
        """
 
        Return a boolean that indicates whether or not we should set the user's
 
        password to a random value when it is authenticated by this plugin.
 
        If your plugin provides authentication, then you will generally want this.
 

	
 
        :returns: boolean
 
        """
 
        raise NotImplementedError("Not implemented in base class")
 

	
 
    def _authenticate(self, userobj, username, passwd, settings, **kwargs):
 
        user_data = super(KallitheaExternalAuthPlugin, self)._authenticate(
 
            userobj, username, passwd, settings, **kwargs)
 
        if user_data is not None:
 
            if userobj is None: # external authentication of unknown user that will be created soon
 
                def_user_perms = AuthUser(dbuser=User.get_default_user()).permissions['global']
 
                active = 'hg.extern_activate.auto' in def_user_perms
 
            else:
 
                active = userobj.active
 

	
 
            if self.use_fake_password():
 
                # Randomize the PW because we don't need it, but don't want
 
                # them blank either
 
                passwd = PasswordGenerator().gen_password(length=8)
 

	
 
            log.debug('Updating or creating user info from %s plugin',
 
                      self.name)
 
            user = UserModel().create_or_update(
 
                username=user_data['username'],
 
                password=passwd,
 
                email=user_data["email"],
 
                firstname=user_data["firstname"],
 
                lastname=user_data["lastname"],
 
                active=active,
 
                admin=user_data["admin"],
 
                extern_name=user_data["extern_name"],
 
                extern_type=self.name,
 
            )
 
            # enforce user is just in given groups, all of them has to be ones
 
            # created from plugins. We store this info in _group_data JSON field
 
            groups = user_data['groups'] or []
 
            UserGroupModel().enforce_groups(user, groups, self.name)
 
            Session().commit()
 
        return user_data
 

	
 

	
 
def loadplugin(plugin):
 
    """
 
    Imports, instantiates, and returns the authentication plugin in the module named by plugin
 
    (e.g., plugin='kallithea.lib.auth_modules.auth_internal'). Returns an instance of the
 
    KallitheaAuthPluginBase subclass on success, raises exceptions on failure.
 

	
 
    raises:
 
        AttributeError -- no KallitheaAuthPlugin class in the module
 
        TypeError -- if the KallitheaAuthPlugin is not a subclass of ours KallitheaAuthPluginBase
 
        ImportError -- if we couldn't import the plugin at all
 
    """
 
    log.debug("Importing %s", plugin)
 
    if not plugin.startswith(u'kallithea.lib.auth_modules.auth_'):
 
        parts = plugin.split(u'.lib.auth_modules.auth_', 1)
 
    if not plugin.startswith('kallithea.lib.auth_modules.auth_'):
 
        parts = plugin.split('.lib.auth_modules.auth_', 1)
 
        if len(parts) == 2:
 
            _module, pn = parts
 
            plugin = u'kallithea.lib.auth_modules.auth_' + pn
 
            plugin = 'kallithea.lib.auth_modules.auth_' + pn
 
    PLUGIN_CLASS_NAME = "KallitheaAuthPlugin"
 
    try:
 
        module = importlib.import_module(plugin)
 
    except (ImportError, TypeError):
 
        log.error(traceback.format_exc())
 
        # TODO: make this more error prone, if by some accident we screw up
 
        # the plugin name, the crash is pretty bad and hard to recover
 
        raise
 

	
 
    log.debug("Loaded auth plugin from %s (module:%s, file:%s)",
 
              plugin, module.__name__, module.__file__)
 

	
 
    pluginclass = getattr(module, PLUGIN_CLASS_NAME)
 
    if not issubclass(pluginclass, KallitheaAuthPluginBase):
 
        raise TypeError("Authentication class %s.KallitheaAuthPlugin is not "
 
                        "a subclass of %s" % (plugin, KallitheaAuthPluginBase))
 

	
 
    plugin = pluginclass()
 
    if plugin.plugin_settings.__func__ != KallitheaAuthPluginBase.plugin_settings:
 
        raise TypeError("Authentication class %s.KallitheaAuthPluginBase "
 
                        "has overridden the plugin_settings method, which is "
 
                        "forbidden." % plugin)
 
    return plugin
 

	
 

	
 
def get_auth_plugins():
 
    """Return a list of instances of plugins that are available and enabled"""
 
    auth_plugins = []
 
    for plugin_name in Setting.get_by_name("auth_plugins").app_settings_value:
 
        try:
 
            plugin = loadplugin(plugin_name)
 
        except Exception:
 
            log.exception('Failed to load authentication module %s' % (plugin_name))
 
        else:
 
            auth_plugins.append(plugin)
 
    return auth_plugins
 

	
 

	
 
def authenticate(username, password, environ=None):
 
    """
 
    Authentication function used for access control,
 
    It tries to authenticate based on enabled authentication modules.
 

	
 
    :param username: username can be empty for container auth
 
    :param password: password can be empty for container auth
 
    :param environ: environ headers passed for container auth
 
    :returns: None if auth failed, user_data dict if auth is correct
 
    """
 

	
 
    auth_plugins = get_auth_plugins()
 
    for plugin in auth_plugins:
 
        module = plugin.__class__.__module__
 
        log.debug('Trying authentication using %s', module)
 
        # load plugin settings from Kallithea database
 
        plugin_name = plugin.name
 
        plugin_settings = {}
 
        for v in plugin.plugin_settings():
 
            conf_key = "auth_%s_%s" % (plugin_name, v["name"])
 
            setting = Setting.get_by_name(conf_key)
 
            plugin_settings[v["name"]] = setting.app_settings_value if setting else None
 
        log.debug('Settings for auth plugin %s:\n%s', plugin_name, formatted_json(plugin_settings))
 

	
 
        if not str2bool(plugin_settings["enabled"]):
 
            log.info("Authentication plugin %s is disabled, skipping for %s",
 
                     module, username)
 
            continue
 

	
 
        # use plugin's method of user extraction.
 
        user = plugin.get_user(username, environ=environ,
 
                               settings=plugin_settings)
 
        log.debug('Plugin %s extracted user `%s`', module, user)
 

	
 
        if user is not None and not user.active: # give up, way before creating AuthUser
 
            log.error("Rejecting authentication of in-active user %s", user)
 
            continue
 

	
 
        if not plugin.accepts(user):
 
            log.debug('Plugin %s does not accept user `%s` for authentication',
 
                      module, user)
 
            continue
 
        else:
 
            log.debug('Plugin %s accepted user `%s` for authentication',
 
                      module, user)
 
            # The user might have tried to authenticate using their email address,
 
            # then the username variable wouldn't contain a valid username.
 
            # But as the plugin has accepted the user, .username field should
 
            # have a valid username, so use it for authentication purposes.
 
            if user is not None:
 
                username = user.username
 

	
 
        log.info('Authenticating user using %s plugin', module)
 

	
 
        # _authenticate is a wrapper for .auth() method of plugin.
 
        # it checks if .auth() sends proper data. For KallitheaExternalAuthPlugin
 
        # it also maps users to Database and maps the attributes returned
 
        # from .auth() to Kallithea database. If this function returns data
 
        # then auth is correct.
 
        user_data = plugin._authenticate(user, username, password,
 
                                           plugin_settings,
 
                                           environ=environ or {})
 
        log.debug('Plugin user data: %s', user_data)
 

	
 
        if user_data is not None:
 
            log.debug('Plugin returned proper authentication data')
 
            return user_data
 

	
 
        # we failed to Auth because .auth() method didn't return the user
 
        if username:
 
            log.warning("User `%s` failed to authenticate against %s",
 
                        username, module)
 
    return None
 

	
 

	
 
def get_managed_fields(user):
 
    """return list of fields that are managed by the user's auth source, usually some of
 
    'username', 'firstname', 'lastname', 'email', 'password'
 
    """
 
    auth_plugins = get_auth_plugins()
 
    for plugin in auth_plugins:
 
        module = plugin.__class__.__module__
 
        log.debug('testing %s (%s) with auth plugin %s', user, user.extern_type, module)
 
        if plugin.name == user.extern_type:
 
            return plugin.get_managed_fields()
 
    log.error('no auth plugin %s found for %s', user.extern_type, user)
 
    return [] # TODO: Fail badly instead of allowing everything to be edited?
kallithea/lib/db_manage.py
Show inline comments
 
@@ -4,419 +4,419 @@
 
# 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.db_manage
 
~~~~~~~~~~~~~~~~~~~~~~~
 

	
 
Database creation, and setup module for Kallithea. Used for creation
 
of database as well as for migration operations
 

	
 
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: Apr 10, 2010
 
:author: marcink
 
:copyright: (c) 2013 RhodeCode GmbH, and others.
 
:license: GPLv3, see LICENSE.md for more details.
 
"""
 

	
 
from __future__ import print_function
 

	
 
import logging
 
import os
 
import sys
 
import uuid
 

	
 
import alembic.command
 
import alembic.config
 
import sqlalchemy
 
from sqlalchemy.engine import create_engine
 

	
 
from kallithea.model.base import init_model
 
from kallithea.model.db import Permission, RepoGroup, Repository, Setting, Ui, User, UserRepoGroupToPerm, UserToPerm
 
#from kallithea.model import meta
 
from kallithea.model.meta import Base, Session
 
from kallithea.model.permission import PermissionModel
 
from kallithea.model.repo_group import RepoGroupModel
 
from kallithea.model.user import UserModel
 

	
 

	
 
log = logging.getLogger(__name__)
 

	
 

	
 
class DbManage(object):
 
    def __init__(self, dbconf, root, tests=False, SESSION=None, cli_args=None):
 
        self.dbname = dbconf.split('/')[-1]
 
        self.tests = tests
 
        self.root = root
 
        self.dburi = dbconf
 
        self.db_exists = False
 
        self.cli_args = cli_args or {}
 
        self.init_db(SESSION=SESSION)
 

	
 
    def _ask_ok(self, msg):
 
        """Invoke ask_ok unless the force_ask option provides the answer"""
 
        force_ask = self.cli_args.get('force_ask')
 
        if force_ask is not None:
 
            return force_ask
 
        from kallithea.lib.utils2 import ask_ok
 
        return ask_ok(msg)
 

	
 
    def init_db(self, SESSION=None):
 
        if SESSION:
 
            self.sa = SESSION
 
        else:
 
            # init new sessions
 
            engine = create_engine(self.dburi)
 
            init_model(engine)
 
            self.sa = Session()
 

	
 
    def create_tables(self, override=False):
 
        """
 
        Create a auth database
 
        """
 

	
 
        log.info("Any existing database is going to be destroyed")
 
        if self.tests:
 
            destroy = True
 
        else:
 
            destroy = self._ask_ok('Are you sure to destroy old database ? [y/n]')
 
        if not destroy:
 
            print('Nothing done.')
 
            sys.exit(0)
 
        if destroy:
 
            # drop and re-create old schemas
 

	
 
            url = sqlalchemy.engine.url.make_url(self.dburi)
 
            database = url.database
 

	
 
            # Some databases enforce foreign key constraints and Base.metadata.drop_all() doesn't work
 
            if url.drivername == 'mysql':
 
                url.database = None  # don't connect to the database (it might not exist)
 
                engine = sqlalchemy.create_engine(url)
 
                with engine.connect() as conn:
 
                    conn.execute('DROP DATABASE IF EXISTS ' + database)
 
                    conn.execute('CREATE DATABASE ' + database)
 
            elif url.drivername == 'postgresql':
 
                from psycopg2.extensions import ISOLATION_LEVEL_AUTOCOMMIT
 
                url.database = 'postgres'  # connect to the system database (as the real one might not exist)
 
                engine = sqlalchemy.create_engine(url)
 
                with engine.connect() as conn:
 
                    conn.connection.set_isolation_level(ISOLATION_LEVEL_AUTOCOMMIT)
 
                    conn.execute('DROP DATABASE IF EXISTS ' + database)
 
                    conn.execute('CREATE DATABASE ' + database)
 
            else:
 
                # known to work on SQLite - possibly not on other databases with strong referential integrity
 
                Base.metadata.drop_all()
 

	
 
        checkfirst = not override
 
        Base.metadata.create_all(checkfirst=checkfirst)
 

	
 
        # Create an Alembic configuration and generate the version table,
 
        # "stamping" it with the most recent Alembic migration revision, to
 
        # tell Alembic that all the schema upgrades are already in effect.
 
        alembic_cfg = alembic.config.Config()
 
        alembic_cfg.set_main_option('script_location', 'kallithea:alembic')
 
        alembic_cfg.set_main_option('sqlalchemy.url', self.dburi)
 
        # This command will give an error in an Alembic multi-head scenario,
 
        # but in practice, such a scenario should not come up during database
 
        # creation, even during development.
 
        alembic.command.stamp(alembic_cfg, 'head')
 

	
 
        log.info('Created tables for %s', self.dbname)
 

	
 
    def fix_repo_paths(self):
 
        """
 
        Fixes a old kallithea version path into new one without a '*'
 
        """
 

	
 
        paths = Ui.query() \
 
                .filter(Ui.ui_key == '/') \
 
                .scalar()
 

	
 
        paths.ui_value = paths.ui_value.replace('*', '')
 

	
 
        self.sa.commit()
 

	
 
    def fix_default_user(self):
 
        """
 
        Fixes a old default user with some 'nicer' default values,
 
        used mostly for anonymous access
 
        """
 
        def_user = User.query().filter_by(is_default_user=True).one()
 

	
 
        def_user.name = 'Anonymous'
 
        def_user.lastname = 'User'
 
        def_user.email = 'anonymous@kallithea-scm.org'
 

	
 
        self.sa.commit()
 

	
 
    def fix_settings(self):
 
        """
 
        Fixes kallithea settings adds ga_code key for google analytics
 
        """
 

	
 
        hgsettings3 = Setting('ga_code', '')
 

	
 
        self.sa.add(hgsettings3)
 
        self.sa.commit()
 

	
 
    def admin_prompt(self, second=False):
 
        if not self.tests:
 
            import getpass
 

	
 
            username = self.cli_args.get('username')
 
            password = self.cli_args.get('password')
 
            email = self.cli_args.get('email')
 

	
 
            def get_password():
 
                password = getpass.getpass('Specify admin password '
 
                                           '(min 6 chars):')
 
                confirm = getpass.getpass('Confirm password:')
 

	
 
                if password != confirm:
 
                    log.error('passwords mismatch')
 
                    return False
 
                if len(password) < 6:
 
                    log.error('password is to short use at least 6 characters')
 
                    return False
 

	
 
                return password
 
            if username is None:
 
                username = input('Specify admin username:')
 
            if password is None:
 
                password = get_password()
 
                if not password:
 
                    # second try
 
                    password = get_password()
 
                    if not password:
 
                        sys.exit()
 
            if email is None:
 
                email = input('Specify admin email:')
 
            self.create_user(username, password, email, True)
 
        else:
 
            log.info('creating admin and regular test users')
 
            from kallithea.tests.base import TEST_USER_ADMIN_LOGIN, \
 
                TEST_USER_ADMIN_PASS, TEST_USER_ADMIN_EMAIL, \
 
                TEST_USER_REGULAR_LOGIN, TEST_USER_REGULAR_PASS, \
 
                TEST_USER_REGULAR_EMAIL, TEST_USER_REGULAR2_LOGIN, \
 
                TEST_USER_REGULAR2_PASS, TEST_USER_REGULAR2_EMAIL
 

	
 
            self.create_user(TEST_USER_ADMIN_LOGIN, TEST_USER_ADMIN_PASS,
 
                             TEST_USER_ADMIN_EMAIL, True)
 

	
 
            self.create_user(TEST_USER_REGULAR_LOGIN, TEST_USER_REGULAR_PASS,
 
                             TEST_USER_REGULAR_EMAIL, False)
 

	
 
            self.create_user(TEST_USER_REGULAR2_LOGIN, TEST_USER_REGULAR2_PASS,
 
                             TEST_USER_REGULAR2_EMAIL, False)
 

	
 
    def create_auth_plugin_options(self, skip_existing=False):
 
        """
 
        Create default auth plugin settings, and make it active
 

	
 
        :param skip_existing:
 
        """
 

	
 
        for k, v, t in [('auth_plugins', 'kallithea.lib.auth_modules.auth_internal', 'list'),
 
                        ('auth_internal_enabled', 'True', 'bool')]:
 
            if skip_existing and Setting.get_by_name(k) is not None:
 
                log.debug('Skipping option %s', k)
 
                continue
 
            setting = Setting(k, v, t)
 
            self.sa.add(setting)
 

	
 
    def create_default_options(self, skip_existing=False):
 
        """Creates default settings"""
 

	
 
        for k, v, t in [
 
            ('default_repo_enable_downloads', False, 'bool'),
 
            ('default_repo_enable_statistics', False, 'bool'),
 
            ('default_repo_private', False, 'bool'),
 
            ('default_repo_type', 'hg', 'unicode')
 
        ]:
 
            if skip_existing and Setting.get_by_name(k) is not None:
 
                log.debug('Skipping option %s', k)
 
                continue
 
            setting = Setting(k, v, t)
 
            self.sa.add(setting)
 

	
 
    def fixup_groups(self):
 
        def_usr = User.get_default_user()
 
        for g in RepoGroup.query().all():
 
            g.group_name = g.get_new_name(g.name)
 
            # get default perm
 
            default = UserRepoGroupToPerm.query() \
 
                .filter(UserRepoGroupToPerm.group == g) \
 
                .filter(UserRepoGroupToPerm.user == def_usr) \
 
                .scalar()
 

	
 
            if default is None:
 
                log.debug('missing default permission for group %s adding', g)
 
                RepoGroupModel()._create_default_perms(g)
 

	
 
    def reset_permissions(self, username):
 
        """
 
        Resets permissions to default state, useful when old systems had
 
        bad permissions, we must clean them up
 

	
 
        :param username:
 
        """
 
        default_user = User.get_by_username(username)
 
        if not default_user:
 
            return
 

	
 
        u2p = UserToPerm.query() \
 
            .filter(UserToPerm.user == default_user).all()
 
        fixed = False
 
        if len(u2p) != len(Permission.DEFAULT_USER_PERMISSIONS):
 
            for p in u2p:
 
                Session().delete(p)
 
            fixed = True
 
            self.populate_default_permissions()
 
        return fixed
 

	
 
    def update_repo_info(self):
 
        for repo in Repository.query():
 
            repo.update_changeset_cache()
 

	
 
    def prompt_repo_root_path(self, test_repo_path='', retries=3):
 
        _path = self.cli_args.get('repos_location')
 
        if retries == 3:
 
            log.info('Setting up repositories config')
 

	
 
        if _path is not None:
 
            path = _path
 
        elif not self.tests and not test_repo_path:
 
            path = input(
 
                 'Enter a valid absolute path to store repositories. '
 
                 'All repositories in that path will be added automatically:'
 
            )
 
        else:
 
            path = test_repo_path
 
        path_ok = True
 

	
 
        # check proper dir
 
        if not os.path.isdir(path):
 
            path_ok = False
 
            log.error('Given path %s is not a valid directory', path)
 

	
 
        elif not os.path.isabs(path):
 
            path_ok = False
 
            log.error('Given path %s is not an absolute path', path)
 

	
 
        # check if path is at least readable.
 
        if not os.access(path, os.R_OK):
 
            path_ok = False
 
            log.error('Given path %s is not readable', path)
 

	
 
        # check write access, warn user about non writeable paths
 
        elif not os.access(path, os.W_OK) and path_ok:
 
            log.warning('No write permission to given path %s', path)
 
            if not self._ask_ok('Given path %s is not writeable, do you want to '
 
                          'continue with read only mode ? [y/n]' % (path,)):
 
                log.error('Canceled by user')
 
                sys.exit(-1)
 

	
 
        if retries == 0:
 
            sys.exit('max retries reached')
 
        if not path_ok:
 
            if _path is not None:
 
                sys.exit('Invalid repo path: %s' % _path)
 
            retries -= 1
 
            return self.prompt_repo_root_path(test_repo_path, retries) # recursing!!!
 

	
 
        real_path = os.path.normpath(os.path.realpath(path))
 

	
 
        if real_path != os.path.normpath(path):
 
            log.warning('Using normalized path %s instead of %s', real_path, path)
 

	
 
        return real_path
 

	
 
    def create_settings(self, repo_root_path):
 
        ui_config = [
 
            ('paths', '/', repo_root_path, True),
 
            #('phases', 'publish', 'false', False)
 
            ('hooks', Ui.HOOK_UPDATE, 'hg update >&2', False),
 
            ('hooks', Ui.HOOK_REPO_SIZE, 'python:kallithea.lib.hooks.repo_size', True),
 
            ('extensions', 'largefiles', '', True),
 
            ('largefiles', 'usercache', os.path.join(repo_root_path, '.cache', 'largefiles'), True),
 
            ('extensions', 'hgsubversion', '', False),
 
            ('extensions', 'hggit', '', False),
 
        ]
 
        for ui_section, ui_key, ui_value, ui_active in ui_config:
 
            ui_conf = Ui(
 
                ui_section=ui_section,
 
                ui_key=ui_key,
 
                ui_value=ui_value,
 
                ui_active=ui_active)
 
            self.sa.add(ui_conf)
 

	
 
        settings = [
 
            ('realm', 'Kallithea', 'unicode'),
 
            ('title', '', 'unicode'),
 
            ('ga_code', '', 'unicode'),
 
            ('show_public_icon', True, 'bool'),
 
            ('show_private_icon', True, 'bool'),
 
            ('stylify_metalabels', False, 'bool'),
 
            ('dashboard_items', 100, 'int'), # TODO: call it page_size
 
            ('admin_grid_items', 25, 'int'),
 
            ('show_version', True, 'bool'),
 
            ('use_gravatar', True, 'bool'),
 
            ('gravatar_url', User.DEFAULT_GRAVATAR_URL, 'unicode'),
 
            ('clone_uri_tmpl', Repository.DEFAULT_CLONE_URI, 'unicode'),
 
            ('clone_ssh_tmpl', Repository.DEFAULT_CLONE_SSH, 'unicode'),
 
        ]
 
        for key, val, type_ in settings:
 
            sett = Setting(key, val, type_)
 
            self.sa.add(sett)
 

	
 
        self.create_auth_plugin_options()
 
        self.create_default_options()
 

	
 
        log.info('Populated Ui and Settings defaults')
 

	
 
    def create_user(self, username, password, email='', admin=False):
 
        log.info('creating user %s', username)
 
        UserModel().create_or_update(username, password, email,
 
                                     firstname=u'Kallithea', lastname=u'Admin',
 
                                     firstname='Kallithea', lastname='Admin',
 
                                     active=True, admin=admin,
 
                                     extern_type=User.DEFAULT_AUTH_TYPE)
 

	
 
    def create_default_user(self):
 
        log.info('creating default user')
 
        # create default user for handling default permissions.
 
        user = UserModel().create_or_update(username=User.DEFAULT_USER,
 
                                            password=str(uuid.uuid1())[:20],
 
                                            email='anonymous@kallithea-scm.org',
 
                                            firstname=u'Anonymous',
 
                                            lastname=u'User')
 
                                            firstname='Anonymous',
 
                                            lastname='User')
 
        # based on configuration options activate/deactivate this user which
 
        # controls anonymous access
 
        if self.cli_args.get('public_access') is False:
 
            log.info('Public access disabled')
 
            user.active = False
 
            Session().commit()
 

	
 
    def create_permissions(self):
 
        """
 
        Creates all permissions defined in the system
 
        """
 
        # module.(access|create|change|delete)_[name]
 
        # module.(none|read|write|admin)
 
        log.info('creating permissions')
 
        PermissionModel().create_permissions()
 

	
 
    def populate_default_permissions(self):
 
        """
 
        Populate default permissions. It will create only the default
 
        permissions that are missing, and not alter already defined ones
 
        """
 
        log.info('creating default user permissions')
 
        PermissionModel().create_default_permissions(user=User.DEFAULT_USER)
kallithea/lib/feeds.py
Show inline comments
 
# -*- 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.feeds
 
