Bring back the separate web-side entity cache
Prior to changeset 635cfac73d28 "[repoapi] fold ClientConnection into
Connection", we had two entity caches: one on the client/dbapi/web side,
and one on the server/repo side. This is a waste, but it is actually
needed as long as we have the magic _cw attribute on entities which must
sometimes be a request and sometimes a cnx. Removing the duplication
caused weird problems with entity._cw alternating between both types of
objects, which is unexpected by both the repo and the web sides.
We add an entity cache on ConnectionCubicWebRequestBase, separate from
the Connection's, and bring back the
_cw_update_attr_cache/dont-cache-attrs mechanism, to let the
server/edition code let caches know which attributes have been modified
Entity.as_rset can be cached again, as ResultSet no longer modifies the
entity when fetching it from a cache.
Contrary to the pre-3.21 code, _cw_update_attr_cache now handles web
requests and connections in the same way (otherwise the cache ends up
with wrong values if a hook modifies attributes), but dont-cache-attrs
is never set for (inlined) relations.
Closes #6863543
# copyright 2003-2013 LOGILAB S.A. (Paris, FRANCE), all rights reserved.
# contact http://www.logilab.fr/ -- mailto:contact@logilab.fr
#
# This file is part of CubicWeb.
#
# CubicWeb is free software: you can redistribute it and/or modify it under the
# terms of the GNU Lesser General Public License as published by the Free
# Software Foundation, either version 2.1 of the License, or (at your option)
# any later version.
#
# CubicWeb 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 Lesser General Public License for more
# details.
#
# You should have received a copy of the GNU Lesser General Public License along
# with CubicWeb. If not, see <http://www.gnu.org/licenses/>.
"""rest publishing functions
contains some functions and setup of docutils for cubicweb. Provides the
following ReST directives:
* `eid`, create link to entity in the repository by their eid
* `card`, create link to card entity in the repository by their wikiid
(proposing to create it when the refered card doesn't exist yet)
* `winclude`, reference to a web documentation file (in wdoc/ directories)
* `sourcecode` (if pygments is installed), source code colorization
* `rql-table`, create a table from a RQL query
"""
__docformat__ = "restructuredtext en"
import sys
from itertools import chain
from logging import getLogger
from os.path import join
from six import text_type
from six.moves.urllib.parse import urlsplit
from docutils import statemachine, nodes, utils, io
from docutils.core import Publisher
from docutils.parsers.rst import Parser, states, directives, Directive
from docutils.parsers.rst.roles import register_canonical_role, set_classes
from logilab.mtconverter import ESC_UCAR_TABLE, ESC_CAR_TABLE, xml_escape
from cubicweb import UnknownEid
from cubicweb.ext.html4zope import Writer
from cubicweb.web.views import vid_from_rset # XXX better not to import c.w.views here...
# We provide our own parser as an attempt to get rid of
# state machine reinstanciation
import re
# compile states.Body patterns
for k, v in states.Body.patterns.items():
if isinstance(v, str):
states.Body.patterns[k] = re.compile(v)
# register ReStructured Text mimetype / extensions
import mimetypes
mimetypes.add_type('text/rest', '.rest')
mimetypes.add_type('text/rest', '.rst')
LOGGER = getLogger('cubicweb.rest')
def eid_reference_role(role, rawtext, text, lineno, inliner,
options={}, content=[]):
try:
try:
eid_num, rest = text.split(u':', 1)
except ValueError:
eid_num, rest = text, '#'+text
eid_num = int(eid_num)
if eid_num < 0:
raise ValueError
except ValueError:
msg = inliner.reporter.error(
'EID number must be a positive number; "%s" is invalid.'
