blob: 4740a508f4d7bef59bebeb5a929d65407341b0e0 [file] [log] [blame]
# Copyright 2012 The Chromium Authors
# Use of this source code is governed by a BSD-style license that can be
# found in the LICENSE file.
"""Miscellaneous node types.
import os.path
import re
import sys
from grit import constants
from grit import exception
from grit import util
from grit.extern import FP
from grit.node import base
from grit.node import message
from grit.node import node_io
# Python 3 doesn't have long() as int() works everywhere. But we really do need
# the long() behavior on Python 2 as our ids are much too large for int().
except NameError:
long = int
# RTL languages
# TODO(jennyz): remove this fixed set of RTL language array
# now that generic expand_variable code exists.
'ar', # Arabic
'fa', # Farsi
'iw', # Hebrew
'ks', # Kashmiri
'ku', # Kurdish
'ps', # Pashto
'ur', # Urdu
'yi', # Yiddish
def _GetIdMapKeyFromFilename(filename, src_root_dir):
new_grd_filename = filename
abs_grd_filename = os.path.abspath(filename)
if abs_grd_filename[:len(src_root_dir)] == src_root_dir:
new_grd_filename = abs_grd_filename[len(src_root_dir) + 1:]
new_grd_filename = abs_grd_filename
return new_grd_filename.replace('\\', '/')
def _ReadFirstIdsFromFile(filename, defines):
"""Read the starting resource id values from |filename|. We also
expand variables of the form <(FOO) based on defines passed in on
the command line.
Returns a tuple, the absolute path of SRCDIR followed by the
first_ids dictionary.
first_ids_dict = eval(util.ReadFile(filename, 'utf-8'))
src_root_dir = os.path.abspath(os.path.join(os.path.dirname(filename),
def ReplaceVariable(matchobj):
for key, value in defines.items():
if == key:
return value if len(value) > 0 else EMPTY_REPLACEMENT
renames = []
for grd_filename in first_ids_dict:
new_grd_filename = re.sub(r'<\(([A-Za-z_]+)\)', ReplaceVariable,
new_grd_filename = new_grd_filename.replace('\\', '/')
if new_grd_filename != grd_filename:
# Empty or failed replacements at the start of the path can cause the path
# to appear to be absolute, when it in fact represents the path from the
# current (output) directory. Replace these with '.' instead.
if (new_grd_filename.startswith(EMPTY_REPLACEMENT + '/')):
new_grd_filename = '.' + new_grd_filename[len(EMPTY_REPLACEMENT):]
# Any other empty replacements in the middle of the path can be left
# as-is.
new_grd_filename = new_grd_filename.replace(EMPTY_REPLACEMENT, '')
new_grd_filename = _GetIdMapKeyFromFilename(new_grd_filename,
renames.append((grd_filename, new_grd_filename))
for grd_filename, new_grd_filename in renames:
assert new_grd_filename not in first_ids_dict
first_ids_dict[new_grd_filename] = first_ids_dict[grd_filename]
return (src_root_dir, first_ids_dict)
def _ComputeIds(root, predetermined_tids):
"""Returns a dict of textual id -> numeric id for all nodes in root.
IDs are mostly assigned sequentially, but will vary based on:
* first_id node attribute (from first_ids_file)
* hash of textual id (if not first_id is defined)
* offset node attribute
* whether the textual id matches a system id
* whether the node generates its own ID via GetId()
predetermined_tids: Dict of textual id -> numeric id to use in return dict.