~~~~~~~~~~~~~~~~~~~
 

	
 
Shared code for providing RSS and ATOM feeds.
 
"""
 

	
 
import datetime
 
import re
 

	
 
import mako.template
 

	
 

	
 
language = 'en-us'
 
ttl = "5"
 

	
 

	
 
# From ``django.utils.feedgenerator`` via webhelpers.feedgenerator
 
def rfc2822_date(date):
 
    # We do this ourselves to be timezone aware, email.Utils is not tz aware.
 
    if getattr(date, "tzinfo", False):
 
        time_str = date.strftime('%a, %d %b %Y %H:%M:%S ')
 
        offset = date.tzinfo.utcoffset(date)
 
        timezone = (offset.days * 24 * 60) + (offset.seconds / 60)
 
        hour, minute = divmod(timezone, 60)
 
        return time_str + "%+03d%02d" % (hour, minute)
 
    else:
 
        return date.strftime('%a, %d %b %Y %H:%M:%S -0000')
 

	
 
# From ``django.utils.feedgenerator`` via webhelpers.feedgenerator
 
def rfc3339_date(date):
 
    if getattr(date, "tzinfo", False):
 
        time_str = date.strftime('%Y-%m-%dT%H:%M:%S')
 
        offset = date.tzinfo.utcoffset(date)
 
        timezone = (offset.days * 24 * 60) + (offset.seconds / 60)
 
        hour, minute = divmod(timezone, 60)
 
        return time_str + "%+03d:%02d" % (hour, minute)
 
    else:
 
        return date.strftime('%Y-%m-%dT%H:%M:%SZ')
 

	
 
# From ``django.utils.feedgenerator`` via webhelpers.feedgenerator
 
def get_tag_uri(url, date):
 
    "Creates a TagURI. See http://diveintomark.org/archives/2004/05/28/howto-atom-id"
 
    tag = re.sub('^http://', '', url)
 
    if date is not None:
 
        tag = re.sub('/', ',%s:/' % date.strftime('%Y-%m-%d'), tag, 1)
 
    tag = re.sub('#', '/', tag)
 
    return u'tag:' + tag
 
    return 'tag:' + tag
 

	
 

	
 
class Attributes(object):
 
    """Simple namespace for attribute dict access in mako and elsewhere"""
 
    def __init__(self, a_dict):
 
        self.__dict__ = a_dict
 

	
 

	
 
class _Feeder(object):
 

	
 
    content_type = None
 
    template = None  # subclass must provide a mako.template.Template
 

	
 
    @classmethod
 
    def render(cls, header, entries):
 
        try:
 
            latest_pubdate = max(
 
                pubdate for pubdate in (e.get('pubdate') for e in entries)
 
                if pubdate
 
            )
 
        except ValueError:  # max() arg is an empty sequence ... or worse
 
            latest_pubdate = datetime.datetime.now()
 

	
 
        return cls.template.render(
 
            language=language,
 
            ttl=ttl,  # rss only
 
            latest_pubdate=latest_pubdate,
 
            rfc2822_date=rfc2822_date,  # for RSS
 
            rfc3339_date=rfc3339_date,  # for Atom
 
            get_tag_uri=get_tag_uri,
 
            entries=[Attributes(e) for e in entries],
 
            **header
 
        )
 

	
 

	
 
class AtomFeed(_Feeder):
 

	
 
    content_type = 'application/atom+xml'
 

	
 
    template = mako.template.Template('''\
 
<?xml version="1.0" encoding="utf-8"?>
 
<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="${language}">
 
  <title>${title}</title>
 
  <link href="${link}" rel="alternate"></link>
 
  <id>${link}</id>
 
  <updated>${rfc3339_date(latest_pubdate)}</updated>
 
  % for entry in entries:
 
  <entry>
 
    <title>${entry.title}</title>
 
    <link href="${entry.link}" rel="alternate"></link>
 
    <updated>${rfc3339_date(entry.pubdate)}</updated>
 
    <published>${rfc3339_date(entry.pubdate)}</published>
 
    <author>
 
      <name>${entry.author_name}</name>
 
      <email>${entry.author_email}</email>
 
    </author>
 
    <id>${get_tag_uri(entry.link, entry.pubdate)}</id>
 
    <summary type="html">${entry.description}</summary>
 
  </entry>
 
  % endfor
 
</feed>
 
''', default_filters=['x'], output_encoding='utf-8', encoding_errors='replace')
 

	
 

	
 
class RssFeed(_Feeder):
 

	
 
    content_type = 'application/rss+xml'
 

	
 
    template = mako.template.Template('''\
 
<?xml version="1.0" encoding="utf-8"?>
 
<rss version="2.0">
 
  <channel>
 
    <title>${title}</title>
 
    <link>${link}</link>
 
    <description>${description}</description>
 
    <language>${language}</language>
 
    <lastBuildDate>${rfc2822_date(latest_pubdate)}</lastBuildDate>
 
    <ttl>${ttl}</ttl>
 
    % for entry in entries:
 
    <item>
 
      <title>${entry.title}</title>
 
      <link>${entry.link}</link>
 
      <description>${entry.description}</description>
 
      <author>${entry.author_email} (${entry.author_name})</author>
 
      <pubDate>${rfc2822_date(entry.pubdate)}</pubDate>
 
    </item>
 
    % endfor
 
  </channel>
 
</rss>
 
''', default_filters=['x'], output_encoding='utf-8', encoding_errors='replace')
kallithea/lib/indexers/__init__.py
Show inline comments
 