% text, line=lineno)
prb = inliner.problematic(rawtext, rawtext, msg)
return [prb], [msg]
# Base URL mainly used by inliner.pep_reference; so this is correct:
context = inliner.document.settings.context
try:
refedentity = context._cw.entity_from_eid(eid_num)
except UnknownEid:
ref = '#'
rest += u' ' + context._cw._('(UNEXISTANT EID)')
else:
ref = refedentity.absolute_url()
set_classes(options)
return [nodes.reference(rawtext, utils.unescape(rest), refuri=ref,
**options)], []
def rql_role(role, rawtext, text, lineno, inliner, options={}, content=[]):
"""``:rql:`<rql-expr>``` or ``:rql:`<rql-expr>:<vid>```
Example: ``:rql:`Any X,Y WHERE X is CWUser, X login Y:table```
Replace the directive with the output of applying the view to the resultset
returned by the query.
"X eid %(userid)s" can be used in the RQL query for this query will be
executed with the argument {'userid': _cw.user.eid}.
"""
_cw = inliner.document.settings.context._cw
text = text.strip()
if ':' in text:
rql, vid = text.rsplit(u':', 1)
rql = rql.strip()
else:
rql, vid = text, None
_cw.ensure_ro_rql(rql)
try:
rset = _cw.execute(rql, {'userid': _cw.user.eid})
if rset:
if vid is None:
vid = vid_from_rset(_cw, rset, _cw.vreg.schema)
else:
vid = 'noresult'
view = _cw.vreg['views'].select(vid, _cw, rset=rset)
content = view.render()
except Exception as exc:
content = 'an error occurred while interpreting this rql directive: %r' % exc
set_classes(options)
return [nodes.raw('', content, format='html')], []
def bookmark_role(role, rawtext, text, lineno, inliner, options={}, content=[]):
"""``:bookmark:`<bookmark-eid>``` or ``:bookmark:`<eid>:<vid>```
Example: ``:bookmark:`1234:table```
Replace the directive with the output of applying the view to the resultset
returned by the query stored in the bookmark. By default, the view is the one
stored in the bookmark, but it can be overridden by the directive as in the
example above.
"X eid %(userid)s" can be used in the RQL query stored in the Bookmark, for
this query will be executed with the argument {'userid': _cw.user.eid}.
"""
_cw = inliner.document.settings.context._cw
text = text.strip()
try:
if ':' in text:
eid, vid = text.rsplit(u':', 1)
eid = int(eid)
else:
eid, vid = int(text), None
except ValueError:
msg = inliner.reporter.error(
'EID number must be a positive number; "%s" is invalid.'
% text, line=lineno)
prb = inliner.problematic(rawtext, rawtext, msg)
return [prb], [msg]
try:
bookmark = _cw.entity_from_eid(eid)
except UnknownEid:
msg = inliner.reporter.error('Unknown EID %s.' % text, line=lineno)
prb = inliner.problematic(rawtext, rawtext, msg)
return [prb], [msg]
try:
params = dict(_cw.url_parse_qsl(urlsplit(bookmark.path).query))
rql = params['rql']
if vid is None:
vid = params.get('vid')
except (ValueError, KeyError) as exc:
msg = inliner.reporter.error('Could not parse bookmark path %s [%s].'
% (bookmark.path, exc), line=lineno)
prb = inliner.problematic(rawtext, rawtext, msg)
return [prb], [msg]
try:
rset = _cw.execute(rql, {'userid': _cw.user.eid})
if rset:
if vid is None:
vid = vid_from_rset(_cw, rset, _cw.vreg.schema)
else:
vid = 'noresult'
view = _cw.vreg['views'].select(vid, _cw, rset=rset)
content = view.render()
except Exception as exc:
content = 'An error occurred while interpreting directive bookmark: %r' % exc
set_classes(options)
return [nodes.raw('', content, format='html')], []
def winclude_directive(name, arguments, options, content, lineno,
content_offset, block_text, state, state_machine):
"""Include a reST file as part of the content of this reST file.
same as standard include directive but using config.locate_doc_resource to
get actual file to include.