from grit.node import empty, include, misc, structure
ids = {} # Maps numeric id to textual id
tids = {} # Maps textual id to numeric id
id_reasons = {} # Maps numeric id to text id and a human-readable explanation
group = None
last_id = None
predetermined_ids = {value: key
for key, value in predetermined_tids.items()}
# For entries that set in their first_ids_file (generated .grd files):
# "META": {"sizes": {"includes": [50]}}
# Check that "sizes" was big enough.
def check_group_count():
if root.skip_validation_checks or group is None or last_id is None:
max_ids = group.attrs.get('max_ids')
if max_ids is not None:
first_id = int(group.attrs['first_id'])
num_ids = last_id - first_id + 1
if num_ids > int(max_ids):
msg = ('Generated .grd file used more IDs (%d) than were allocated '
'for it (%s) for type %s. You need to update %s '
'for this .grd.') % (num_ids, max_ids,,
raise exception.IdRangeOverflow(msg)
for item in root:
if isinstance(item, empty.GroupingNode):
# Note: this won't work if any GroupingNode can be contained inside
# another.
group = item
last_id = None
assert not item.GetTextualIds() or isinstance(item,
(include.IncludeNode, message.MessageNode,
misc.IdentifierNode, structure.StructureNode))
# Resources that use the RES protocol don't need
# any numerical ids generated, so we skip them altogether.
# This is accomplished by setting the flag 'generateid' to false
# in the GRD file.
if item.attrs.get('generateid', 'true') == 'false':
for tid in item.GetTextualIds():
if util.SYSTEM_IDENTIFIERS.match(tid):
# Don't emit a new ID for predefined IDs
if tid in tids:
if predetermined_tids and tid in predetermined_tids:
id = predetermined_tids[tid]
reason = "from predetermined_tids map"
# Some identifier nodes can provide their own id,
# and we use that id in the generated header in that case.
elif hasattr(item, 'GetId') and item.GetId():
id = long(item.GetId())
reason = 'returned by GetId() method'
elif ('offset' in item.attrs and group and
group.attrs.get('first_id', '') != ''):
offset_text = item.attrs['offset']
parent_text = group.attrs['first_id']
offset_id = long(offset_text)
except ValueError:
offset_id = tids[offset_text]
parent_id = long(parent_text)
except ValueError:
parent_id = tids[parent_text]
id = parent_id + offset_id
reason = 'first_id %d + offset %d' % (parent_id, offset_id)
# We try to allocate IDs sequentially for blocks of items that might
# be related, for instance strings in a stringtable (as their IDs might be
# used e.g. as IDs for some radio buttons, in which case the IDs must
# be sequential).
# We do this by having the first item in a section store its computed ID
# (computed from a fingerprint) in its parent object. Subsequent children
# of the same parent will then try to get IDs that sequentially follow
# the currently stored ID (on the parent) and increment it.
elif last_id is None:
# First check if the starting ID is explicitly specified by the parent.
if group and group.attrs.get('first_id', '') != '':
id = long(group.attrs['first_id'])
reason = "from parent's first_id attribute"
# Automatically generate the ID based on the first clique from the
# first child of the first child node of our parent (i.e. when we
# first get to this location in the code).
# According to
# the safe usable range for resource IDs in Windows is from decimal
# 101 to 0x7FFF.
id = FP.UnsignedFingerPrint(tid)
id = id % (0x7FFF - 101) + 101
reason = 'chosen by random fingerprint -- use first_id to override'
last_id = id
id = last_id = last_id + 1
reason = 'sequentially assigned'
reason = "%s (%s)" % (tid, reason)
# Don't fail when 'offset' is specified, as the base and the 0th
# offset will have the same ID.
if id in id_reasons and not 'offset' in item.attrs:
raise exception.IdRangeOverlap('ID %d was assigned to both %s and %s.'
% (id, id_reasons[id], reason))
if id < 101:
print('WARNING: Numeric resource IDs should be greater than 100 to\n'
'avoid conflicts with system-defined resource IDs.')
if tid not in predetermined_tids and id in predetermined_ids:
raise exception.IdRangeOverlap('ID %d overlaps between %s and %s' %
(id, tid, predetermined_ids[id]))
ids[id] = tid
tids[tid] = id
id_reasons[id] = reason
return tids
class SplicingNode(base.Node):
"""A node whose children should be considered to be at the same level as
its siblings for most purposes. This includes <if> and <part> nodes.
def _IsValidChild(self, child):
assert self.parent, '<%s> node should never be root.' %
if isinstance(child, SplicingNode):
return True # avoid O(n^2) behavior
return self.parent._IsValidChild(child)
class IfNode(SplicingNode):
"""A node for conditional inclusion of resources.