# -*- 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.indexers
 
~~~~~~~~~~~~~~~~~~~~~~
 

	
 
Whoosh indexing module for 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 17, 2010
 
:author: marcink
 
:copyright: (c) 2013 RhodeCode GmbH, and others.
 
:license: GPLv3, see LICENSE.md for more details.
 
"""
 

	
 
import logging
 
import os
 
import sys
 
from os.path import dirname
 

	
 
from whoosh.analysis import IDTokenizer, LowercaseFilter, RegexTokenizer
 
from whoosh.fields import BOOLEAN, DATETIME, ID, NUMERIC, STORED, TEXT, FieldType, Schema
 
from whoosh.formats import Characters
 
from whoosh.highlight import ContextFragmenter, HtmlFormatter
 
from whoosh.highlight import highlight as whoosh_highlight
 

	
 
from kallithea.lib.utils2 import LazyProperty
 

	
 

	
 
# Add location of top level folder to sys.path
 
sys.path.append(dirname(dirname(dirname(os.path.realpath(__file__)))))
 

	
 

	
 
log = logging.getLogger(__name__)
 

	
 
# CUSTOM ANALYZER wordsplit + lowercase filter
 
ANALYZER = RegexTokenizer(expression=r"\w+") | LowercaseFilter()
 

	
 
# CUSTOM ANALYZER wordsplit + lowercase filter, for emailaddr-like text
 
#
 
# This is useful to:
 
# - avoid removing "stop words" from text
 
# - search case-insensitively
 
#
 
EMAILADDRANALYZER = RegexTokenizer() | LowercaseFilter()
 

	
 
# CUSTOM ANALYZER raw-string + lowercase filter
 
#
 
# This is useful to:
 
# - avoid tokenization
 
# - avoid removing "stop words" from text
 
# - search case-insensitively
 
#
 
ICASEIDANALYZER = IDTokenizer() | LowercaseFilter()
 

	
 
# CUSTOM ANALYZER raw-string
 
#
 
# This is useful to:
 
# - avoid tokenization
 
# - avoid removing "stop words" from text
 
#
 
IDANALYZER = IDTokenizer()
 

	
 
# CUSTOM ANALYZER wordsplit + lowercase filter, for pathname-like text
 
#
 
# This is useful to:
 
# - avoid removing "stop words" from text
 
# - search case-insensitively
 
#
 
PATHANALYZER = RegexTokenizer() | LowercaseFilter()
 

	
 
# INDEX SCHEMA DEFINITION
 
SCHEMA = Schema(
 
    fileid=ID(unique=True),
 
    owner=TEXT(analyzer=EMAILADDRANALYZER),
 
    # this field preserves case of repository name for exact matching
 
    repository_rawname=TEXT(analyzer=IDANALYZER),
 
    repository=TEXT(stored=True, analyzer=ICASEIDANALYZER),
 
    path=TEXT(stored=True, analyzer=PATHANALYZER),
 
    content=FieldType(format=Characters(), analyzer=ANALYZER,
 
                      scorable=True, stored=True),
 
    modtime=STORED(),
 
    extension=TEXT(stored=True, analyzer=PATHANALYZER)
 
)
 

	
 
IDX_NAME = 'HG_INDEX'
 
FORMATTER = HtmlFormatter('span', between='\n<span class="break">...</span>\n')
 
FRAGMENTER = ContextFragmenter(200)
 

	
 
CHGSETS_SCHEMA = Schema(
 
    raw_id=ID(unique=True, stored=True),
 
    date=NUMERIC(stored=True),
 
    last=BOOLEAN(),
 
    owner=TEXT(analyzer=EMAILADDRANALYZER),
 
    # this field preserves case of repository name for exact matching
 
    # and unique-ness in index table
 
    repository_rawname=ID(unique=True),
 
    repository=ID(stored=True, analyzer=ICASEIDANALYZER),
 
    author=TEXT(stored=True, analyzer=EMAILADDRANALYZER),
 
    message=FieldType(format=Characters(), analyzer=ANALYZER,
 
                      scorable=True, stored=True),
 
    parents=TEXT(),
 
    added=TEXT(analyzer=PATHANALYZER),
 
    removed=TEXT(analyzer=PATHANALYZER),
 
    changed=TEXT(analyzer=PATHANALYZER),
 
)
 

	
 
CHGSET_IDX_NAME = 'CHGSET_INDEX'
 

	
 
# used only to generate queries in journal
 
JOURNAL_SCHEMA = Schema(
 
    username=ID(),
 
    date=DATETIME(),
 
    action=TEXT(),
 
    repository=ID(),
 
    ip=TEXT(),
 
)
 

	
 

	
 
class WhooshResultWrapper(object):
 
    def __init__(self, search_type, searcher, matcher, highlight_items,
 
                 repo_location):
 
        self.search_type = search_type
 
        self.searcher = searcher
 
        self.matcher = matcher
 
        self.highlight_items = highlight_items
 
        self.fragment_size = 200
 
        self.repo_location = repo_location
 

	
 
    @LazyProperty
 
    def doc_ids(self):
 
        docs_id = []
 
        while self.matcher.is_active():
 
            docnum = self.matcher.id()
 
            chunks = [offsets for offsets in self.get_chunks()]
 
            docs_id.append([docnum, chunks])
 
            self.matcher.next()  # this looks like a py2 iterator ... but it isn't
 
        return docs_id
 

	
 
    def __str__(self):
 
        return '<%s at %s>' % (self.__class__.__name__, len(self.doc_ids))
 

	
 
    def __repr__(self):
 
        return self.__str__()
 

	
 
    def __len__(self):
 
        return len(self.doc_ids)
 

	
 
    def __iter__(self):
 
        """
 
        Allows Iteration over results,and lazy generate content
 

	
 
        *Requires* implementation of ``__getitem__`` method.
 
        """
 
        for docid in self.doc_ids:
 
            yield self.get_full_content(docid)
 

	
 
    def __getitem__(self, key):
 
        """
 
        Slicing of resultWrapper
 
        """
 
        i, j = key.start, key.stop
 

	
 
        slices = []
 
        for docid in self.doc_ids[i:j]:
 
            slices.append(self.get_full_content(docid))
 
        return slices
 

	
 
    def get_full_content(self, docid):
 
        res = self.searcher.stored_fields(docid[0])
 
        log.debug('result: %s', res)
 
        if self.search_type == 'content':
 
            full_repo_path = os.path.join(self.repo_location, res['repository'])
 
            f_path = res['path'].split(full_repo_path)[-1]
 
            f_path = f_path.lstrip(os.sep)
 
            content_short = self.get_short_content(res, docid[1])
 
            res.update({'content_short': content_short,
 
                        'content_short_hl': self.highlight(content_short),
 
                        'f_path': f_path
 
            })
 
        elif self.search_type == 'path':
 
            full_repo_path = os.path.join(self.repo_location, res['repository'])
 
            f_path = res['path'].split(full_repo_path)[-1]
 
            f_path = f_path.lstrip(os.sep)
 
            res.update({'f_path': f_path})
 
        elif self.search_type == 'message':
 
            res.update({'message_hl': self.highlight(res['message'])})
 

	
 
        log.debug('result: %s', res)
 

	
 
        return res
 

	
 
    def get_short_content(self, res, chunks):
 
        return u''.join([res['content'][chunk[0]:chunk[1]] for chunk in chunks])
 
        return ''.join([res['content'][chunk[0]:chunk[1]] for chunk in chunks])
 

	
 
    def get_chunks(self):
 
        """
 
        Smart function that implements chunking the content
 
        but not overlap chunks so it doesn't highlight the same
 
        close occurrences twice.
 
        """
 
        memory = [(0, 0)]
 
        try:
 
            supports_positions = self.matcher.supports('positions')
 
        except AttributeError:  # 'NoneType' object has no attribute 'supports' (because matcher never get a format)
 
            supports_positions = False
 
        if supports_positions:
 
            for span in self.matcher.spans():
 
                start = span.startchar or 0
 
                end = span.endchar or 0
 
                start_offseted = max(0, start - self.fragment_size)
 
                end_offseted = end + self.fragment_size
 

	
 
                if start_offseted < memory[-1][1]:
 
                    start_offseted = memory[-1][1]
 
                memory.append((start_offseted, end_offseted,))
 
                yield (start_offseted, end_offseted,)
 

	
 
    def highlight(self, content, top=5):
 
        if self.search_type not in ['content', 'message']:
 
            return ''
 
        hl = whoosh_highlight(
 
            text=content,
 
            terms=self.highlight_items,
 
            analyzer=ANALYZER,
 
            fragmenter=FRAGMENTER,
 
            formatter=FORMATTER,
 
            top=top
 
        )
 
        return hl
kallithea/lib/indexers/daemon.py
Show inline comments
 
# -*- 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.indexers.daemon
 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 

	
 
A daemon will read from task table and run tasks
 

	
 
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: Jan 26, 2010
 
:author: marcink
 
:copyright: (c) 2013 RhodeCode GmbH, and others.
 
:license: GPLv3, see LICENSE.md for more details.
 
"""
 

	
 

	
 
import logging
 
import os
 
import sys
 
import traceback
 
from os.path import dirname
 
from shutil import rmtree
 
from time import mktime
 

	
 
from whoosh.index import create_in, exists_in, open_dir
 
from whoosh.qparser import QueryParser
 

	
 
from kallithea.config.conf import INDEX_EXTENSIONS, INDEX_FILENAMES
 
from kallithea.lib.indexers import CHGSET_IDX_NAME, CHGSETS_SCHEMA, IDX_NAME, SCHEMA
 
from kallithea.lib.utils2 import safe_str
 
from kallithea.lib.vcs.exceptions import ChangesetError, NodeDoesNotExistError, RepositoryError
 
from kallithea.model.db import Repository
 
from kallithea.model.scm import ScmModel
 

	
 

	
 
# Add location of top level folder to sys.path
 
project_path = dirname(dirname(dirname(dirname(os.path.realpath(__file__)))))
 
sys.path.append(project_path)
 

	
 

	
 

	
 

	
 
log = logging.getLogger('whoosh_indexer')
 

	
 

	
 
class WhooshIndexingDaemon(object):
 
    """
 
    Daemon for atomic indexing jobs
 
    """
 

	
 
    def __init__(self, indexname=IDX_NAME, index_location=None,
 
                 repo_location=None, repo_list=None,
 
                 repo_update_list=None):
 
        self.indexname = indexname
 

	
 
        self.index_location = index_location
 
        if not index_location:
 
            raise Exception('You have to provide index location')
 

	
 
        self.repo_location = repo_location
 
        if not repo_location:
 
            raise Exception('You have to provide repositories location')
 

	
 
        self.repo_paths = ScmModel().repo_scan(self.repo_location)
 

	
 
        # filter repo list
 
        if repo_list:
 
            repo_list = set(repo_list)
 
            self.filtered_repo_paths = {}
 
            for repo_name, repo in self.repo_paths.items():
 
                if repo_name in repo_list:
 
                    self.filtered_repo_paths[repo_name] = repo
 

	
 
            self.repo_paths = self.filtered_repo_paths
 

	
 
        # filter update repo list
 
        self.filtered_repo_update_paths = {}
 
        if repo_update_list:
 
            self.filtered_repo_update_paths = {}
 
            for repo_name, repo in self.repo_paths.items():
 
                if repo_name in repo_update_list:
 
                    self.filtered_repo_update_paths[repo_name] = repo
 
            self.repo_paths = self.filtered_repo_update_paths
 

	
 
        self.initial = True
 
        if not os.path.isdir(self.index_location):
 
            os.makedirs(self.index_location)
 
            log.info('Cannot run incremental index since it does not '
 
                     'yet exist - running full build')
 
        elif not exists_in(self.index_location, IDX_NAME):
 
            log.info('Running full index build, as the file content '
 
                     'index does not exist')
 
        elif not exists_in(self.index_location, CHGSET_IDX_NAME):
 
            log.info('Running full index build, as the changeset '
 
                     'index does not exist')
 
        else:
 
            self.initial = False
 

	
 
    def _get_index_revision(self, repo):
 
        db_repo = Repository.get_by_repo_name(repo.name)
 
        landing_rev = 'tip'
 
        if db_repo:
 
            _rev_type, _rev = db_repo.landing_rev
 
            landing_rev = _rev
 
        return landing_rev
 

	
 
    def _get_index_changeset(self, repo, index_rev=None):
 
        if not index_rev:
 
            index_rev = self._get_index_revision(repo)
 
        cs = repo.get_changeset(index_rev)
 
        return cs
 

	
 
    def get_paths(self, repo):
 
        """
 
        recursive walk in root dir and return a set of all path in that dir
 
        based on repository walk function
 
        """
 
        index_paths_ = set()
 
        try:
 
            cs = self._get_index_changeset(repo)
 
            for _topnode, _dirs, files in cs.walk('/'):
 
                for f in files:
 
                    index_paths_.add(os.path.join(repo.path, f.path))
 

	
 
        except RepositoryError:
 
            log.debug(traceback.format_exc())
 
            pass
 
        return index_paths_
 

	
 
    def get_node(self, repo, path, index_rev=None):
 
        """
 
        gets a filenode based on given full path.
 

	
 
        :param repo: scm repo instance
 
        :param path: full path including root location
 
        :return: FileNode
 
        """
 
        # FIXME: paths should be normalized ... or even better: don't include repo.path
 
        assert path.startswith(repo.path)
 
        assert path[len(repo.path)] in (os.path.sep, os.path.altsep)
 
        node_path = path[len(repo.path) + 1:]
 
        cs = self._get_index_changeset(repo, index_rev=index_rev)
 
        node = cs.get_node(node_path)
 
        return node
 

	
 
    def is_indexable_node(self, node):
 
        """
 
        Just index the content of chosen files, skipping binary files
 
        """
 
        return (node.extension in INDEX_EXTENSIONS or node.name in INDEX_FILENAMES) and \
 
               not node.is_binary
 

	
 
    def get_node_mtime(self, node):
 
        return mktime(node.last_changeset.date.timetuple())
 

	
 
    def add_doc(self, writer, path, repo, repo_name, index_rev=None):
 
        """
 
        Adding doc to writer this function itself fetches data from
 
        the instance of vcs backend
 
        """
 
        try:
 
            node = self.get_node(repo, path, index_rev)
 
        except (ChangesetError, NodeDoesNotExistError):
 
            log.debug("    >> %s - not found in %s %s", path, repo, index_rev)
 
            return 0, 0
 

	
 
        indexed = indexed_w_content = 0
 
        if self.is_indexable_node(node):
 
            bytes_content = node.content
 
            if b'\0' in bytes_content:
 
                log.warning('    >> %s - no text content', path)
 
                u_content = u''
 
                u_content = ''
 
            else:
 
                log.debug('    >> %s', path)
 
                u_content = safe_str(bytes_content)
 
                indexed_w_content += 1
 

	
 
        else:
 
            log.debug('    >> %s - not indexable', path)
 
            # just index file name without it's content
 
            u_content = u''
 
            u_content = ''
 
            indexed += 1
 

	
 
        writer.add_document(
 
            fileid=path,
 
            owner=repo.contact,
 
            repository_rawname=repo_name,
 
            repository=repo_name,
 
            path=path,
 
            content=u_content,
 
            modtime=self.get_node_mtime(node),
 
            extension=node.extension
 
        )
 
        return indexed, indexed_w_content
 

	
 
    def index_changesets(self, writer, repo_name, repo, start_rev=None):
 
        """
 
        Add all changeset in the vcs repo starting at start_rev
 
        to the index writer
 

	
 
        :param writer: the whoosh index writer to add to
 
        :param repo_name: name of the repository from whence the
 
          changeset originates including the repository group
 
        :param repo: the vcs repository instance to index changesets for,
 
          the presumption is the repo has changesets to index
 
        :param start_rev=None: the full sha id to start indexing from
 
          if start_rev is None then index from the first changeset in
 
          the repo
 
        """
 

	
 
        if start_rev is None:
 
            start_rev = repo[0].raw_id
 

	
 
        log.debug('Indexing changesets in %s, starting at rev %s',
 
                  repo_name, start_rev)
 

	
 
        indexed = 0
 
        cs_iter = repo.get_changesets(start=start_rev)
 
        total = len(cs_iter)
 
        for cs in cs_iter:
 
            indexed += 1
 
            log.debug('    >> %s %s/%s', cs, indexed, total)
 
            writer.add_document(
 
                raw_id=cs.raw_id,
 
                owner=repo.contact,
 
                date=cs._timestamp,
 
                repository_rawname=repo_name,
 
                repository=repo_name,
 
                author=cs.author,
 
                message=cs.message,
 
                last=cs.last,
 
                added=u' '.join(node.path for node in cs.added).lower(),
 
                removed=u' '.join(node.path for node in cs.removed).lower(),
 
                changed=u' '.join(node.path for node in cs.changed).lower(),
 
                parents=u' '.join(cs.raw_id for cs in cs.parents),
 
                added=' '.join(node.path for node in cs.added).lower(),
 
                removed=' '.join(node.path for node in cs.removed).lower(),
 
                changed=' '.join(node.path for node in cs.changed).lower(),
 
                parents=' '.join(cs.raw_id for cs in cs.parents),
 
            )
 

	
 
        return indexed
 

	
 
    def index_files(self, file_idx_writer, repo_name, repo):
 
        """
 
        Index files for given repo_name
 

	
 
        :param file_idx_writer: the whoosh index writer to add to
 
        :param repo_name: name of the repository we're indexing
 
        :param repo: instance of vcs repo
 
        """
 
        i_cnt = iwc_cnt = 0
 
        log.debug('Building file index for %s @revision:%s', repo_name,
 
                                                self._get_index_revision(repo))
 
        index_rev = self._get_index_revision(repo)
 
        for idx_path in self.get_paths(repo):
 
            i, iwc = self.add_doc(file_idx_writer, idx_path, repo, repo_name, index_rev)
 
            i_cnt += i
 
            iwc_cnt += iwc
 

	
 
        log.debug('added %s files %s with content for repo %s',
 
                  i_cnt + iwc_cnt, iwc_cnt, repo.path)
 
        return i_cnt, iwc_cnt
 

	
 
    def update_changeset_index(self):
 
        idx = open_dir(self.index_location, indexname=CHGSET_IDX_NAME)
 

	
 
        with idx.searcher() as searcher:
 
            writer = idx.writer()
 
            writer_is_dirty = False
 
            try:
 
                indexed_total = 0
 
                repo_name = None
 
                for repo_name, repo in sorted(self.repo_paths.items()):
 
                    log.debug('Updating changeset index for repo %s', repo_name)
 
                    # skip indexing if there aren't any revs in the repo
 
                    num_of_revs = len(repo)
 
                    if num_of_revs < 1:
 
                        continue
 

	
 
                    qp = QueryParser('repository', schema=CHGSETS_SCHEMA)
 
                    q = qp.parse(u"last:t AND %s" % repo_name)
 
                    q = qp.parse("last:t AND %s" % repo_name)
 

	
 
                    results = searcher.search(q)
 

	
 
                    # default to scanning the entire repo
 
                    last_rev = 0
 
                    start_id = None
 

	
 
                    if len(results) > 0:
 
                        # assuming that there is only one result, if not this
 
                        # may require a full re-index.
 
                        start_id = results[0]['raw_id']
 
                        last_rev = repo.get_changeset(revision=start_id).revision
 

	
 
                    # there are new changesets to index or a new repo to index
 
                    if last_rev == 0 or num_of_revs > last_rev + 1:
 
                        # delete the docs in the index for the previous
 
                        # last changeset(s)
 
                        for hit in results:
 
                            q = qp.parse(u"last:t AND %s AND raw_id:%s" %
 
                            q = qp.parse("last:t AND %s AND raw_id:%s" %
 
                                            (repo_name, hit['raw_id']))
 
                            writer.delete_by_query(q)
 

	
 
                        # index from the previous last changeset + all new ones
 
                        indexed_total += self.index_changesets(writer,
 
                                                repo_name, repo, start_id)
 
                        writer_is_dirty = True
 
                log.debug('indexed %s changesets for repo %s',
 
                          indexed_total, repo_name
 
                )
 
            finally:
 
                if writer_is_dirty:
 
                    log.debug('>> COMMITING CHANGES TO CHANGESET INDEX<<')
 
                    writer.commit(merge=True)
 
                    log.debug('>>> FINISHED REBUILDING CHANGESET INDEX <<<')
 
                else:
 
                    log.debug('>> NOTHING TO COMMIT TO CHANGESET INDEX<<')
 

	
 
    def update_file_index(self):
 
        log.debug(u'STARTING INCREMENTAL INDEXING UPDATE FOR EXTENSIONS %s '
 
        log.debug('STARTING INCREMENTAL INDEXING UPDATE FOR EXTENSIONS %s '
 
                  'AND REPOS %s', INDEX_EXTENSIONS, ' and '.join(self.repo_paths))
 

	
 
        idx = open_dir(self.index_location, indexname=self.indexname)
 
        # The set of all paths in the index
 
        indexed_paths = set()
 
        # The set of all paths we need to re-index
 
        to_index = set()
 

	
 
        writer = idx.writer()
 
        writer_is_dirty = False
 
        try:
 
            with idx.reader() as reader:
 

	
 
                # Loop over the stored fields in the index
 
                for fields in reader.all_stored_fields():
 
                    indexed_path = fields['path']
 
                    indexed_repo_path = fields['repository']
 
                    indexed_paths.add(indexed_path)
 

	
 
                    if indexed_repo_path not in self.filtered_repo_update_paths:
 
                        continue
 

	
 
                    repo = self.repo_paths[indexed_repo_path]
 

	
 
                    try:
 
                        node = self.get_node(repo, indexed_path)
 
                        # Check if this file was changed since it was indexed
 
                        indexed_time = fields['modtime']
 
                        mtime = self.get_node_mtime(node)
 
                        if mtime > indexed_time:
 
                            # The file has changed, delete it and add it to
 
                            # the list of files to reindex
 
                            log.debug(
 
                                'adding to reindex list %s mtime: %s vs %s',
 
                                    indexed_path, mtime, indexed_time
 
                            )
 
                            writer.delete_by_term('fileid', indexed_path)
 
                            writer_is_dirty = True
 

	
 
                            to_index.add(indexed_path)
 
                    except (ChangesetError, NodeDoesNotExistError):
 
                        # This file was deleted since it was indexed
 
                        log.debug('removing from index %s', indexed_path)
 
                        writer.delete_by_term('path', indexed_path)
 
                        writer_is_dirty = True
 

	
 
            # Loop over the files in the filesystem
 
            # Assume we have a function that gathers the filenames of the
 
            # documents to be indexed
 
            ri_cnt_total = 0  # indexed
 
            riwc_cnt_total = 0  # indexed with content
 
            for repo_name, repo in sorted(self.repo_paths.items()):
 
                log.debug('Updating file index for repo %s', repo_name)
 
                # skip indexing if there aren't any revisions
 
                if len(repo) < 1:
 
                    continue
 
                ri_cnt = 0   # indexed
 
                riwc_cnt = 0  # indexed with content
 
                for path in self.get_paths(repo):
 
                    if path in to_index or path not in indexed_paths:
 
                        # This is either a file that's changed, or a new file
 
                        # that wasn't indexed before. So index it!
 
                        i, iwc = self.add_doc(writer, path, repo, repo_name)
 
                        writer_is_dirty = True
 
                        ri_cnt += i
 
                        ri_cnt_total += 1
 
                        riwc_cnt += iwc
 
                        riwc_cnt_total += iwc
 
                log.debug('added %s files %s with content for repo %s',
 
                             ri_cnt + riwc_cnt, riwc_cnt, repo.path
 
                )
 
            log.debug('indexed %s files in total and %s with content',
 
                        ri_cnt_total, riwc_cnt_total
 
            )
 
        finally:
 
            if writer_is_dirty:
 
                log.debug('>> COMMITING CHANGES TO FILE INDEX <<')
 
                writer.commit(merge=True)
 
                log.debug('>>> FINISHED REBUILDING FILE INDEX <<<')
 
            else:
 
                log.debug('>> NOTHING TO COMMIT TO FILE INDEX <<')
 
                writer.cancel()
 

	
 
    def build_indexes(self):
 
        if os.path.exists(self.index_location):
 
            log.debug('removing previous index')
 
            rmtree(self.index_location)
 

	
 
        if not os.path.exists(self.index_location):
 
            os.mkdir(self.index_location)
 

	
 
        chgset_idx = create_in(self.index_location, CHGSETS_SCHEMA,
 
                               indexname=CHGSET_IDX_NAME)
 
        chgset_idx_writer = chgset_idx.writer()
 

	
 
        file_idx = create_in(self.index_location, SCHEMA, indexname=IDX_NAME)
 
        file_idx_writer = file_idx.writer()
 
        log.debug('BUILDING INDEX FOR EXTENSIONS %s '
 
                  'AND REPOS %s', INDEX_EXTENSIONS, ' and '.join(self.repo_paths))
 

	
 
        for repo_name, repo in sorted(self.repo_paths.items()):
 
            log.debug('Updating indices for repo %s', repo_name)
 
            # skip indexing if there aren't any revisions
 
            if len(repo) < 1:
 
                continue
 

	
 
            self.index_files(file_idx_writer, repo_name, repo)
 
            self.index_changesets(chgset_idx_writer, repo_name, repo)
 

	
 
        log.debug('>> COMMITING CHANGES <<')
 
        file_idx_writer.commit(merge=True)
 
        chgset_idx_writer.commit(merge=True)
 
        log.debug('>>> FINISHED BUILDING INDEX <<<')
 

	
 
    def update_indexes(self):
 
        self.update_file_index()
 
        self.update_changeset_index()
 

	
 
    def run(self, full_index=False):
 
        """Run daemon"""
 
        if full_index or self.initial:
 
            self.build_indexes()
 
        else:
 
            self.update_indexes()
kallithea/lib/locale.py
Show inline comments
 
# -*- 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/>.
 
import logging
 
import os
 
import sys
 

	
 

	
 
log = logging.getLogger(__name__)
 

	
 
def current_locale_is_valid():
 
    """Verify that things work when Dulwich passes unicode paths to the file system layer.
 

	
 
    Note: UTF-8 is preferred, but for example ISO-8859-1 or mbcs should also
 
    work under the right circumstances."""
 
    try:
 
        u'\xe9'.encode(sys.getfilesystemencoding()) # Test using é (&eacute;)
 
        '\xe9'.encode(sys.getfilesystemencoding()) # Test using é (&eacute;)
 
    except UnicodeEncodeError:
 
        log.error("Cannot encode Unicode paths to file system encoding %r", sys.getfilesystemencoding())
 
        for var in ['LC_ALL', 'LC_CTYPE', 'LANG']:
 
            if var in os.environ:
 
                val = os.environ[var]
 
                log.error("Note: Environment variable %s is %r - perhaps change it to some other value from 'locale -a', like 'C.UTF-8' or 'en_US.UTF-8'", var, val)
 
                break
 
        else:
 
            log.error("Note: No locale setting found in environment variables - perhaps set LC_CTYPE to some value from 'locale -a', like 'C.UTF-8' or 'en_US.UTF-8'")
 
        return False
 
    return True
 

	
 
def get_current_locale():
 
    """Return the current locale based on environment variables.
 
    There does not seem to be a good (and functional) way to get it via Python.
 
    """
 
    for var in ['LC_ALL', 'LC_CTYPE', 'LANG']:
 
        val = os.environ.get(var)
 
        if val:
 
            log.debug('Determined current locale via environment variable %s (%s)', var, val)
 
            return val
 
    return None
kallithea/lib/middleware/pygrack.py
Show inline comments
 
# -*- 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.middleware.pygrack
 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 

	
 
Python implementation of git-http-backend's Smart HTTP protocol
 

	
 
Based on original code from git_http_backend.py project.
 

	
 
Copyright (c) 2010 Daniel Dotsenko <dotsa@hotmail.com>
 
Copyright (c) 2012 Marcin Kuzminski <marcin@python-works.com>
 

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

	
 
import logging
 
import os
 
import socket
 
import traceback
 

	
 
from webob import Request, Response, exc
 

	
 
import kallithea
 
from kallithea.lib.utils2 import ascii_bytes
 
from kallithea.lib.vcs import subprocessio
 

	
 

	
 
log = logging.getLogger(__name__)
 

	
 

	
 
class FileWrapper(object):
 

	
 
    def __init__(self, fd, content_length):
 
        self.fd = fd
 
        self.content_length = content_length
 
        self.remain = content_length
 

	
 
    def read(self, size):
 
        if size <= self.remain:
 
            try:
 
                data = self.fd.read(size)
 
            except socket.error:
 
                raise IOError(self)
 
            self.remain -= size
 
        elif self.remain:
 
            data = self.fd.read(self.remain)
 
            self.remain = 0
 
        else:
 
            data = None
 
        return data
 

	
 
    def __repr__(self):
 
        return '<FileWrapper %s len: %s, read: %s>' % (
 
            self.fd, self.content_length, self.content_length - self.remain
 
        )
 

	
 

	
 
class GitRepository(object):
 
    git_folder_signature = set(['config', 'head', 'info', 'objects', 'refs'])
 
    commands = ['git-upload-pack', 'git-receive-pack']
 

	
 
    def __init__(self, repo_name, content_path):
 
        files = set([f.lower() for f in os.listdir(content_path)])
 
        if not (self.git_folder_signature.intersection(files)
 
                == self.git_folder_signature):
 
            raise OSError('%s missing git signature' % content_path)
 
        self.content_path = content_path
 
        self.valid_accepts = ['application/x-%s-result' %
 
                              c for c in self.commands]
 
        self.repo_name = repo_name
 

	
 
    def _get_fixedpath(self, path):
 
        """
 
        Small fix for repo_path
 

	
 
        :param path:
 
        """
 
        assert path.startswith('/' + self.repo_name + '/')
 
        return path[len(self.repo_name) + 2:].strip('/')
 

	
 
    def inforefs(self, req, environ):
 
        """
 
        WSGI Response producer for HTTP GET Git Smart
 
        HTTP /info/refs request.
 
        """
 

	
 
        git_command = req.GET.get('service')
 
        if git_command not in self.commands:
 
            log.debug('command %s not allowed', git_command)
 
            return exc.HTTPMethodNotAllowed()
 

	
 
        # From Documentation/technical/http-protocol.txt shipped with Git:
 
        #
 
        # Clients MUST verify the first pkt-line is `# service=$servicename`.
 
        # Servers MUST set $servicename to be the request parameter value.
 
        # Servers SHOULD include an LF at the end of this line.
 
        # Clients MUST ignore an LF at the end of the line.
 
        #
 
        #  smart_reply     =  PKT-LINE("# service=$servicename" LF)
 
        #                     ref_list
 
        #                     "0000"
 
        server_advert = '# service=%s\n' % git_command
 
        packet_len = hex(len(server_advert) + 4)[2:].rjust(4, '0').lower()
 
        _git_path = kallithea.CONFIG.get('git_path', 'git')
 
        cmd = [_git_path, git_command[4:],
 
               '--stateless-rpc', '--advertise-refs', self.content_path]
 
        log.debug('handling cmd %s', cmd)
 
        try:
 
            out = subprocessio.SubprocessIOChunker(cmd,
 
                starting_values=[ascii_bytes(packet_len + server_advert + '0000')]
 
            )
 
        except EnvironmentError as e:
 
            log.error(traceback.format_exc())
 
            raise exc.HTTPExpectationFailed()
 
        resp = Response()
 
        resp.content_type = 'application/x-%s-advertisement' % str(git_command)
 
        resp.charset = None
 
        resp.app_iter = out
 
        return resp
 

	
 
    def backend(self, req, environ):
 
        """
 
        WSGI Response producer for HTTP POST Git Smart HTTP requests.
 
        Reads commands and data from HTTP POST's body.
 
        returns an iterator obj with contents of git command's
 
        response to stdout
 
        """
 
        _git_path = kallithea.CONFIG.get('git_path', 'git')
 
        git_command = self._get_fixedpath(req.path_info)
 
        if git_command not in self.commands:
 
            log.debug('command %s not allowed', git_command)
 
            return exc.HTTPMethodNotAllowed()
 

	
 
        if 'CONTENT_LENGTH' in environ:
 
            inputstream = FileWrapper(environ['wsgi.input'],
 
                                      req.content_length)
 
        else:
 
            inputstream = environ['wsgi.input']
 

	
 
        gitenv = dict(os.environ)
 
        # forget all configs
 
        gitenv['GIT_CONFIG_NOGLOBAL'] = '1'
 
        cmd = [_git_path, git_command[4:], '--stateless-rpc', self.content_path]
 
        log.debug('handling cmd %s', cmd)
 
        try:
 
            out = subprocessio.SubprocessIOChunker(
 
                cmd,
 
                inputstream=inputstream,
 
                env=gitenv,
 
                cwd=self.content_path,
 
            )
 
        except EnvironmentError as e:
 
            log.error(traceback.format_exc())
 
            raise exc.HTTPExpectationFailed()
 

	
 
        if git_command in [u'git-receive-pack']:
 
        if git_command in ['git-receive-pack']:
 
            # updating refs manually after each push.
 
            # Needed for pre-1.7.0.4 git clients using regular HTTP mode.
 
            from kallithea.lib.vcs import get_repo
 
            from dulwich.server import update_server_info
 
            repo = get_repo(self.content_path)
 
            if repo:
 
                update_server_info(repo._repo)
 

	
 
        resp = Response()
 
        resp.content_type = 'application/x-%s-result' % git_command.encode('utf-8')
 
        resp.charset = None
 
        resp.app_iter = out
 
        return resp
 

	
 
    def __call__(self, environ, start_response):
 
        req = Request(environ)
 
        _path = self._get_fixedpath(req.path_info)
 
        if _path.startswith('info/refs'):
 
            app = self.inforefs
 
        elif req.accept.acceptable_offers(self.valid_accepts):
 
            app = self.backend
 
        try:
 
            resp = app(req, environ)
 
        except exc.HTTPException as e:
 
            resp = e
 
            log.error(traceback.format_exc())
 
        except Exception as e:
 
            log.error(traceback.format_exc())
 
            resp = exc.HTTPInternalServerError()
 
        return resp(environ, start_response)
 

	
 

	
 
class GitDirectory(object):
 

	
 
    def __init__(self, repo_root, repo_name):
 
        repo_location = os.path.join(repo_root, repo_name)
 
        if not os.path.isdir(repo_location):
 
            raise OSError(repo_location)
 

	
 
        self.content_path = repo_location
 
        self.repo_name = repo_name
 
        self.repo_location = repo_location
 

	
 
    def __call__(self, environ, start_response):
 
        content_path = self.content_path
 
        try:
 
            app = GitRepository(self.repo_name, content_path)
 
        except (AssertionError, OSError):
 
            content_path = os.path.join(content_path, '.git')
 
            if os.path.isdir(content_path):
 
                app = GitRepository(self.repo_name, content_path)
 
            else:
 
                return exc.HTTPNotFound()(environ, start_response)
 
        return app(environ, start_response)
 

	
 

	
 
def make_wsgi_app(repo_name, repo_root):
 
    from dulwich.web import LimitedInputFilter, GunzipFilter
 
    app = GitDirectory(repo_root, repo_name)
 
    return GunzipFilter(LimitedInputFilter(app))
kallithea/lib/utils.py
Show inline comments
 
# -*- 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.utils
 