Most part of this implementation is copied from `include` directive defined
in `docutils.parsers.rst.directives.misc`
"""
context = state.document.settings.context
cw = context._cw
source = state_machine.input_lines.source(
lineno - state_machine.input_offset - 1)
#source_dir = os.path.dirname(os.path.abspath(source))
fid = arguments[0]
for lang in chain((cw.lang, cw.vreg.property_value('ui.language')),
cw.vreg.config.available_languages()):
rid = '%s_%s.rst' % (fid, lang)
resourcedir = cw.vreg.config.locate_doc_file(rid)
if resourcedir:
break
else:
severe = state_machine.reporter.severe(
'Problems with "%s" directive path:\nno resource matching %s.'
% (name, fid),
nodes.literal_block(block_text, block_text), line=lineno)
return [severe]
path = join(resourcedir, rid)
encoding = options.get('encoding', state.document.settings.input_encoding)
try:
state.document.settings.record_dependencies.add(path)
include_file = io.FileInput(
source_path=path, encoding=encoding,
error_handler=state.document.settings.input_encoding_error_handler,
handle_io_errors=None)
except IOError as error:
severe = state_machine.reporter.severe(
'Problems with "%s" directive path:\n%s: %s.'
% (name, error.__class__.__name__, error),
nodes.literal_block(block_text, block_text), line=lineno)
return [severe]
try:
include_text = include_file.read()
except UnicodeError as error:
severe = state_machine.reporter.severe(
'Problem with "%s" directive:\n%s: %s'
% (name, error.__class__.__name__, error),
nodes.literal_block(block_text, block_text), line=lineno)
return [severe]
if 'literal' in options:
literal_block = nodes.literal_block(include_text, include_text,
source=path)
literal_block.line = 1
return literal_block
else:
include_lines = statemachine.string2lines(include_text,
convert_whitespace=1)
state_machine.insert_input(include_lines, path)
return []
winclude_directive.arguments = (1, 0, 1)
winclude_directive.options = {'literal': directives.flag,
'encoding': directives.encoding}
class RQLTableDirective(Directive):
"""rql-table directive
Example:
.. rql-table::
:vid: mytable
:headers: , , progress
:colvids: 2=progress
Any X,U,X WHERE X is Project, X url U
All fields but the RQL string are optionnal. The ``:headers:`` option can
contain empty column names.
"""
required_arguments = 0
optional_arguments = 0
has_content= True
final_argument_whitespace = True
option_spec = {'vid': directives.unchanged,
'headers': directives.unchanged,
'colvids': directives.unchanged}
def run(self):
errid = "rql-table directive"
self.assert_has_content()
if self.arguments:
raise self.warning('%s does not accept arguments' % errid)
rql = ' '.join([l.strip() for l in self.content])
_cw = self.state.document.settings.context._cw
_cw.ensure_ro_rql(rql)
try:
rset = _cw.execute(rql)
except Exception as exc:
raise self.error("fail to execute RQL query in %s: %r" %
(errid, exc))
if not rset:
raise self.warning("empty result set")
vid = self.options.get('vid', 'table')
try:
view = _cw.vreg['views'].select(vid, _cw, rset=rset)
except Exception as exc:
raise self.error("fail to select '%s' view in %s: %r" %
(vid, errid, exc))
headers = None
if 'headers' in self.options:
headers = [h.strip() for h in self.options['headers'].split(',')]
while headers.count(''):
headers[headers.index('')] = None
if len(headers) != len(rset[0]):
raise self.error("the number of 'headers' does not match the "
"number of columns in %s" % errid)
cellvids = None
if 'colvids' in self.options:
cellvids = {}
for f in self.options['colvids'].split(','):
try:
idx, vid = f.strip().split('=')
except ValueError:
raise self.error("malformatted 'colvids' option in %s" %
errid)
cellvids[int(idx.strip())] = vid.strip()
try:
content = view.render(headers=headers, cellvids=cellvids)
except Exception as exc:
raise self.error("Error rendering %s (%s)" % (errid, exc))
return [nodes.raw('', content, format='html')]
try:
from pygments import highlight
from pygments.lexers import get_lexer_by_name
from pygments.formatters.html import HtmlFormatter
except ImportError:
pygments_directive = None
else:
_PYGMENTS_FORMATTER = HtmlFormatter()
def pygments_directive(name, arguments, options, content, lineno,
content_offset, block_text, state, state_machine):
try:
lexer = get_lexer_by_name(arguments[0])
except ValueError:
# no lexer found
lexer = get_lexer_by_name('text')
parsed = highlight(u'\n'.join(content), lexer, _PYGMENTS_FORMATTER)
# don't fail if no context set on the sourcecode directive
try:
context = state.document.settings.context
context._cw.add_css('pygments.css')
except AttributeError:
# used outside cubicweb XXX use hasattr instead
pass
return [nodes.raw('', parsed, format='html')]
pygments_directive.arguments = (1, 0, 1)
pygments_directive.content = 1
class CubicWebReSTParser(Parser):
"""The (customized) reStructuredText parser."""