def MandatoryAttributes(self):
return ['expr']
def _IsValidChild(self, child):
return (isinstance(child, (ThenNode, ElseNode)) or
def EndParsing(self):
children = self.children
self.if_then_else = False
if any(isinstance(node, (ThenNode, ElseNode)) for node in children):
if (len(children) != 2 or not isinstance(children[0], ThenNode) or
not isinstance(children[1], ElseNode)):
raise exception.UnexpectedChild(
'<if> element must be <if><then>...</then><else>...</else></if>')
self.if_then_else = True
def ActiveChildren(self):
cond = self.EvaluateCondition(self.attrs['expr'])
if self.if_then_else:
return self.children[0 if cond else 1].ActiveChildren()
# Equivalent to having all children inside <then> with an empty <else>
return super().ActiveChildren() if cond else []
class ThenNode(SplicingNode):
"""A <then> node. Can only appear directly inside an <if> node."""
class ElseNode(SplicingNode):
"""An <else> node. Can only appear directly inside an <if> node."""
class PartNode(SplicingNode):
"""A node for inclusion of sub-grd (*.grp) files.
def __init__(self):
self.started_inclusion = False
def MandatoryAttributes(self):
return ['file']
def _IsValidChild(self, child):
return self.started_inclusion and super()._IsValidChild(child)
class ReleaseNode(base.Node):
"""The <release> element."""
def _IsValidChild(self, child):
from grit.node import empty
return isinstance(child, (empty.IncludesNode, empty.MessagesNode,
empty.StructuresNode, empty.IdentifiersNode))
def _IsValidAttribute(self, name, value):
return (
(name == 'seq' and int(value) <= self.GetRoot().GetCurrentRelease()) or
name == 'allow_pseudo'
def MandatoryAttributes(self):
return ['seq']
def DefaultAttributes(self):
return { 'allow_pseudo' : 'true' }
class GritNode(base.Node):
"""The <grit> root element."""
def __init__(self):
self.output_language = ''
self.defines = {}
self.substituter = None
self.target_platform = sys.platform
self.allowlist_support = False
self._predetermined_ids_file = None
self._id_map = None # Dict of textual_id -> numeric_id.
# Option to skip any post-parsing validation, useful when only interested
# in getting a reference to the parse tree.
self.skip_validation_checks = False
def _IsValidChild(self, child):
from grit.node import empty
return isinstance(child, (ReleaseNode, empty.TranslationsNode,
def _IsValidAttribute(self, name, value):
if name not in ['base_dir', 'first_ids_file', 'source_lang_id',
'latest_public_release', 'current_release',
'enc_check', 'tc_project', 'grit_version',
return False
if name in ['latest_public_release', 'current_release'] and value.strip(
'0123456789') != '':
return False
return True
def MandatoryAttributes(self):
return ['latest_public_release', 'current_release']
def DefaultAttributes(self):
return {
'base_dir' : '.',
'first_ids_file': '',
'grit_version': 1,
'source_lang_id' : 'en',
'enc_check' : constants.ENCODING_CHECK,
'tc_project' : 'NEED_TO_SET_tc_project_ATTRIBUTE',
def EndParsing(self):
if (int(self.attrs['latest_public_release'])
> int(self.attrs['current_release'])):
raise exception.Parsing('latest_public_release cannot have a greater '
'value than current_release')
if self.skip_validation_checks:
# Add the encoding check if it's not present (should ensure that it's always
# present in all .grd files generated by GRIT). If it's present, assert if
# it's not correct.
if 'enc_check' not in self.attrs or self.attrs['enc_check'] == '':
self.attrs['enc_check'] = constants.ENCODING_CHECK
assert self.attrs['enc_check'] == constants.ENCODING_CHECK, (
'Are you sure your .grd file is in the correct encoding (UTF-8)?')
def ValidateUniqueIds(self):
"""Validate that 'name' attribute is unique in all nodes in this tree
except for nodes that are children of <if> nodes.