~~~~~~~~~~~~~~~~~~~
 

	
 
Utilities library for 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: Apr 18, 2010
 
:author: marcink
 
:copyright: (c) 2013 RhodeCode GmbH, and others.
 
:license: GPLv3, see LICENSE.md for more details.
 
"""
 

	
 
import datetime
 
import logging
 
import os
 
import re
 
import sys
 
import traceback
 
from distutils.version import StrictVersion
 

	
 
import beaker.cache
 
import mercurial.config
 
import mercurial.ui
 
from tg.i18n import ugettext as _
 

	
 
import kallithea.config.conf
 
from kallithea.lib.exceptions import HgsubversionImportError
 
from kallithea.lib.utils2 import ascii_bytes, aslist, get_current_authuser, safe_bytes, safe_str
 
from kallithea.lib.vcs.backends.git.repository import GitRepository
 
from kallithea.lib.vcs.backends.hg.repository import MercurialRepository
 
from kallithea.lib.vcs.conf import settings
 
from kallithea.lib.vcs.exceptions import RepositoryError, VCSError
 
from kallithea.lib.vcs.utils.fakemod import create_module
 
from kallithea.lib.vcs.utils.helpers import get_scm
 
from kallithea.model import meta
 
from kallithea.model.db import RepoGroup, Repository, Setting, Ui, User, UserGroup, UserLog
 

	
 

	
 
log = logging.getLogger(__name__)
 

	
 
REMOVED_REPO_PAT = re.compile(r'rm__\d{8}_\d{6}_\d{6}_.*')
 

	
 

	
 
#==============================================================================
 
# PERM DECORATOR HELPERS FOR EXTRACTING NAMES FOR PERM CHECKS
 
#==============================================================================
 
def get_repo_slug(request):
 
    _repo = request.environ['pylons.routes_dict'].get('repo_name')
 
    if _repo:
 
        _repo = _repo.rstrip('/')
 
    return _repo
 

	
 

	
 
def get_repo_group_slug(request):
 
    _group = request.environ['pylons.routes_dict'].get('group_name')
 
    if _group:
 
        _group = _group.rstrip('/')
 
    return _group
 

	
 

	
 
def get_user_group_slug(request):
 
    _group = request.environ['pylons.routes_dict'].get('id')
 
    _group = UserGroup.get(_group)
 
    if _group:
 
        return _group.users_group_name
 
    return None
 

	
 

	
 
def _get_permanent_id(s):
 
    """Helper for decoding stable URLs with repo ID. For a string like '_123'
 
    return 123.
 
    """
 
    by_id_match = re.match(r'^_(\d+)$', s)
 
    if by_id_match is None:
 
        return None
 
    return int(by_id_match.group(1))
 

	
 

	
 
def fix_repo_id_name(path):
 
    """
 
    Rewrite repo_name for _<ID> permanent URLs.
 

	
 
    Given a path, if the first path element is like _<ID>, return the path with
 
    this part expanded to the corresponding full repo name, else return the
 
    provided path.
 
    """
 
    first, rest = path, ''
 
    if '/' in path:
 
        first, rest_ = path.split('/', 1)
 
        rest = '/' + rest_
 
    repo_id = _get_permanent_id(first)
 
    if repo_id is not None:
 
        repo = Repository.get(repo_id)
 
        if repo is not None:
 
            return repo.repo_name + rest
 
    return path
 

	
 

	
 
def action_logger(user, action, repo, ipaddr='', commit=False):
 
    """
 
    Action logger for various actions made by users
 

	
 
    :param user: user that made this action, can be a unique username string or
 
        object containing user_id attribute
 
    :param action: action to log, should be on of predefined unique actions for
 
        easy translations
 
    :param repo: string name of repository or object containing repo_id,
 
        that action was made on
 
    :param ipaddr: optional IP address from what the action was made
 

	
 
    """
 

	
 
    # if we don't get explicit IP address try to get one from registered user
 
    # in tmpl context var
 
    if not ipaddr:
 
        ipaddr = getattr(get_current_authuser(), 'ip_addr', '')
 

	
 
    if getattr(user, 'user_id', None):
 
        user_obj = User.get(user.user_id)
 
    elif isinstance(user, str):
 
        user_obj = User.get_by_username(user)
 
    else:
 
        raise Exception('You have to provide a user object or a username')
 

	
 
    if getattr(repo, 'repo_id', None):
 
        repo_obj = Repository.get(repo.repo_id)
 
        repo_name = repo_obj.repo_name
 
    elif isinstance(repo, str):
 
        repo_name = repo.lstrip('/')
 
        repo_obj = Repository.get_by_repo_name(repo_name)
 
    else:
 
        repo_obj = None
 
        repo_name = u''
 
        repo_name = ''
 

	
 
    user_log = UserLog()
 
    user_log.user_id = user_obj.user_id
 
    user_log.username = user_obj.username
 
    user_log.action = action
 

	
 
    user_log.repository = repo_obj
 
    user_log.repository_name = repo_name
 

	
 
    user_log.action_date = datetime.datetime.now()
 
    user_log.user_ip = ipaddr
 
    meta.Session().add(user_log)
 

	
 
    log.info('Logging action:%s on %s by user:%s ip:%s',
 
             action, repo, user_obj, ipaddr)
 
    if commit:
 
        meta.Session().commit()
 

	
 

	
 
def get_filesystem_repos(path):
 
    """
 
    Scans given path for repos and return (name,(type,path)) tuple
 

	
 
    :param path: path to scan for repositories
 
    :param recursive: recursive search and return names with subdirs in front
 
    """
 

	
 
    # remove ending slash for better results
 
    path = path.rstrip(os.sep)
 
    log.debug('now scanning in %s', path)
 

	
 
    def isdir(*n):
 
        return os.path.isdir(os.path.join(*n))
 

	
 
    for root, dirs, _files in os.walk(path):
 
        recurse_dirs = []
 
        for subdir in dirs:
 
            # skip removed repos
 
            if REMOVED_REPO_PAT.match(subdir):
 
                continue
 

	
 
            # skip .<something> dirs TODO: rly? then we should prevent creating them ...
 
            if subdir.startswith('.'):
 
                continue
 

	
 
            cur_path = os.path.join(root, subdir)
 
            if isdir(cur_path, '.git'):
 
                log.warning('ignoring non-bare Git repo: %s', cur_path)
 
                continue
 

	
 
            if (isdir(cur_path, '.hg') or
 
                isdir(cur_path, '.svn') or
 
                isdir(cur_path, 'objects') and (isdir(cur_path, 'refs') or
 
                                                os.path.isfile(os.path.join(cur_path, 'packed-refs')))):
 

	
 
                if not os.access(cur_path, os.R_OK) or not os.access(cur_path, os.X_OK):
 
                    log.warning('ignoring repo path without access: %s', cur_path)
 
                    continue
 

	
 
                if not os.access(cur_path, os.W_OK):
 
                    log.warning('repo path without write access: %s', cur_path)
 

	
 
                try:
 
                    scm_info = get_scm(cur_path)
 
                    assert cur_path.startswith(path)
 
                    repo_path = cur_path[len(path) + 1:]
 
                    yield repo_path, scm_info
 
                    continue # no recursion
 
                except VCSError:
 
                    # We should perhaps ignore such broken repos, but especially
 
                    # the bare git detection is unreliable so we dive into it
 
                    pass
 

	
 
            recurse_dirs.append(subdir)
 

	
 
        dirs[:] = recurse_dirs
 

	
 

	
 
def is_valid_repo_uri(repo_type, url, ui):
 
    """Check if the url seems like a valid remote repo location - raise an Exception if any problems"""
 
    if repo_type == 'hg':
 
        if url.startswith('http') or url.startswith('ssh'):
 
            # initially check if it's at least the proper URL
 
            # or does it pass basic auth
 
            MercurialRepository._check_url(url, ui)
 
        elif url.startswith('svn+http'):
 
            try:
 
                from hgsubversion.svnrepo import svnremoterepo
 
            except ImportError:
 
                raise HgsubversionImportError(_('Unable to activate hgsubversion support. '
 
                                                'The "hgsubversion" library is missing'))
 
            svnremoterepo(ui, url).svn.uuid
 
        elif url.startswith('git+http'):
 
            raise NotImplementedError()
 
        else:
 
            raise Exception('URI %s not allowed' % (url,))
 

	
 
    elif repo_type == 'git':
 
        if url.startswith('http') or url.startswith('git'):
 
            # initially check if it's at least the proper URL
 
            # or does it pass basic auth
 
            GitRepository._check_url(url)
 
        elif url.startswith('svn+http'):
 
            raise NotImplementedError()
 
        elif url.startswith('hg+http'):
 
            raise NotImplementedError()
 
        else:
 
            raise Exception('URI %s not allowed' % (url))
 

	
 

	
 
def is_valid_repo(repo_name, base_path, scm=None):
 
    """
 
    Returns True if given path is a valid repository False otherwise.
 
    If scm param is given also compare if given scm is the same as expected
 
    from scm parameter
 

	
 
    :param repo_name:
 
    :param base_path:
 
    :param scm:
 

	
 
    :return True: if given path is a valid repository
 
    """
 
    # TODO: paranoid security checks?
 
    full_path = os.path.join(base_path, repo_name)
 

	
 
    try:
 
        scm_ = get_scm(full_path)
 
        if scm:
 
            return scm_[0] == scm
 
        return True
 
    except VCSError:
 
        return False
 

	
 

	
 
def is_valid_repo_group(repo_group_name, base_path, skip_path_check=False):
 
    """
 
    Returns True if given path is a repository group False otherwise
 

	
 
    :param repo_name:
 
    :param base_path:
 
    """
 
    full_path = os.path.join(base_path, repo_group_name)
 

	
 
    # check if it's not a repo
 
    if is_valid_repo(repo_group_name, base_path):
 
        return False
 

	
 
    try:
 
        # we need to check bare git repos at higher level
 
        # since we might match branches/hooks/info/objects or possible
 
        # other things inside bare git repo
 
        get_scm(os.path.dirname(full_path))
 
        return False
 
    except VCSError:
 
        pass
 

	
 
    # check if it's a valid path
 
    if skip_path_check or os.path.isdir(full_path):
 
        return True
 

	
 
    return False
 

	
 

	
 
# propagated from mercurial documentation
 
ui_sections = ['alias', 'auth',
 
                'decode/encode', 'defaults',
 
                'diff', 'email',
 
                'extensions', 'format',
 
                'merge-patterns', 'merge-tools',
 
                'hooks', 'http_proxy',
 
                'smtp', 'patch',
 
                'paths', 'profiling',
 
                'server', 'trusted',
 
                'ui', 'web', ]
 

	
 

	
 
def make_ui(repo_path=None):
 
    """
 
    Create an Mercurial 'ui' object based on database Ui settings, possibly
 
    augmenting with content from a hgrc file.
 