def __init__(self):
self.initial_state = 'Body'
self.state_classes = states.state_classes
self.inliner = states.Inliner()
self.statemachine = states.RSTStateMachine(
state_classes=self.state_classes,
initial_state=self.initial_state,
debug=0)
def parse(self, inputstring, document):
"""Parse `inputstring` and populate `document`, a document tree."""
self.setup_parse(inputstring, document)
inputlines = statemachine.string2lines(inputstring,
convert_whitespace=1)
self.statemachine.run(inputlines, document, inliner=self.inliner)
self.finish_parse()
# XXX docutils keep a ref on context, can't find a correct way to remove it
class CWReSTPublisher(Publisher):
def __init__(self, context, settings, **kwargs):
Publisher.__init__(self, **kwargs)
self.set_components('standalone', 'restructuredtext', 'pseudoxml')
self.process_programmatic_settings(None, settings, None)
self.settings.context = context
def rest_publish(context, data):
"""publish a string formatted as ReStructured Text to HTML
:type context: a cubicweb application object
:type data: str
:param data: some ReST text
:rtype: unicode
:return:
the data formatted as HTML or the original data if an error occurred
"""
req = context._cw
if isinstance(data, text_type):
encoding = 'unicode'
# remove unprintable characters unauthorized in xml
data = data.translate(ESC_UCAR_TABLE)
else:
encoding = req.encoding
# remove unprintable characters unauthorized in xml
data = data.translate(ESC_CAR_TABLE)
settings = {'input_encoding': encoding, 'output_encoding': 'unicode',
'warning_stream': False,
'traceback': True, # don't sys.exit
'stylesheet': None, # don't try to embed stylesheet (may cause
# obscure bug due to docutils computing
# relative path according to the directory
# used *at import time*
# dunno what's the max, severe is 4, and we never want a crash
# (though try/except may be a better option...). May be the
# above traceback option will avoid this?
'halt_level': 10,
# disable stupid switch to colspan=2 if field name is above a size limit
'field_name_limit': sys.maxsize,
}
if context:
if hasattr(req, 'url'):
base_url = req.url()
elif hasattr(context, 'absolute_url'):
base_url = context.absolute_url()
else:
base_url = req.base_url()
else:
base_url = None
try:
pub = CWReSTPublisher(context, settings,
parser=CubicWebReSTParser(),
writer=Writer(base_url=base_url),
source_class=io.StringInput,
destination_class=io.StringOutput)
pub.set_source(data)
pub.set_destination()
res = pub.publish(enable_exit_status=None)
# necessary for proper garbage collection, else a ref is kept somewhere in docutils...
del pub.settings.context
return res
except BaseException:
LOGGER.exception('error while publishing ReST text')
if not isinstance(data, text_type):
data = text_type(data, encoding, 'replace')
return xml_escape(req._('error while publishing ReST text')
+ '\n\n' + data)
_INITIALIZED = False
def cw_rest_init():
global _INITIALIZED
if _INITIALIZED:
return
_INITIALIZED = True
register_canonical_role('eid', eid_reference_role)
register_canonical_role('rql', rql_role)
register_canonical_role('bookmark', bookmark_role)
directives.register_directive('winclude', winclude_directive)
if pygments_directive is not None:
directives.register_directive('sourcecode', pygments_directive)
directives.register_directive('rql-table', RQLTableDirective)