unique_names = {}
duplicate_names = []
# To avoid false positives from mutually exclusive <if> clauses, check
# against whatever the output condition happens to be right now.
# TODO(benrg): do something better.
for node in self.ActiveDescendants():
if node.attrs.get('generateid', 'true') == 'false':
continue # Duplication not relevant in that case
for node_id in node.GetTextualIds():
if util.SYSTEM_IDENTIFIERS.match(node_id):
continue # predefined IDs are sometimes used more than once
if node_id in unique_names and node_id not in duplicate_names:
unique_names[node_id] = 1
if len(duplicate_names):
raise exception.DuplicateKey(', '.join(duplicate_names))
def GetCurrentRelease(self):
"""Returns the current release number."""
return int(self.attrs['current_release'])
def GetLatestPublicRelease(self):
"""Returns the latest public release number."""
return int(self.attrs['latest_public_release'])
def GetSourceLanguage(self):
"""Returns the language code of the source language."""
return self.attrs['source_lang_id']
def GetTcProject(self):
"""Returns the name of this project in the TranslationConsole, or
'NEED_TO_SET_tc_project_ATTRIBUTE' if it is not defined."""
return self.attrs['tc_project']
def SetOwnDir(self, dir):
"""Informs the 'grit' element of the directory the file it is in resides.
This allows it to calculate relative paths from the input file, which is
what we desire (rather than from the current path).
dir: r'c:\bla'
assert dir
self.base_dir = os.path.normpath(os.path.join(dir, self.attrs['base_dir']))
def GetBaseDir(self):
"""Returns the base directory, relative to the working directory. To get
the base directory as set in the .grd file, use GetOriginalBaseDir()
if hasattr(self, 'base_dir'):
return self.base_dir
return self.GetOriginalBaseDir()
def GetOriginalBaseDir(self):
"""Returns the base directory, as set in the .grd file.
return self.attrs['base_dir']
def IsAllowlistSupportEnabled(self):
return self.allowlist_support
def SetAllowlistSupportEnabled(self, allowlist_support):
self.allowlist_support = allowlist_support
def GetInputFiles(self):
"""Returns the list of files that are read to produce the output."""
# Importing this here avoids a circular dependency in the imports.
# pylint: disable-msg=C6204
from grit.node import include
from grit.node import misc
from grit.node import structure
from grit.node import variant
# Check if the input is required for any output configuration.
input_files = set()
# Collect even inactive PartNodes since they affect ID assignments.
for node in self:
if isinstance(node, misc.PartNode):
old_output_language = self.output_language
for lang, ctx, fallback in self.GetConfigurations():
self.SetOutputLanguage(lang or self.GetSourceLanguage())
for node in self.ActiveDescendants():
if isinstance(node, (node_io.FileNode, include.IncludeNode,
structure.StructureNode, variant.SkeletonNode)):
input_path = node.GetInputPath()
if input_path is not None:
# If it's a flattened node, grab inlined resources too.
if (( == 'structure' or == 'include')
and node.attrs['flattenhtml'] == 'true'):
if == 'structure':
return sorted(input_files)
def GetFirstIdsFile(self):
"""Returns a usable path to the first_ids file, if set, otherwise
returns None.
The first_ids_file attribute is by default relative to the
base_dir of the .grd file, but may be prefixed by GRIT_DIR/,
which makes it relative to the directory of
(e.g. GRIT_DIR/../gritsettings/resource_ids).
if not self.attrs['first_ids_file']:
return None
path = self.attrs['first_ids_file']
if (path.startswith(GRIT_DIR_PREFIX)
and path[len(GRIT_DIR_PREFIX)] in ['/', '\\']):
return util.PathFromRoot(path[len(GRIT_DIR_PREFIX) + 1:])
return self.ToRealPath(path)
def GetOutputFiles(self):
"""Returns the list of <output> nodes that are descendants of this node's
<outputs> child and are not enclosed by unsatisfied <if> conditionals.
for child in self.children:
if == 'outputs':
return [node for node in child.ActiveDescendants()
if == 'output']
raise exception.MissingElement()
def GetConfigurations(self):
"""Returns the distinct (language, context, fallback_to_default_layout)
triples from the output nodes.
return {(n.GetLanguage(), n.GetContext(), n.GetFallbackToDefaultLayout())
for n in self.GetOutputFiles()}
def GetSubstitutionMessages(self):
"""Returns the list of <message sub_variable="true"> nodes."""
return [n for n in self.ActiveDescendants()
if isinstance(n, message.MessageNode)
and n.attrs['sub_variable'] == 'true']
def SetOutputLanguage(self, output_language):
"""Set the output language. Prepares substitutions.