    """
 
    baseui = mercurial.ui.ui()
 

	
 
    # clean the baseui object
 
    baseui._ocfg = mercurial.config.config()
 
    baseui._ucfg = mercurial.config.config()
 
    baseui._tcfg = mercurial.config.config()
 

	
 
    sa = meta.Session()
 
    for ui_ in sa.query(Ui).all():
 
        if ui_.ui_active:
 
            log.debug('config from db: [%s] %s=%r', ui_.ui_section,
 
                      ui_.ui_key, ui_.ui_value)
 
            baseui.setconfig(ascii_bytes(ui_.ui_section), ascii_bytes(ui_.ui_key),
 
                             b'' if ui_.ui_value is None else safe_bytes(ui_.ui_value))
 

	
 
    # force set push_ssl requirement to False, Kallithea handles that
 
    baseui.setconfig(b'web', b'push_ssl', False)
 
    baseui.setconfig(b'web', b'allow_push', b'*')
 
    # prevent interactive questions for ssh password / passphrase
 
    ssh = baseui.config(b'ui', b'ssh', default=b'ssh')
 
    baseui.setconfig(b'ui', b'ssh', b'%s -oBatchMode=yes -oIdentitiesOnly=yes' % ssh)
 
    # push / pull hooks
 
    baseui.setconfig(b'hooks', b'changegroup.kallithea_log_push_action', b'python:kallithea.lib.hooks.log_push_action')
 
    baseui.setconfig(b'hooks', b'outgoing.kallithea_log_pull_action', b'python:kallithea.lib.hooks.log_pull_action')
 

	
 
    if repo_path is not None:
 
        hgrc_path = os.path.join(repo_path, '.hg', 'hgrc')
 
        if os.path.isfile(hgrc_path):
 
            log.debug('reading hgrc from %s', hgrc_path)
 
            cfg = mercurial.config.config()
 
            cfg.read(safe_bytes(hgrc_path))
 
            for section in ui_sections:
 
                for k, v in cfg.items(section):
 
                    log.debug('config from file: [%s] %s=%s', section, k, v)
 
                    baseui.setconfig(ascii_bytes(section), ascii_bytes(k), safe_bytes(v))
 
        else:
 
            log.debug('hgrc file is not present at %s, skipping...', hgrc_path)
 

	
 
    return baseui
 

	
 

	
 
def set_app_settings(config):
 
    """
 
    Updates app config with new settings from database
 

	
 
    :param config:
 
    """
 
    hgsettings = Setting.get_app_settings()
 
    for k, v in hgsettings.items():
 
        config[k] = v
 

	
 

	
 
def set_vcs_config(config):
 
    """
 
    Patch VCS config with some Kallithea specific stuff
 

	
 
    :param config: kallithea.CONFIG
 
    """
 
    settings.BACKENDS = {
 
        'hg': 'kallithea.lib.vcs.backends.hg.MercurialRepository',
 
        'git': 'kallithea.lib.vcs.backends.git.GitRepository',
 
    }
 

	
 
    settings.GIT_EXECUTABLE_PATH = config.get('git_path', 'git')
 
    settings.GIT_REV_FILTER = config.get('git_rev_filter', '--all').strip()
 
    settings.DEFAULT_ENCODINGS = aslist(config.get('default_encoding',
 
                                                        'utf-8'), sep=',')
 

	
 

	
 
def set_indexer_config(config):
 
    """
 
    Update Whoosh index mapping
 

	
 
    :param config: kallithea.CONFIG
 
    """
 
    log.debug('adding extra into INDEX_EXTENSIONS')
 
    kallithea.config.conf.INDEX_EXTENSIONS.extend(re.split(r'\s+', config.get('index.extensions', '')))
 

	
 
    log.debug('adding extra into INDEX_FILENAMES')
 
    kallithea.config.conf.INDEX_FILENAMES.extend(re.split(r'\s+', config.get('index.filenames', '')))
 

	
 

	
 
def map_groups(path):
 
    """
 
    Given a full path to a repository, create all nested groups that this
 
    repo is inside. This function creates parent-child relationships between
 
    groups and creates default perms for all new groups.
 

	
 
    :param paths: full path to repository
 
    """
 
    from kallithea.model.repo_group import RepoGroupModel
 
    sa = meta.Session()
 
    groups = path.split(Repository.url_sep())
 
    parent = None
 
    group = None
 

	
 
    # last element is repo in nested groups structure
 
    groups = groups[:-1]
 
    rgm = RepoGroupModel()
 
    owner = User.get_first_admin()
 
    for lvl, group_name in enumerate(groups):
 
        group_name = u'/'.join(groups[:lvl] + [group_name])
 
        group_name = '/'.join(groups[:lvl] + [group_name])
 
        group = RepoGroup.get_by_group_name(group_name)
 
        desc = '%s group' % group_name
 

	
 
        # skip folders that are now removed repos
 
        if REMOVED_REPO_PAT.match(group_name):
 
            break
 

	
 
        if group is None:
 
            log.debug('creating group level: %s group_name: %s',
 
                      lvl, group_name)
 
            group = RepoGroup(group_name, parent)
 
            group.group_description = desc
 
            group.owner = owner
 
            sa.add(group)
 
            rgm._create_default_perms(group)
 
            sa.flush()
 

	
 
        parent = group
 
    return group
 

	
 

	
 
def repo2db_mapper(initial_repo_dict, remove_obsolete=False,
 
                   install_git_hooks=False, user=None, overwrite_git_hooks=False):
 
    """
 
    maps all repos given in initial_repo_dict, non existing repositories
 
    are created, if remove_obsolete is True it also check for db entries
 
    that are not in initial_repo_dict and removes them.
 

	
 
    :param initial_repo_dict: mapping with repositories found by scanning methods
 
    :param remove_obsolete: check for obsolete entries in database
 
    :param install_git_hooks: if this is True, also check and install git hook
 
        for a repo if missing
 
    :param overwrite_git_hooks: if this is True, overwrite any existing git hooks
 
        that may be encountered (even if user-deployed)
 
    """
 
    from kallithea.model.repo import RepoModel
 
    from kallithea.model.scm import ScmModel
 
    sa = meta.Session()
 
    repo_model = RepoModel()
 
    if user is None:
 
        user = User.get_first_admin()
 
    added = []
 

	
 
    # creation defaults
 
    defs = Setting.get_default_repo_settings(strip_prefix=True)
 
    enable_statistics = defs.get('repo_enable_statistics')
 
    enable_downloads = defs.get('repo_enable_downloads')
 
    private = defs.get('repo_private')
 

	
 
    for name, repo in initial_repo_dict.items():
 
        group = map_groups(name)
 
        db_repo = repo_model.get_by_repo_name(name)
 
        # found repo that is on filesystem not in Kallithea database
 
        if not db_repo:
 
            log.info('repository %s not found, creating now', name)
 
            added.append(name)
 
            desc = (repo.description
 
                    if repo.description != 'unknown'
 
                    else '%s repository' % name)
 

	
 
            new_repo = repo_model._create_repo(
 
                repo_name=name,
 
                repo_type=repo.alias,
 
                description=desc,
 
                repo_group=getattr(group, 'group_id', None),
 
                owner=user,
 
                enable_downloads=enable_downloads,
 
                enable_statistics=enable_statistics,
 
                private=private,
 
                state=Repository.STATE_CREATED
 
            )
 
            sa.commit()
 
            # we added that repo just now, and make sure it has githook
 
            # installed, and updated server info
 
            if new_repo.repo_type == 'git':
 
                git_repo = new_repo.scm_instance
 
                ScmModel().install_git_hooks(git_repo)
 
                # update repository server-info
 
                log.debug('Running update server info')
 
                git_repo._update_server_info()
 
            new_repo.update_changeset_cache()
 
        elif install_git_hooks:
 
            if db_repo.repo_type == 'git':
 
                ScmModel().install_git_hooks(db_repo.scm_instance, force_create=overwrite_git_hooks)
 

	
 
    removed = []
 
    # remove from database those repositories that are not in the filesystem
 
    for repo in sa.query(Repository).all():
 
        if repo.repo_name not in initial_repo_dict:
 
            if remove_obsolete:
 
                log.debug("Removing non-existing repository found in db `%s`",
 
                          repo.repo_name)
 
                try:
 
                    RepoModel().delete(repo, forks='detach', fs_remove=False)
 
                    sa.commit()
 
                except Exception:
 
                    #don't hold further removals on error
 
                    log.error(traceback.format_exc())
 
                    sa.rollback()
 
            removed.append(repo.repo_name)
 
    return added, removed
 

	
 

	
 
def load_rcextensions(root_path):
 
    path = os.path.join(root_path, 'rcextensions', '__init__.py')
 
    if os.path.isfile(path):
 
        rcext = create_module('rc', path)
 
        EXT = kallithea.EXTENSIONS = rcext
 
        log.debug('Found rcextensions now loading %s...', rcext)
 

	
 
        # Additional mappings that are not present in the pygments lexers
 
        kallithea.config.conf.LANGUAGES_EXTENSIONS_MAP.update(getattr(EXT, 'EXTRA_MAPPINGS', {}))
 

	
 
        # OVERRIDE OUR EXTENSIONS FROM RC-EXTENSIONS (if present)
 

	
 
        if getattr(EXT, 'INDEX_EXTENSIONS', []):
 
            log.debug('settings custom INDEX_EXTENSIONS')
 
            kallithea.config.conf.INDEX_EXTENSIONS = getattr(EXT, 'INDEX_EXTENSIONS', [])
 

	
 
        # ADDITIONAL MAPPINGS
 
        log.debug('adding extra into INDEX_EXTENSIONS')
 
        kallithea.config.conf.INDEX_EXTENSIONS.extend(getattr(EXT, 'EXTRA_INDEX_EXTENSIONS', []))
 

	
 
        # auto check if the module is not missing any data, set to default if is
 
        # this will help autoupdate new feature of rcext module
 
        #from kallithea.config import rcextensions
 
        #for k in dir(rcextensions):
 
        #    if not k.startswith('_') and not hasattr(EXT, k):
 
        #        setattr(EXT, k, getattr(rcextensions, k))
 

	
 

	
 
#==============================================================================
 
# MISC
 
#==============================================================================
 

	
 
git_req_ver = StrictVersion('1.7.4')
 

	
 
def check_git_version():
 
    """
 
    Checks what version of git is installed on the system, and raise a system exit
 
    if it's too old for Kallithea to work properly.
 
    """
 
    if 'git' not in kallithea.BACKENDS:
 
        return None
 

	
 
    if not settings.GIT_EXECUTABLE_PATH:
 
        log.warning('No git executable configured - check "git_path" in the ini file.')
 
        return None
 

	
 
    try:
 
        stdout, stderr = GitRepository._run_git_command(['--version'])
 
    except RepositoryError as e:
 
        # message will already have been logged as error
 
        log.warning('No working git executable found - check "git_path" in the ini file.')
 
        return None
 

	
 
    if stderr:
 
        log.warning('Error/stderr from "%s --version":\n%s', settings.GIT_EXECUTABLE_PATH, safe_str(stderr))
 

	
 
    if not stdout:
 
        log.warning('No working git executable found - check "git_path" in the ini file.')
 
        return None
 

	
 
    output = safe_str(stdout).strip()
 
    m = re.search(r"\d+.\d+.\d+", output)
 
    if m:
 
        ver = StrictVersion(m.group(0))
 
        log.debug('Git executable: "%s", version %s (parsed from: "%s")',
 
                  settings.GIT_EXECUTABLE_PATH, ver, output)
 
        if ver < git_req_ver:
 
            log.error('Kallithea detected %s version %s, which is too old '
 
                      'for the system to function properly. '
 
                      'Please upgrade to version %s or later. '
 
                      'If you strictly need Mercurial repositories, you can '
 
                      'clear the "git_path" setting in the ini file.',
 
                      settings.GIT_EXECUTABLE_PATH, ver, git_req_ver)
 
            log.error("Terminating ...")
 
            sys.exit(1)
 
    else:
 
        ver = StrictVersion('0.0.0')
 
        log.warning('Error finding version number in "%s --version" stdout:\n%s',
 
                    settings.GIT_EXECUTABLE_PATH, output)
 

	
 
    return ver
 

	
 

	
 
#===============================================================================
 
# CACHE RELATED METHODS
 
#===============================================================================
 

	
 
def conditional_cache(region, prefix, condition, func):
 
    """
 

	
 
    Conditional caching function use like::
 
        def _c(arg):
 
            #heavy computation function
 
            return data
 

	
 
        # depending from condition the compute is wrapped in cache or not
 
        compute = conditional_cache('short_term', 'cache_desc', condition=True, func=func)
 
        return compute(arg)
 

	
 
    :param region: name of cache region
 
    :param prefix: cache region prefix
 
    :param condition: condition for cache to be triggered, and return data cached
 
    :param func: wrapped heavy function to compute
 

	
 
    """
 
    wrapped = func
 
    if condition:
 
        log.debug('conditional_cache: True, wrapping call of '
 
                  'func: %s into %s region cache' % (region, func))
 
        wrapped = beaker.cache._cache_decorate((prefix,), None, None, region)(func)
 

	
 
    return wrapped
kallithea/lib/utils2.py
Show inline comments
 
# -*- 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.utils2
 
~~~~~~~~~~~~~~~~~~~~
 

	
 
Some simple helper functions.
 
Note: all these functions should be independent of Kallithea classes, i.e.
 
models, controllers, etc.  to prevent import cycles.
 

	
 
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: Jan 5, 2011
 
:author: marcink
 
:copyright: (c) 2013 RhodeCode GmbH, and others.
 
:license: GPLv3, see LICENSE.md for more details.
 
"""
 

	
 
from __future__ import print_function
 

	
 
import binascii
 
import datetime
 
import json
 
import os
 
import pwd
 
import re
 
import time
 
import urllib.parse
 

	
 
import urlobject
 
from tg.i18n import ugettext as _
 
from tg.i18n import ungettext
 
from webhelpers2.text import collapse, remove_formatting, strip_tags
 

	
 
from kallithea.lib.vcs.utils import ascii_bytes, ascii_str, safe_bytes, safe_str  # re-export
 
from kallithea.lib.vcs.utils.lazy import LazyProperty
 

	
 

	
 
def str2bool(_str):
 
    """
 
    returns True/False value from given string, it tries to translate the
 
    string into boolean
 

	
 
    :param _str: string value to translate into boolean
 
    :rtype: boolean
 
    :returns: boolean from given string
 
    """
 
    if _str is None:
 
        return False
 
    if _str in (True, False):
 
        return _str
 
    _str = str(_str).strip().lower()
 
    return _str in ('t', 'true', 'y', 'yes', 'on', '1')
 

	
 

	
 
def aslist(obj, sep=None, strip=True):
 
    """
 
    Returns given string separated by sep as list
 

	
 
    :param obj:
 
    :param sep:
 
    :param strip:
 
    """
 
    if isinstance(obj, (str)):
 
        lst = obj.split(sep)
 
        if strip:
 
            lst = [v.strip() for v in lst]
 
        return lst
 
    elif isinstance(obj, (list, tuple)):
 
        return obj
 
    elif obj is None:
 
        return []
 
    else:
 
        return [obj]
 

	
 

	
 
def convert_line_endings(line, mode):
 
    """
 
    Converts a given line  "line end" according to given mode
 

	
 
    Available modes are::
 
        0 - Unix
 
        1 - Mac
 
        2 - DOS
 

	
 
    :param line: given line to convert
 
    :param mode: mode to convert to
 
    :rtype: str
 
    :return: converted line according to mode
 
    """
 
    if mode == 0:
 
        line = line.replace('\r\n', '\n')
 
        line = line.replace('\r', '\n')
 
    elif mode == 1:
 
        line = line.replace('\r\n', '\r')
 
        line = line.replace('\n', '\r')
 
    elif mode == 2:
 
        line = re.sub("\r(?!\n)|(?<!\r)\n", "\r\n", line)
 
    return line
 

	
 

	
 
def detect_mode(line, default):
 
    """
 
    Detects line break for given line, if line break couldn't be found
 
    given default value is returned
 

	
 
    :param line: str line
 
    :param default: default
 
    :rtype: int
 
    :return: value of line end on of 0 - Unix, 1 - Mac, 2 - DOS
 
    """
 
    if line.endswith('\r\n'):
 
        return 2
 
    elif line.endswith('\n'):
 
        return 0
 
    elif line.endswith('\r'):
 
        return 1
 
    else:
 
        return default
 

	
 

	
 
def generate_api_key():
 
    """
 
    Generates a random (presumably unique) API key.
 

	
 
    This value is used in URLs and "Bearer" HTTP Authorization headers,
 
    which in practice means it should only contain URL-safe characters
 
    (RFC 3986):
 

	
 
        unreserved = ALPHA / DIGIT / "-" / "." / "_" / "~"
 
    """
 
    # Hexadecimal certainly qualifies as URL-safe.
 
    return ascii_str(binascii.hexlify(os.urandom(20)))
 

	
 

	
 
def safe_int(val, default=None):
 
    """
 
    Returns int() of val if val is not convertable to int use default
 
    instead
 

	
 
    :param val:
 
    :param default:
 
    """
 
    try:
 
        val = int(val)
 
    except (ValueError, TypeError):
 
        val = default
 
    return val
 

	
 

	
 
def remove_suffix(s, suffix):
 
    if s.endswith(suffix):
 
        s = s[:-1 * len(suffix)]
 
    return s
 

	
 

	
 
def remove_prefix(s, prefix):
 
    if s.startswith(prefix):
 
        s = s[len(prefix):]
 
    return s
 

	
 

	
 
def age(prevdate, show_short_version=False, now=None):
 
    """
 
    turns a datetime into an age string.
 
    If show_short_version is True, then it will generate a not so accurate but shorter string,
 
    example: 2days ago, instead of 2 days and 23 hours ago.
 

	
 
    :param prevdate: datetime object
 
    :param show_short_version: if it should approximate the date and return a shorter string
 
    :rtype: str
 
    :returns: str words describing age
 
    """
 
    now = now or datetime.datetime.now()
 
    order = ['year', 'month', 'day', 'hour', 'minute', 'second']
 
    deltas = {}
 
    future = False
 

	
 
    if prevdate > now:
 
        now, prevdate = prevdate, now
 
        future = True
 
    if future:
 
        prevdate = prevdate.replace(microsecond=0)
 
    # Get date parts deltas
 
    from dateutil import relativedelta
 
    for part in order:
 
        d = relativedelta.relativedelta(now, prevdate)
 
        deltas[part] = getattr(d, part + 's')
 

	
 
    # Fix negative offsets (there is 1 second between 10:59:59 and 11:00:00,
 
    # not 1 hour, -59 minutes and -59 seconds)
 
    for num, length in [(5, 60), (4, 60), (3, 24)]:  # seconds, minutes, hours
 
        part = order[num]
 
        carry_part = order[num - 1]
 

	
 
        if deltas[part] < 0:
 
            deltas[part] += length
 
            deltas[carry_part] -= 1
 

	
 
    # Same thing for days except that the increment depends on the (variable)
 
    # number of days in the month
 
    month_lengths = [31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31]
 
    if deltas['day'] < 0:
 
        if prevdate.month == 2 and (prevdate.year % 4 == 0 and
 
            (prevdate.year % 100 != 0 or prevdate.year % 400 == 0)
 
        ):
 
            deltas['day'] += 29
 
        else:
 
            deltas['day'] += month_lengths[prevdate.month - 1]
 

	
 
        deltas['month'] -= 1
 

	
 
    if deltas['month'] < 0:
 
        deltas['month'] += 12
 
        deltas['year'] -= 1
 

	
 
    # In short version, we want nicer handling of ages of more than a year
 
    if show_short_version:
 
        if deltas['year'] == 1:
 
            # ages between 1 and 2 years: show as months
 
            deltas['month'] += 12
 
            deltas['year'] = 0
 
        if deltas['year'] >= 2:
 
            # ages 2+ years: round
 
            if deltas['month'] > 6:
 
                deltas['year'] += 1
 
                deltas['month'] = 0
 

	
 
    # Format the result
 
    fmt_funcs = {
 
        'year': lambda d: ungettext(u'%d year', '%d years', d) % d,
 
        'month': lambda d: ungettext(u'%d month', '%d months', d) % d,
 
        'day': lambda d: ungettext(u'%d day', '%d days', d) % d,
 
        'hour': lambda d: ungettext(u'%d hour', '%d hours', d) % d,
 
        'minute': lambda d: ungettext(u'%d minute', '%d minutes', d) % d,
 
        'second': lambda d: ungettext(u'%d second', '%d seconds', d) % d,
 
        'year': lambda d: ungettext('%d year', '%d years', d) % d,
 
        'month': lambda d: ungettext('%d month', '%d months', d) % d,
 
        'day': lambda d: ungettext('%d day', '%d days', d) % d,
 
        'hour': lambda d: ungettext('%d hour', '%d hours', d) % d,
 
        'minute': lambda d: ungettext('%d minute', '%d minutes', d) % d,
 
        'second': lambda d: ungettext('%d second', '%d seconds', d) % d,
 
    }
 

	
 
    for i, part in enumerate(order):
 
        value = deltas[part]
 
        if value == 0:
 
            continue
 

	
 
        if i < 5:
 
            sub_part = order[i + 1]
 
            sub_value = deltas[sub_part]
 
        else:
 
            sub_value = 0
 

	
 
        if sub_value == 0 or show_short_version:
 
            if future:
 
                return _('in %s') % fmt_funcs[part](value)
 
            else:
 
                return _('%s ago') % fmt_funcs[part](value)
 
        if future:
 
            return _('in %s and %s') % (fmt_funcs[part](value),
 
                fmt_funcs[sub_part](sub_value))
 
        else:
 
            return _('%s and %s ago') % (fmt_funcs[part](value),
 
                fmt_funcs[sub_part](sub_value))
 

	
 
    return _('just now')
 

	
 

	
 
def uri_filter(uri):
 
    """
 
    Removes user:password from given url string
 

	
 
    :param uri:
 
    :rtype: str
 
    :returns: filtered list of strings
 
    """
 
    if not uri:
 
        return []
 

	
 
    proto = ''
 

	
 
    for pat in ('https://', 'http://', 'git://'):
 
        if uri.startswith(pat):
 
            uri = uri[len(pat):]
 
            proto = pat
 
            break
 

	
 
    # remove passwords and username
 
    uri = uri[uri.find('@') + 1:]
 

	
 
    # get the port
 
    cred_pos = uri.find(':')
 
    if cred_pos == -1:
 
        host, port = uri, None
 
    else:
 
        host, port = uri[:cred_pos], uri[cred_pos + 1:]
 

	
 
    return [_f for _f in [proto, host, port] if _f]
 

	
 

	
 
def credentials_filter(uri):
 
    """
 
    Returns a url with removed credentials
 

	
 
    :param uri:
 
    """
 

	
 
    uri = uri_filter(uri)
 
    # check if we have port
 
    if len(uri) > 2 and uri[2]:
 
        uri[2] = ':' + uri[2]
 

	
 
    return ''.join(uri)
 

	
 

	
 
def get_clone_url(clone_uri_tmpl, prefix_url, repo_name, repo_id, username=None):
 
    parsed_url = urlobject.URLObject(prefix_url)
 
    prefix = urllib.parse.unquote(parsed_url.path.rstrip('/'))
 
    try:
 
        system_user = pwd.getpwuid(os.getuid()).pw_name
 
    except Exception: # TODO: support all systems - especially Windows
 
        system_user = 'kallithea' # hardcoded default value ...
 
    args = {
 
        'scheme': parsed_url.scheme,
 
        'user': urllib.parse.quote(username or ''),
 
        'netloc': parsed_url.netloc + prefix,  # like "hostname:port/prefix" (with optional ":port" and "/prefix")
 
        'prefix': prefix, # undocumented, empty or starting with /
 
        'repo': repo_name,
 
        'repoid': str(repo_id),
 
        'system_user': system_user,
 
        'hostname': parsed_url.hostname,
 
    }
 
    url = re.sub('{([^{}]+)}', lambda m: args.get(m.group(1), m.group(0)), clone_uri_tmpl)
 

	
 
    # remove leading @ sign if it's present. Case of empty user
 
    url_obj = urlobject.URLObject(url)
 
    if not url_obj.username:
 
        url_obj = url_obj.with_username(None)
 

	
 
    return str(url_obj)
 

	
 

	
 
def get_changeset_safe(repo, rev):
 
    """
 
    Safe version of get_changeset if this changeset doesn't exists for a
 
    repo it returns a Dummy one instead
 

	
 
    :param repo:
 
    :param rev:
 
    """
 
    from kallithea.lib.vcs.backends.base import BaseRepository
 
    from kallithea.lib.vcs.exceptions import RepositoryError
 
    from kallithea.lib.vcs.backends.base import EmptyChangeset
 
    if not isinstance(repo, BaseRepository):
 
        raise Exception('You must pass an Repository '
 
                        'object as first argument got %s' % type(repo))
 

	
 
    try:
 
        cs = repo.get_changeset(rev)
 
    except (RepositoryError, LookupError):
 
        cs = EmptyChangeset(requested_revision=rev)
 
    return cs
 

	
 

	
 
def datetime_to_time(dt):
 
    if dt:
 
        return time.mktime(dt.timetuple())
 

	
 

	
 
def time_to_datetime(tm):
 
    if tm:
 
        if isinstance(tm, str):
 
            try:
 
                tm = float(tm)
 
            except ValueError:
 
                return
 
        return datetime.datetime.fromtimestamp(tm)
 

	
 

	
 
# Must match regexp in kallithea/public/js/base.js MentionsAutoComplete()
 
# Check char before @ - it must not look like we are in an email addresses.
 
# Matching is greedy so we don't have to look beyond the end.
 
MENTIONS_REGEX = re.compile(r'(?:^|(?<=[^a-zA-Z0-9]))@([a-zA-Z0-9][-_.a-zA-Z0-9]*[a-zA-Z0-9])')
 

	
 

	
 
def extract_mentioned_usernames(text):
 
    r"""
 
    Returns list of (possible) usernames @mentioned in given text.
 

	
 
    >>> extract_mentioned_usernames('@1-2.a_X,@1234 not@not @ddd@not @n @ee @ff @gg, @gg;@hh @n\n@zz,')
 
    ['1-2.a_X', '1234', 'ddd', 'ee', 'ff', 'gg', 'gg', 'hh', 'zz']
 
    """
 
    return MENTIONS_REGEX.findall(text)
 

	
 

	
 
def extract_mentioned_users(text):
 
    """ Returns set of actual database Users @mentioned in given text. """
 
    from kallithea.model.db import User
 
    result = set()
 
    for name in extract_mentioned_usernames(text):
 
        user = User.get_by_username(name, case_insensitive=True)
 
        if user is not None and not user.is_default_user:
 
            result.add(user)
 
    return result
 

	
 

	
 
class AttributeDict(dict):
 
    def __getattr__(self, attr):
 
        return self.get(attr, None)
 
    __setattr__ = dict.__setitem__
 
    __delattr__ = dict.__delitem__
 

	
 

	
 
def obfuscate_url_pw(engine):
 
    from sqlalchemy.engine import url as sa_url
 
    from sqlalchemy.exc import ArgumentError
 
    try:
 
        _url = sa_url.make_url(engine or '')
 
    except ArgumentError:
 
        return engine
 
    if _url.password:
 
        _url.password = 'XXXXX'
 
    return str(_url)
 

	
 

	
 
class HookEnvironmentError(Exception): pass
 

	
 

	
 
def get_hook_environment():
 
    """
 
    Get hook context by deserializing the global KALLITHEA_EXTRAS environment
 
    variable.
 

	
 
    Called early in Git out-of-process hooks to get .ini config path so the
 
    basic environment can be configured properly. Also used in all hooks to get
 
    information about the action that triggered it.
 
    """
 

	
 
    try:
 
        kallithea_extras = os.environ['KALLITHEA_EXTRAS']
 
    except KeyError:
 
        raise HookEnvironmentError("Environment variable KALLITHEA_EXTRAS not found")
 

	
 
    extras = json.loads(kallithea_extras)
 
    for k in ['username', 'repository', 'scm', 'action', 'ip', 'config']:
 
        try:
 
            extras[k]
 
        except KeyError:
 
            raise HookEnvironmentError('Missing key %s in KALLITHEA_EXTRAS %s' % (k, extras))
 

	
 
    return AttributeDict(extras)
 

	
 

	
 
def set_hook_environment(username, ip_addr, repo_name, repo_alias, action=None):
 
    """Prepare global context for running hooks by serializing data in the
 
    global KALLITHEA_EXTRAS environment variable.
 

	
 
    Most importantly, this allow Git hooks to do proper logging and updating of
 
    caches after pushes.
 

	
 
    Must always be called before anything with hooks are invoked.
 
    """
 
    from kallithea import CONFIG
 
    extras = {
 
        'ip': ip_addr, # used in log_push/pull_action action_logger
 
        'username': username,
 
        'action': action or 'push_local', # used in log_push_action_raw_ids action_logger
 
        'repository': repo_name,
 
        'scm': repo_alias, # used to pick hack in log_push_action_raw_ids
 
        'config': CONFIG['__file__'], # used by git hook to read config
 
    }
 
    os.environ['KALLITHEA_EXTRAS'] = json.dumps(extras)
 

	
 

	
 
def get_current_authuser():
 
    """
 
    Gets kallithea user from threadlocal tmpl_context variable if it's
 
    defined, else returns None.
 
    """
 
    from tg import tmpl_context
 
    try:
 
        return getattr(tmpl_context, 'authuser', None)
 
    except TypeError:  # No object (name: context) has been registered for this thread
 
        return None
 

	
 

	
 
class OptionalAttr(object):
 
    """
 
    Special Optional Option that defines other attribute. Example::
 

	
 
        def test(apiuser, userid=Optional(OAttr('apiuser')):
 
            user = Optional.extract(userid)
 
            # calls
 

	
 
    """
 

	
 
    def __init__(self, attr_name):
 
        self.attr_name = attr_name
 

	
 
    def __repr__(self):
 
        return '<OptionalAttr:%s>' % self.attr_name
 

	
 
    def __call__(self):
 
        return self
 

	
 

	
 
# alias
 
OAttr = OptionalAttr
 

	
 

	
 
class Optional(object):
 
    """
 
    Defines an optional parameter::
 

	
 
        param = param.getval() if isinstance(param, Optional) else param
 
        param = param() if isinstance(param, Optional) else param
 

	
 
    is equivalent of::
 

	
 
        param = Optional.extract(param)
 

	
 
    """
 

	
 
    def __init__(self, type_):
 
        self.type_ = type_
 

	
 
    def __repr__(self):
 
        return '<Optional:%s>' % self.type_.__repr__()
 

	
 
    def __call__(self):
 
        return self.getval()
 

	
 
    def getval(self):
 
        """
 
        returns value from this Optional instance
 
        """
 
        if isinstance(self.type_, OAttr):
 
            # use params name
 
            return self.type_.attr_name
 
        return self.type_
 

	
 
    @classmethod
 
    def extract(cls, val):
 
        """
 
        Extracts value from Optional() instance
 

	
 
        :param val:
 
        :return: original value if it's not Optional instance else
 
            value of instance
 
        """
 
        if isinstance(val, cls):
 
            return val.getval()
 
        return val
 

	
 

	
 
def urlreadable(s, _cleanstringsub=re.compile('[^-a-zA-Z0-9./]+').sub):
 
    return _cleanstringsub('_', s).rstrip('_')
 

	
 

	
 
def recursive_replace(str_, replace=' '):
 
    """
 
    Recursive replace of given sign to just one instance
 

	
 
    :param str_: given string
 
    :param replace: char to find and replace multiple instances
 

	
 
    Examples::
 
    >>> recursive_replace("Mighty---Mighty-Bo--sstones",'-')
 
    'Mighty-Mighty-Bo-sstones'
 
    """
 

	
 
    if str_.find(replace * 2) == -1:
 
        return str_
 
    else:
 
        str_ = str_.replace(replace * 2, replace)
 
        return recursive_replace(str_, replace)
 

	
 

	
 
def repo_name_slug(value):
 
    """
 
    Return slug of name of repository
 
    This function is called on each creation/modification
 
    of repository to prevent bad names in repo
 
    """
 

	
 
    slug = remove_formatting(value)
 
    slug = strip_tags(slug)
 

	
 
    for c in r"""`?=[]\;'"<>,/~!@#$%^&*()+{}|: """:
 
        slug = slug.replace(c, '-')
 
    slug = recursive_replace(slug, '-')
 
    slug = collapse(slug, '-')
 
    return slug
 

	
 

	
 
def ask_ok(prompt, retries=4, complaint='Yes or no please!'):
 
    while True:
 
        ok = input(prompt)
 
        if ok in ('y', 'ye', 'yes'):
 
            return True
 
        if ok in ('n', 'no', 'nop', 'nope'):
 
            return False
 
        retries = retries - 1
 
        if retries < 0:
 
            raise IOError
 
        print(complaint)
kallithea/lib/vcs/backends/base.py
Show inline comments
 
@@ -645,413 +645,413 @@ class BaseChangeset(object):
 
        data['author'] = {'name': self.author_name, 'email': self.author_email}
 
        data['added'] = [node.path for node in self.added]
 
        data['changed'] = [node.path for node in self.changed]
 
        data['removed'] = [node.path for node in self.removed]
 
        return data
 

	
 
    @LazyProperty
 
    def closesbranch(self):
 
        return False
 

	
 
    @LazyProperty
 
    def obsolete(self):
 
        return False
 

	
 
    @LazyProperty
 
    def bumped(self):
 
        return False
 

	
 
    @LazyProperty
 
    def divergent(self):
 
        return False
 

	
 
    @LazyProperty
 
    def extinct(self):
 
        return False
 

	
 
    @LazyProperty
 
    def unstable(self):
 
        return False
 

	
 
    @LazyProperty
 
    def phase(self):
 
        return ''
 

	
 

	
 
class BaseWorkdir(object):
 
    """
 
    Working directory representation of single repository.
 

	
 
    :attribute: repository: repository object of working directory
 
    """
 

	
 
    def __init__(self, repository):
 
        self.repository = repository
 

	
 
    def get_branch(self):
 
        """
 
        Returns name of current branch.
 
        """
 
        raise NotImplementedError
 

	
 
    def get_changeset(self):
 
        """
 
        Returns current changeset.
 
        """
 
        raise NotImplementedError
 

	
 
    def get_added(self):
 
        """
 
        Returns list of ``FileNode`` objects marked as *new* in working
 
        directory.
 
        """
 
        raise NotImplementedError
 

	
 
    def get_changed(self):
 
        """
 
        Returns list of ``FileNode`` objects *changed* in working directory.
 
        """
 
        raise NotImplementedError
 

	
 
    def get_removed(self):
 
        """
 
        Returns list of ``RemovedFileNode`` objects marked as *removed* in
 
        working directory.
 
        """
 
        raise NotImplementedError
 

	
 
    def get_untracked(self):
 
        """
 
        Returns list of ``FileNode`` objects which are present within working
 
        directory however are not tracked by repository.
 
        """
 
        raise NotImplementedError
 

	
 
    def get_status(self):
 
        """
 
        Returns dict with ``added``, ``changed``, ``removed`` and ``untracked``
 
        lists.
 
        """
 
        raise NotImplementedError
 

	
 
    def commit(self, message, **kwargs):
 
        """
 
        Commits local (from working directory) changes and returns newly
 
        created
 
        ``Changeset``. Updates repository's ``revisions`` list.
 

	
 
        :raises ``CommitError``: if any error occurs while committing
 
        """
 
        raise NotImplementedError
 

	
 
    def update(self, revision=None):
 
        """
 
        Fetches content of the given revision and populates it within working
 
        directory.
 
        """
 
        raise NotImplementedError
 

	
 
    def checkout_branch(self, branch=None):
 
        """
 
        Checks out ``branch`` or the backend's default branch.
 

	
 
        Raises ``BranchDoesNotExistError`` if the branch does not exist.
 
        """
 
        raise NotImplementedError
 

	
 

	
 
class BaseInMemoryChangeset(object):
 
    """
 
    Represents differences between repository's state (most recent head) and
 
    changes made *in place*.
 

	
 
    **Attributes**
 

	
 
        ``repository``
 
            repository object for this in-memory-changeset
 

	
 
        ``added``
 
            list of ``FileNode`` objects marked as *added*
 

	
 
        ``changed``
 
            list of ``FileNode`` objects marked as *changed*
 

	
 
        ``removed``
 
            list of ``FileNode`` or ``RemovedFileNode`` objects marked to be
 
            *removed*
 

	
 
        ``parents``
 
            list of ``Changeset`` representing parents of in-memory changeset.
 
            Should always be 2-element sequence.
 

	
 
    """
 

	
 
    def __init__(self, repository):
 
        self.repository = repository
 
        self.added = []
 
        self.changed = []
 
        self.removed = []
 
        self.parents = []
 

	
 
    def add(self, *filenodes):
 
        """
 
        Marks given ``FileNode`` objects as *to be committed*.
 

	
 
        :raises ``NodeAlreadyExistsError``: if node with same path exists at
 
          latest changeset
 
        :raises ``NodeAlreadyAddedError``: if node with same path is already
 
          marked as *added*
 
        """
 
        # Check if not already marked as *added* first
 
        for node in filenodes:
 
            if node.path in (n.path for n in self.added):
 
                raise NodeAlreadyAddedError("Such FileNode %s is already "
 
                    "marked for addition" % node.path)
 
        for node in filenodes:
 
            self.added.append(node)
 

	
 
    def change(self, *filenodes):
 
        """
 
        Marks given ``FileNode`` objects to be *changed* in next commit.
 

	
 
        :raises ``EmptyRepositoryError``: if there are no changesets yet
 
        :raises ``NodeAlreadyExistsError``: if node with same path is already
 
          marked to be *changed*
 
        :raises ``NodeAlreadyRemovedError``: if node with same path is already
 
          marked to be *removed*
 
        :raises ``NodeDoesNotExistError``: if node doesn't exist in latest
 
          changeset
 
        :raises ``NodeNotChangedError``: if node hasn't really be changed
 
        """
 
        for node in filenodes:
 
            if node.path in (n.path for n in self.removed):
 
                raise NodeAlreadyRemovedError("Node at %s is already marked "
 
                    "as removed" % node.path)
 
        try:
 
            self.repository.get_changeset()
 
        except EmptyRepositoryError:
 
            raise EmptyRepositoryError("Nothing to change - try to *add* new "
 
                "nodes rather than changing them")
 
        for node in filenodes:
 
            if node.path in (n.path for n in self.changed):
 
                raise NodeAlreadyChangedError("Node at '%s' is already "
 
                    "marked as changed" % node.path)
 
            self.changed.append(node)
 

	
 
    def remove(self, *filenodes):
 
        """
 
        Marks given ``FileNode`` (or ``RemovedFileNode``) objects to be
 
        *removed* in next commit.
 

	
 
        :raises ``NodeAlreadyRemovedError``: if node has been already marked to
 
          be *removed*
 
        :raises ``NodeAlreadyChangedError``: if node has been already marked to
 
          be *changed*
 
        """
 
        for node in filenodes:
 
            if node.path in (n.path for n in self.removed):
 
                raise NodeAlreadyRemovedError("Node is already marked to "
 
                    "for removal at %s" % node.path)
 
            if node.path in (n.path for n in self.changed):
 
                raise NodeAlreadyChangedError("Node is already marked to "
 
                    "be changed at %s" % node.path)
 
            # We only mark node as *removed* - real removal is done by
 
            # commit method
 
            self.removed.append(node)
 

	
 
    def reset(self):
 
        """
 
        Resets this instance to initial state (cleans ``added``, ``changed``
 
        and ``removed`` lists).
 
        """
 
        self.added = []
 
        self.changed = []
 
        self.removed = []
 
        self.parents = []
 

	
 
    def get_ipaths(self):
 
        """
 
        Returns generator of paths from nodes marked as added, changed or
 
        removed.
 
        """
 
        for node in itertools.chain(self.added, self.changed, self.removed):
 
            yield node.path
 

	
 
    def get_paths(self):
 
        """
 
        Returns list of paths from nodes marked as added, changed or removed.
 
        """
 
        return list(self.get_ipaths())
 

	
 
    def check_integrity(self, parents=None):
 
        """
 
        Checks in-memory changeset's integrity. Also, sets parents if not
 
        already set.
 

	
 
        :raises CommitError: if any error occurs (i.e.
 
          ``NodeDoesNotExistError``).
 
        """
 
        if not self.parents:
 
            parents = parents or []
 
            if len(parents) == 0:
 
                try:
 
                    parents = [self.repository.get_changeset(), None]
 
                except EmptyRepositoryError:
 
                    parents = [None, None]
 
            elif len(parents) == 1:
 
                parents += [None]
 
            self.parents = parents
 

	
 
        # Local parents, only if not None
 
        parents = [p for p in self.parents if p]
 

	
 
        # Check nodes marked as added
 
        for p in parents:
 
            for node in self.added:
 
                try:
 
                    p.get_node(node.path)
 
                except NodeDoesNotExistError:
 
                    pass
 
                else:
 
                    raise NodeAlreadyExistsError("Node at %s already exists "
 
                        "at %s" % (node.path, p))
 

	
 
        # Check nodes marked as changed
 
        missing = set(node.path for node in self.changed)
 
        not_changed = set(node.path for node in self.changed)
 
        if self.changed and not parents:
 
            raise NodeDoesNotExistError(self.changed[0].path)
 
        for p in parents:
 
            for node in self.changed:
 
                try:
 
                    old = p.get_node(node.path)
 
                    missing.remove(node.path)
 
                    # if content actually changed, remove node from unchanged
 
                    if old.content != node.content:
 
                        not_changed.remove(node.path)
 
                except NodeDoesNotExistError:
 
                    pass
 
        if self.changed and missing:
 
            raise NodeDoesNotExistError("Node at %s is missing "
 
                "(parents: %s)" % (node.path, parents))
 

	
 
        if self.changed and not_changed:
 
            raise NodeNotChangedError("Node at %s wasn't actually changed "
 
                "since parents' changesets: %s" % (not_changed.pop(),
 
                    parents)
 
            )
 

	
 
        # Check nodes marked as removed
 
        if self.removed and not parents:
 
            raise NodeDoesNotExistError("Cannot remove node at %s as there "
 
                "were no parents specified" % self.removed[0].path)
 
        really_removed = set()
 
        for p in parents:
 
            for node in self.removed:
 
                try:
 
                    p.get_node(node.path)
 
                    really_removed.add(node.path)
 
                except ChangesetError:
 
                    pass
 
        not_removed = list(set(node.path for node in self.removed) - really_removed)
 
        if not_removed:
 
            raise NodeDoesNotExistError("Cannot remove node at %s from "
 
                "following parents: %s" % (not_removed[0], parents))
 

	
 
    def commit(self, message, author, parents=None, branch=None, date=None,
 
            **kwargs):
 
        """
 
        Performs in-memory commit (doesn't check workdir in any way) and
 
        returns newly created ``Changeset``. Updates repository's
 
        ``revisions``.
 

	
 
        .. note::
 
            While overriding this method each backend's should call
 
            ``self.check_integrity(parents)`` in the first place.
 

	
 
        :param message: message of the commit
 
        :param author: full username, i.e. "Joe Doe <joe.doe@example.com>"
 
        :param parents: single parent or sequence of parents from which commit
 
          would be derived
 
        :param date: ``datetime.datetime`` instance. Defaults to
 
          ``datetime.datetime.now()``.
 
        :param branch: branch name, as string. If none given, default backend's
 
          branch would be used.
 

	
 
        :raises ``CommitError``: if any error occurs while committing
 
        """
 
        raise NotImplementedError
 

	
 

	
 
class EmptyChangeset(BaseChangeset):
 
    """
 
    An dummy empty changeset. It's possible to pass hash when creating
 
    an EmptyChangeset
 
    """
 

	
 
    def __init__(self, cs='0' * 40, repo=None, requested_revision=None,
 
                 alias=None, revision=-1, message='', author='', date=None):
 
        self._empty_cs = cs
 
        self.revision = revision
 
        self.message = message
 
        self.author = author
 
        self.date = date or datetime.datetime.fromtimestamp(0)
 
        self.repository = repo
 
        self.requested_revision = requested_revision
 
        self.alias = alias
 

	
 
    @LazyProperty
 
    def raw_id(self):
 
        """
 
        Returns raw string identifying this changeset, useful for web
 
        representation.
 
        """
 

	
 
        return self._empty_cs
 

	
 
    @LazyProperty
 
    def branch(self):
 
        from kallithea.lib.vcs.backends import get_backend
 
        return get_backend(self.alias).DEFAULT_BRANCH_NAME
 

	
 
    @LazyProperty
 
    def branches(self):
 
        from kallithea.lib.vcs.backends import get_backend
 
        return [get_backend(self.alias).DEFAULT_BRANCH_NAME]
 

	
 
    @LazyProperty
 
    def short_id(self):
 
        return self.raw_id[:12]
 

	
 
    def get_file_changeset(self, path):
 
        return self
 

	
 
    def get_file_content(self, path):
 
        return u''
 
        return ''
 

	
 
    def get_file_size(self, path):
 
        return 0
 

	
 

	
 
class CollectionGenerator(object):
 

	
 
    def __init__(self, repo, revs):
 
        self.repo = repo
 
        self.revs = revs
 

	
 
    def __len__(self):
 
        return len(self.revs)
 

	
 
    def __iter__(self):
 
        for rev in self.revs:
 
            yield self.repo.get_changeset(rev)
 

	
 
    def __getitem__(self, what):
 
        """Return either a single element by index, or a sliced collection."""
 
        if isinstance(what, slice):
 
            return CollectionGenerator(self.repo, self.revs[what])
 
        else:
 
            # single item
 
            return self.repo.get_changeset(self.revs[what])
 

	
 
    def __repr__(self):
 
        return '<CollectionGenerator[len:%s]>' % (len(self))
kallithea/lib/vcs/backends/git/repository.py
Show inline comments
 
# -*- 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 errno
 
import logging
 
import os
 
import re
 
import time
 
import urllib.error
 
import urllib.parse
 
import urllib.request
 
from collections import OrderedDict
 

	
 
import mercurial.url  # import httpbasicauthhandler, httpdigestauthhandler
 
import mercurial.util  # import url as hg_url
 
from dulwich.config import ConfigFile
 
from dulwich.objects import Tag
 
from dulwich.repo import NotGitRepository, Repo
 

	
 
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 ascii_str, date_fromtimestamp, makedate, safe_bytes, safe_str
 
from kallithea.lib.vcs.utils.lazy import LazyProperty
 
from kallithea.lib.vcs.utils.paths import abspath, get_user_home
 

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

	
 
    @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, cwd=None):
 
        """
 
        Runs given ``cmd`` as git command and returns output bytes in a tuple
 
        (stdout, stderr) ... or raise RepositoryError.
 

	
 
        :param cmd: git command to be executed
 
        :param cwd: passed directly to subprocess
 
        """
 
        # need to clean fix GIT_DIR !
 
        gitenv = dict(os.environ)
 
        gitenv.pop('GIT_DIR', None)
 
        gitenv['GIT_CONFIG_NOGLOBAL'] = '1'
 

	
 
        assert isinstance(cmd, list), cmd
 
        cmd = [settings.GIT_EXECUTABLE_PATH, '-c', 'core.quotepath=false'] + cmd
 
        try:
 
            p = subprocessio.SubprocessIOChunker(cmd, cwd=cwd, env=gitenv, shell=False)
 
        except (EnvironmentError, OSError) as err:
 
            # output from the failing process is in str(EnvironmentError)
 
            msg = ("Couldn't run git command %s.\n"
 
                   "Subprocess failed with '%s': %s\n" %
 
                   (cmd, type(err).__name__, err)
 
            ).strip()
 
            log.error(msg)
 
            raise RepositoryError(msg)
 

	
 
        try:
 
            stdout = b''.join(p.output)
 
            stderr = b''.join(p.error)
 
        finally:
 
            p.close()
 
        # TODO: introduce option to make commands fail if they have any stderr output?
 
        if stderr:
 
            log.debug('stderr from %s:\n%s', cmd, stderr)
 
        else:
 
            log.debug('stderr from %s: None', cmd)
 
        return stdout, stderr
 

	
 
    def run_git_command(self, cmd):
 
        """
 
        Runs given ``cmd`` as git command with cwd set to current repo.
 
        Returns stdout as unicode str ... or raise RepositoryError.
 
        """
 
        cwd = None
 
        if os.path.isdir(self.path):
 
            cwd = self.path
 
        stdout, _stderr = self._run_git_command(cmd, cwd=cwd)
 
        return safe_str(stdout)
 

	
 
    @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 url.startswith('git://'):
 
            return True
 

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

	
 
        handlers = []
 
        url_obj = mercurial.util.url(safe_bytes(url))
 
        test_uri, authinfo = url_obj.authinfo()
 
        if not test_uri.endswith('info/refs'):
 
            test_uri = test_uri.rstrip('/') + '/info/refs'
 

	
 
        url_obj.passwd = b'*****'
 
        cleaned_uri = str(url_obj)
 

	
 
        if authinfo:
 
            # create a password manager
 
            passmgr = urllib.request.HTTPPasswordMgrWithDefaultRealm()
 
            passmgr.add_password(*authinfo)
 

	
 
            handlers.extend((mercurial.url.httpbasicauthhandler(passmgr),
 
                             mercurial.url.httpdigestauthhandler(passmgr)))
 

	
 
        o = urllib.request.build_opener(*handlers)
 
        o.addheaders = [('User-Agent', 'git/1.7.8.0')]  # fake some git
 

	
 
        req = urllib.request.Request(
 
            "%s?%s" % (
 
                test_uri,
 
                urllib.parse.urlencode({"service": 'git-upload-pack'})
 
            ))
 

	
 
        try:
 
            resp = o.open(req)
 
            if resp.code != 200:
 
                raise Exception('Return Code is not 200')
 
        except Exception as e:
 
            # means it cannot be cloned
 
            raise urllib.error.URLError("[%s] org_exc: %s" % (cleaned_uri, e))
 

	
 
        # now detect if it's proper git repo
 
        gitdata = resp.read()
 
        if 'service=git-upload-pack' not in gitdata:
 
            raise urllib.error.URLError(
 
                "url [%s] does not look like an git" % cleaned_uri)
 

	
 
        return True
 

	
 
    def _get_repo(self, create, src_url=None, update_after_clone=False,
 
                  bare=False):
 
        if create and os.path.exists(self.path):
 
            raise RepositoryError("Location already exist")
 
        if src_url and not create:
 
            raise RepositoryError("Create should be set to True if src_url is "
 
                                  "given (clone operation creates repository)")
 
        try:
 
            if create and src_url:
 
                GitRepository._check_url(src_url)
 
                self.clone(src_url, update_after_clone, bare)
 
                return Repo(self.path)
 
            elif create:
 
                os.makedirs(self.path)
 
                if bare:
 
                    return Repo.init_bare(self.path)
 
                else:
 
                    return Repo.init(self.path)
 
            else:
 
                return Repo(self.path)
 
        except (NotGitRepository, OSError) as err:
 
            raise RepositoryError(err)
 

	
 
    def _get_all_revisions(self):
 
        # we must check if this repo is not empty, since later command
 
        # fails if it is. And it's cheaper to ask than throw the subprocess
 
        # errors
 
        try:
 
            self._repo.head()
 
        except KeyError:
 
            return []
 

	
 
        rev_filter = settings.GIT_REV_FILTER
 
        cmd = ['rev-list', rev_filter, '--reverse', '--date-order']
 
        try:
 
            so = self.run_git_command(cmd)
 
        except RepositoryError:
 
            # Can be raised for empty repositories
 
            return []
 
        return so.splitlines()
 

	
 
    def _get_all_revisions2(self):
 
        # alternate implementation using dulwich
 
        includes = [ascii_str(sha) for key, (sha, type_) in self._parsed_refs.items()
 
                    if type_ != b'T']
 
        return [c.commit.id for c in self._repo.get_walker(include=includes)]
 

	
 
    def _get_revision(self, revision):
 
        """
 
        Given any revision identifier, returns a 40 char string with revision hash.
 
        """
 
        if self._empty:
 
            raise EmptyRepositoryError("There are no changesets yet")
 

	
 
        if revision in (None, '', 'tip', 'HEAD', 'head', -1):
 
            revision = -1
 

	
 
        if isinstance(revision, int):
 
            try:
 
                return self.revisions[revision]
 
            except IndexError:
 
                msg = "Revision %r does not exist for %s" % (revision, self.name)
 
                raise ChangesetDoesNotExistError(msg)
 

	
 
        if isinstance(revision, str):
 
            if revision.isdigit() and (len(revision) < 12 or len(revision) == revision.count('0')):
 
                try:
 
                    return self.revisions[int(revision)]
 
                except IndexError:
 
                    msg = "Revision %r does not exist for %s" % (revision, self)
 
                    raise ChangesetDoesNotExistError(msg)
 

	
 
            # get by branch/tag name
 
            _ref_revision = self._parsed_refs.get(safe_bytes(revision))
 
            if _ref_revision:  # and _ref_revision[1] in [b'H', b'RH', b'T']:
 
                return ascii_str(_ref_revision[0])
 

	
 
            if revision in self.revisions:
 
                return revision
 

	
 
            # maybe it's a tag ? we don't have them in self.revisions
 
            if revision in self.tags.values():
 
                return revision
 

	
 
            if SHA_PATTERN.match(revision):
 
                msg = "Revision %r does not exist for %s" % (revision, self.name)
 
                raise ChangesetDoesNotExistError(msg)
 

	
 
        raise ChangesetDoesNotExistError("Given revision %r not recognized" % revision)
 

	
 
    def get_ref_revision(self, ref_type, ref_name):
 
        """
 
        Returns ``GitChangeset`` object representing repository's
 
        changeset at the given ``revision``.
 
        """
 
        return self._get_revision(ref_name)
 

	
 
    def _get_archives(self, archive_name='tip'):
 

	
 
        for i in [('zip', '.zip'), ('gz', '.tar.gz'), ('bz2', '.tar.bz2')]:
 
            yield {"type": i[0], "extension": i[1], "node": archive_name}
 

	
 
    def _get_url(self, url):
 
        """
 
        Returns normalized url. If schema is not given, would fall to
 
        filesystem (``file:///``) schema.
 
        """
 
        if url != 'default' and '://' not in url:
 
            url = ':///'.join(('file', url))
 
        return url
 

	
 
    @LazyProperty
 
    def name(self):
 
        return os.path.basename(self.path)
 

	
 
    @LazyProperty
 
    def last_change(self):
 
        """
 
        Returns last change made on this repository as datetime object
 
        """
 
        return date_fromtimestamp(self._get_mtime(), makedate()[1])
 

	
 
    def _get_mtime(self):
 
        try:
 
            return time.mktime(self.get_changeset().date.timetuple())
 
        except RepositoryError:
 
            idx_loc = '' if self.bare else '.git'
 
            # fallback to filesystem
 
            in_path = os.path.join(self.path, idx_loc, "index")
 
            he_path = os.path.join(self.path, idx_loc, "HEAD")
 
            if os.path.exists(in_path):
 
                return os.stat(in_path).st_mtime
 
            else:
 
                return os.stat(he_path).st_mtime
 

	
 
    @LazyProperty
 
    def description(self):
 
        return safe_str(self._repo.get_description() or b'unknown')
 

	
 
    @LazyProperty
 
    def contact(self):
 
        undefined_contact = u'Unknown'
 
        undefined_contact = 'Unknown'
 
        return undefined_contact
 

	
 
    @property
 
    def branches(self):
 
        if not self.revisions:
 
            return {}
 
        sortkey = lambda ctx: ctx[0]
 
        _branches = [(safe_str(key), ascii_str(sha))
 
                     for key, (sha, type_) in self._parsed_refs.items() if type_ == 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 = [(safe_str(key), ascii_str(sha))
 
                 for key, (sha, type_) in self._parsed_refs.items() if type_ == 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" % safe_bytes(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(safe_str(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):
 
        """Return refs as a dict, like:
 
        { b'v0.2.0': [b'599ba911aa24d2981225f3966eb659dfae9e9f30', b'T'] }
 
        """
 
        _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.items():
 
            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.items())
 

	
 
    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
 
        return GitChangeset(repository=self, revision=self._get_revision(revision))
 

	
 
    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
 

	
 
        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.reverse()
 

	
 
        return CollectionGenerator(self, revs)
 

	
 
    def get_diff(self, rev1, rev2, path=None, ignore_whitespace=False,
 
                 context=3):
 
        """
 
        Returns (git like) *diff*, as plain bytes 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``. Due to limitations in Git, if
 
          value passed-in is greater than ``2**31-1``
 
          (``2147483647``), it will be set to ``2147483647``
 
          instead. If negative value is passed-in, it will be set to
 
          ``0`` instead.
 
        """
 

	
 
        # Git internally uses a signed long int for storing context
 
        # size (number of lines to show before and after the
 
        # differences). This can result in integer overflow, so we
 
        # ensure the requested context is smaller by one than the
 
        # number that would cause the overflow. It is highly unlikely
 
        # that a single file will contain that many lines, so this
 
        # kind of change should not cause any realistic consequences.
 
        overflowed_long_int = 2**31
 

	
 
        if context >= overflowed_long_int:
 
            context = overflowed_long_int - 1
 

	
 
        # Negative context values make no sense, and will result in
 
        # errors. Ensure this does not happen.
 
        if context < 0:
 
            context = 0
 

	
 
        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 = ['show'] + flags + [rev2]
 
        else:
 
            rev1 = self.get_changeset(rev1).raw_id
 
            rev2 = self.get_changeset(rev2).raw_id
 
            cmd = ['diff'] + flags + [rev1, rev2]
 

	
 
        if path:
 
            cmd += ['--', path]
 

	
 
        stdout, stderr = self._run_git_command(cmd, cwd=self.path)
 
        # If we used 'show' command, strip first few lines (until actual diff
 
        # starts)
 
        if rev1 == self.EMPTY_CHANGESET:
 
            parts = stdout.split(b'\ndiff ', 1)
 
            if len(parts) > 1:
 
                stdout = b'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 += ['--', url, self.path]
 
        # 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', url]
 
        # 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 = self.run_git_command(['ls-remote', '-h', url])
 
        cmd = ['fetch', url, '--']
 
        for line in (x for x in so.splitlines()):
 
            sha, ref = line.split('\t')
 
            cmd.append('+%s:%s' % (ref, ref))
 
        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
 
        try:
 
            update_server_info(self._repo)
 
        except OSError as e:
 
            if e.errno not in [errno.ENOENT, errno.EROFS]:
 
                raise
 
            # Workaround for dulwich crashing on for example its own dulwich/tests/data/repos/simple_merge.git/info/refs.lock
 
            log.error('Ignoring %s running update-server-info: %s', type(e).__name__, e)
 

	
 
    @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, str):
 
            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:
 
                value = config.get(section, name)
 
            except KeyError:
 
                continue
 
            return None if value is None else safe_str(value)
 
        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)
kallithea/lib/vcs/nodes.py
Show inline comments
 
# -*- coding: utf-8 -*-
 
"""
 
    vcs.nodes
 
    ~~~~~~~~~
 

	
 
    Module holding everything related to vcs nodes.
 

	
 
    :created_on: Apr 8, 2010
 
    :copyright: (c) 2010-2011 by Marcin Kuzminski, Lukasz Balcerzak.
 
"""
 

	
 
import functools
 
import mimetypes
 
import posixpath
 
import stat
 

	
 
from kallithea.lib.vcs.backends.base import EmptyChangeset
 
from kallithea.lib.vcs.exceptions import NodeError, RemovedFileNodeError
 
from kallithea.lib.vcs.utils import safe_bytes, safe_str
 
from kallithea.lib.vcs.utils.lazy import LazyProperty
 

	
 

	
 
class NodeKind:
 
    SUBMODULE = -1
 
    DIR = 1
 
    FILE = 2
 

	
 

	
 
class NodeState:
 
    ADDED = u'added'
 
    CHANGED = u'changed'
 
    NOT_CHANGED = u'not changed'
 
    REMOVED = u'removed'
 
    ADDED = 'added'
 
    CHANGED = 'changed'
 
    NOT_CHANGED = 'not changed'
 
    REMOVED = 'removed'
 

	
 

	
 
class NodeGeneratorBase(object):
 
    """
 
    Base class for removed added and changed filenodes, it's a lazy generator
 
    class that will create filenodes only on iteration or call
 

	
 
    The len method doesn't need to create filenodes at all
 
    """
 

	
 
    def __init__(self, current_paths, cs):
 
        self.cs = cs
 
        self.current_paths = current_paths
 

	
 
    def __getitem__(self, key):
 
        assert isinstance(key, slice), key
 
        for p in self.current_paths[key]:
 
            yield self.cs.get_node(p)
 

	
 
    def __len__(self):
 
        return len(self.current_paths)
 

	
 
    def __iter__(self):
 
        for p in self.current_paths:
 
            yield self.cs.get_node(p)
 

	
 

	
 
class AddedFileNodesGenerator(NodeGeneratorBase):
 
    """
 
    Class holding Added files for current changeset
 
    """
 
    pass
 

	
 

	
 
class ChangedFileNodesGenerator(NodeGeneratorBase):
 
    """
 
    Class holding Changed files for current changeset
 
    """
 
    pass
 

	
 

	
 
class RemovedFileNodesGenerator(NodeGeneratorBase):
 
    """
 
    Class holding removed files for current changeset
 
    """
 
    def __iter__(self):
 
        for p in self.current_paths:
 
            yield RemovedFileNode(path=p)
 

	
 
    def __getitem__(self, key):
 
        assert isinstance(key, slice), key
 
        for p in self.current_paths[key]:
 
            yield RemovedFileNode(path=p)
 

	
 

	
 
@functools.total_ordering
 
class Node(object):
 
    """
 
    Simplest class representing file or directory on repository.  SCM backends
 
    should use ``FileNode`` and ``DirNode`` subclasses rather than ``Node``
 
    directly.
 

	
 
    Node's ``path`` cannot start with slash as we operate on *relative* paths
 
    only. Moreover, every single node is identified by the ``path`` attribute,
 
    so it cannot end with slash, too. Otherwise, path could lead to mistakes.
 
    """
 

	
 
    def __init__(self, path, kind):
 
        if path.startswith('/'):
 
            raise NodeError("Cannot initialize Node objects with slash at "
 
                            "the beginning as only relative paths are supported")
 
        self.path = path.rstrip('/')
 
        if path == '' and kind != NodeKind.DIR:
 
            raise NodeError("Only DirNode and its subclasses may be "
 
                            "initialized with empty path")
 
        self.kind = kind
 
        #self.dirs, self.files = [], []
 
        if self.is_root() and not self.is_dir():
 
            raise NodeError("Root node cannot be FILE kind")
 

	
 
    @LazyProperty
 
    def parent(self):
 
        parent_path = self.get_parent_path()
 
        if parent_path:
 
            if self.changeset:
 
                return self.changeset.get_node(parent_path)
 
            return DirNode(parent_path)
 
        return None
 

	
 
    @LazyProperty
 
    def name(self):
 
        """
 
        Returns name of the node so if its path
 
        then only last part is returned.
 
        """
 
        return self.path.rstrip('/').split('/')[-1]
 

	
 
    def _get_kind(self):
 
        return self._kind
 

	
 
    def _set_kind(self, kind):
 
        if hasattr(self, '_kind'):
 
            raise NodeError("Cannot change node's kind")
 
        else:
 
            self._kind = kind
 
            # Post setter check (path's trailing slash)
 
            if self.path.endswith('/'):
 
                raise NodeError("Node's path cannot end with slash")
 

	
 
    kind = property(_get_kind, _set_kind)
 

	
 
    def __eq__(self, other):
 
        if type(self) is not type(other):
 
            return False
 
        if self._kind != other._kind:
 
            return False
 
        if self.path != other.path:
 
            return False
 
        if self.is_file():
 
            return self.content == other.content
 
        else:
 
            # For DirNode's check without entering each dir
 
            self_nodes_paths = list(sorted(n.path for n in self.nodes))
 
            other_nodes_paths = list(sorted(n.path for n in self.nodes))
 
            return self_nodes_paths == other_nodes_paths
 

	
 
    def __lt__(self, other):
 
        if self._kind < other._kind:
 
            return True
 
        if self._kind > other._kind:
 
            return False
 
        if self.path < other.path:
 
            return True
 
        if self.path > other.path:
 
            return False
 
        if self.is_file():
 
            return self.content < other.content
 
        else:
 
            # For DirNode's check without entering each dir
 
            self_nodes_paths = list(sorted(n.path for n in self.nodes))
 
            other_nodes_paths = list(sorted(n.path for n in self.nodes))
 
            return self_nodes_paths < other_nodes_paths
 

	
 
    def __repr__(self):
 
        return '<%s %r>' % (self.__class__.__name__, self.path)
 

	
 
    def get_parent_path(self):
 
        """
 
        Returns node's parent path or empty string if node is root.
 
        """
 
        if self.is_root():
 
            return ''
 
        return posixpath.dirname(self.path.rstrip('/')) + '/'
 

	
 
    def is_file(self):
 
        """
 
        Returns ``True`` if node's kind is ``NodeKind.FILE``, ``False``
 
        otherwise.
 
        """
 
        return self.kind == NodeKind.FILE
 

	
 
    def is_dir(self):
 
        """
 
        Returns ``True`` if node's kind is ``NodeKind.DIR``, ``False``
 
        otherwise.
 
        """
 
        return self.kind == NodeKind.DIR
 

	
 
    def is_root(self):
 
        """
 
        Returns ``True`` if node is a root node and ``False`` otherwise.
 
        """
 
        return self.kind == NodeKind.DIR and self.path == ''
 

	
 
    def is_submodule(self):
 
        """
 
        Returns ``True`` if node's kind is ``NodeKind.SUBMODULE``, ``False``
 
        otherwise.
 
        """
 
        return self.kind == NodeKind.SUBMODULE
 

	
 
    @LazyProperty
 
    def added(self):
 
        return self.state is NodeState.ADDED
 

	
 
    @LazyProperty
 
    def changed(self):
 
        return self.state is NodeState.CHANGED
 

	
 
    @LazyProperty
 
    def not_changed(self):
 
        return self.state is NodeState.NOT_CHANGED
 

	
 
    @LazyProperty
 
    def removed(self):
 
        return self.state is NodeState.REMOVED
 

	
 

	
 
class FileNode(Node):
 
    """
 
    Class representing file nodes.
 

	
 
    :attribute: path: path to the node, relative to repository's root
 
    :attribute: content: if given arbitrary sets content of the file
 
    :attribute: changeset: if given, first time content is accessed, callback
 
    :attribute: mode: octal stat mode for a node. Default is 0100644.
 
    """
 

	
 
    def __init__(self, path, content=None, changeset=None, mode=None):
 
        """
 
        Only one of ``content`` and ``changeset`` may be given. Passing both
 
        would raise ``NodeError`` exception.
 

	
 
        :param path: relative path to the node
 
        :param content: content may be passed to constructor
 
        :param changeset: if given, will use it to lazily fetch content
 
        :param mode: octal representation of ST_MODE (i.e. 0100644)
 
        """
 

	
 
        if content and changeset:
 
            raise NodeError("Cannot use both content and changeset")
 
        super(FileNode, self).__init__(path, kind=NodeKind.FILE)
 
        self.changeset = changeset
 
        if not isinstance(content, bytes) and content is not None:
 
            # File content is one thing that inherently must be bytes ... but
 
            # VCS module tries to be "user friendly" and support unicode ...
 
            content = safe_bytes(content)
 
        self._content = content
 
        self._mode = mode or 0o100644
 

	
 
    @LazyProperty
 
    def mode(self):
 
        """
 
        Returns lazily mode of the FileNode. If ``changeset`` is not set, would
 
        use value given at initialization or 0100644 (default).
 
        """
 
        if self.changeset:
 
            mode = self.changeset.get_file_mode(self.path)
 
        else:
 
            mode = self._mode
 
        return mode
 

	
 
    @property
 
    def content(self):
 
        """
 
        Returns lazily byte content of the FileNode.
 
        """
 
        if self.changeset:
 
            content = self.changeset.get_file_content(self.path)
 
        else:
 
            content = self._content
 
        return content
 

	
 
    @LazyProperty
 
    def size(self):
 
        if self.changeset:
 
            return self.changeset.get_file_size(self.path)
 
        raise NodeError("Cannot retrieve size of the file without related "
 
            "changeset attribute")
 

	
 
    @LazyProperty
 
    def message(self):
 
        if self.changeset:
 
            return self.last_changeset.message
 
        raise NodeError("Cannot retrieve message of the file without related "
 
            "changeset attribute")
 

	
 
    @LazyProperty
 
    def last_changeset(self):
 
        if self.changeset:
 
            return self.changeset.get_file_changeset(self.path)
 
        raise NodeError("Cannot retrieve last changeset of the file without "
 
            "related changeset attribute")
 

	
 
    def get_mimetype(self):
 
        """
 
        Mimetype is calculated based on the file's content.
 
        """
 

	
 
        mtype, encoding = mimetypes.guess_type(self.name)
 

	
 
        if mtype is None:
 
            if self.is_binary:
 
                mtype = 'application/octet-stream'
 
                encoding = None
 
            else:
 
                mtype = 'text/plain'
 
                encoding = None
 

	
 
                # try with pygments
 
                from pygments import lexers
 
                try:
 
                    mt = lexers.get_lexer_for_filename(self.name).mimetypes
 
                except lexers.ClassNotFound:
 
                    mt = None
 

	
 
                if mt:
 
                    mtype = mt[0]
 

	
 
        return mtype, encoding
 

	
 
    @LazyProperty
 
    def mimetype(self):
 
        """
 
        Wrapper around full mimetype info. It returns only type of fetched
 
        mimetype without the encoding part. use get_mimetype function to fetch
 
        full set of (type,encoding)
 
        """
 
        return self.get_mimetype()[0]
 

	
 
    @LazyProperty
 
    def mimetype_main(self):
 
        return self.mimetype.split('/')[0]
 

	
 
    @LazyProperty
 
    def lexer(self):
 
        """
 
        Returns pygment's lexer class. Would try to guess lexer taking file's
 
        content, name and mimetype.
 
        """
 
        from pygments import lexers
 
        try:
 
            lexer = lexers.guess_lexer_for_filename(self.name, safe_str(self.content), stripnl=False)
 
        except lexers.ClassNotFound:
 
            lexer = lexers.TextLexer(stripnl=False)
 
        # returns first alias
 
        return lexer
 

	
 
    @LazyProperty
 
    def lexer_alias(self):
 
        """
 
        Returns first alias of the lexer guessed for this file.
 
        """
 
        return self.lexer.aliases[0]
 

	
 
    @LazyProperty
 
    def history(self):
 
        """
 
        Returns a list of changeset for this file in which the file was changed
 
        """
 
        if self.changeset is None:
 
            raise NodeError('Unable to get changeset for this FileNode')
 
        return self.changeset.get_file_history(self.path)
 

	
 
    @LazyProperty
 
    def annotate(self):
 
        """
 
        Returns a list of three element tuples with lineno,changeset and line
 
        """
 
        if self.changeset is None:
 
            raise NodeError('Unable to get changeset for this FileNode')
 
        return self.changeset.get_file_annotate(self.path)
 

	
 
    @LazyProperty
 
    def state(self):
 
        if not self.changeset:
 
            raise NodeError("Cannot check state of the node if it's not "
 
                "linked with changeset")
 
        elif self.path in (node.path for node in self.changeset.added):
 
            return NodeState.ADDED
 
        elif self.path in (node.path for node in self.changeset.changed):
 
            return NodeState.CHANGED
 
        else:
 
            return NodeState.NOT_CHANGED
 

	
 
    @property
 
    def is_binary(self):
 
        """
 
        Returns True if file has binary content.
 
        """
 
        return b'\0' in self.content
 

	
 
    def is_browser_compatible_image(self):
 
        return self.mimetype in [
 
            "image/gif",
 
            "image/jpeg",
 
            "image/png",
 
            "image/bmp"
 
        ]
 

	
 
    @LazyProperty
 
    def extension(self):
 
        """Returns filenode extension"""
 
        return self.name.split('.')[-1]
 

	
 
    @property
 
    def is_executable(self):
 
        """
 
        Returns ``True`` if file has executable flag turned on.
 
        """
 
        return bool(self.mode & stat.S_IXUSR)
 

	
 
    def __repr__(self):
 
        return '<%s %r @ %s>' % (self.__class__.__name__, self.path,
 
                                 getattr(self.changeset, 'short_id', ''))
 

	
 

	
 
class RemovedFileNode(FileNode):
 
    """
 
    Dummy FileNode class - trying to access any public attribute except path,
 
    name, kind or state (or methods/attributes checking those two) would raise
 
    RemovedFileNodeError.
 
    """
 
    ALLOWED_ATTRIBUTES = [
 
        'name', 'path', 'state', 'is_root', 'is_file', 'is_dir', 'kind',
 
        'added', 'changed', 'not_changed', 'removed'
 
    ]
 

	
 
    def __init__(self, path):
 
        """
 
        :param path: relative path to the node
 
        """
 
        super(RemovedFileNode, self).__init__(path=path)
 

	
 
    def __getattribute__(self, attr):
 
        if attr.startswith('_') or attr in RemovedFileNode.ALLOWED_ATTRIBUTES:
 
            return super(RemovedFileNode, self).__getattribute__(attr)
 
        raise RemovedFileNodeError("Cannot access attribute %s on "
 
            "RemovedFileNode" % attr)
 

	
 
    @LazyProperty
 
    def state(self):
 
        return NodeState.REMOVED
 

	
 

	
 
class DirNode(Node):
 
    """
 
    DirNode stores list of files and directories within this node.
 
    Nodes may be used standalone but within repository context they
 
    lazily fetch data within same repository's changeset.
 
    """
 

	
 
    def __init__(self, path, nodes=(), changeset=None):
 
        """
 
        Only one of ``nodes`` and ``changeset`` may be given. Passing both
 
        would raise ``NodeError`` exception.
 

	
 
        :param path: relative path to the node
 
        :param nodes: content may be passed to constructor
 
        :param changeset: if given, will use it to lazily fetch content
 
        :param size: always 0 for ``DirNode``
 
        """
 
        if nodes and changeset:
 
            raise NodeError("Cannot use both nodes and changeset")
 
        super(DirNode, self).__init__(path, NodeKind.DIR)
 
        self.changeset = changeset
 
        self._nodes = nodes
 

	
 
    @LazyProperty
 
    def content(self):
 
        raise NodeError("%s represents a dir and has no ``content`` attribute"
 
            % self)
 

	
 
    @LazyProperty
 
    def nodes(self):
 
        if self.changeset:
 
            nodes = self.changeset.get_nodes(self.path)
 
        else:
 
            nodes = self._nodes
 
        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):
 
        # Note: Doesn't call Node.__init__!
 
        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(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 = self.path.rstrip('/').rsplit('/', 1)[-1]
 
        return u'%s @ %s' % (org, self.changeset.short_id)
 
        return '%s @ %s' % (org, self.changeset.short_id)

Changeset was too big and was cut off... Show full diff anyway

0 comments (0 inline, 0 general)