The substitutions are reset every time the language is changed.
They include messages designated as variables, and language codes for html
and rc files.
output_language: a two-letter language code (eg: 'en', 'ar'...) or ''
if not output_language:
# We do not specify the output language for .grh files,
# so we get an empty string as the default.
# The value should match grit.clique.MessageClique.source_language.
output_language = self.GetSourceLanguage()
if output_language != self.output_language:
self.output_language = output_language
self.substituter = None # force recalculate
def SetOutputContext(self, output_context):
self.output_context = output_context
self.substituter = None # force recalculate
def SetFallbackToDefaultLayout(self, fallback_to_default_layout):
self.fallback_to_default_layout = fallback_to_default_layout
self.substituter = None # force recalculate
def SetDefines(self, defines):
self.defines = defines
self.substituter = None # force recalculate
def SetTargetPlatform(self, target_platform):
self.target_platform = target_platform
def GetSubstituter(self):
if self.substituter is None:
self.substituter = util.Substituter()
if self.output_language in _RTL_LANGS:
direction = 'dir="RTL"'
direction = 'dir="LTR"'
'GRITLANGCODE': self.output_language,
'GRITDIR': direction,
from grit.format import rc # avoid circular dep
rc.RcSubstitutions(self.substituter, self.output_language)
return self.substituter
def AssignFirstIds(self, filename_or_stream, defines):
"""Assign first ids to each grouping node based on values from the
first_ids file (if specified on the <grit> node).
assert self._id_map is None, 'AssignFirstIds() after InitializeIds()'
# If the input is a stream, then we're probably in a unit test and
# should skip this step.
if not isinstance(filename_or_stream, str):
# Nothing to do if the first_ids_filename attribute isn't set.
first_ids_filename = self.GetFirstIdsFile()
if not first_ids_filename:
src_root_dir, first_ids = _ReadFirstIdsFromFile(first_ids_filename,
from grit.node import empty
grouping_nodes = [
n for n in self.Preorder() if isinstance(n, empty.GroupingNode)
if not grouping_nodes:
grd_filename = _GetIdMapKeyFromFilename(filename_or_stream, src_root_dir)
first_ids_dict = first_ids.get(grd_filename, {})
for node in grouping_nodes:
if node.attrs['first_id'] != '':
raise Exception(
"Don't set the first_id attribute when using the first_ids_file "
"attribute on the <grit> node. Update %s instead." %
id_list = first_ids_dict[]
node.attrs['first_id'] = str(id_list.pop(0))
except (KeyError, IndexError) as e:
raise Exception('Please update %s and add a first id for %s (%s).' %
(first_ids_filename, grd_filename,
# E.g.: "META": {"sizes": {"includes": [50]}}
sizes = first_ids_dict.get('META', {}).get('sizes', {}).get(
if sizes:
node.attrs['max_ids'] = str(sizes.pop(0))
def GetIdMap(self):
'''Return a dictionary mapping textual ids to numeric ids.'''
return self._id_map
def SetPredeterminedIdsFile(self, predetermined_ids_file):
assert self._id_map is None, (
'SetPredeterminedIdsFile() after InitializeIds()')
self._predetermined_ids_file = predetermined_ids_file
def InitializeIds(self):
'''Initializes the text ID -> numeric ID mapping.'''
predetermined_id_map = {}
if self._predetermined_ids_file:
with open(self._predetermined_ids_file) as f:
for line in f:
tid, nid = line.split()
predetermined_id_map[tid] = int(nid)
self._id_map = _ComputeIds(self, predetermined_id_map)
def RunGatherers(self, debug=False):
'''Call RunPreSubstitutionGatherer() on every node of the tree, then apply
substitutions, then call RunPostSubstitutionGatherer() on every node.
The substitutions step requires that the output language has been set.
Locally, get the Substitution messages and add them to the substituter.
Also add substitutions for language codes in the Rc.
debug: will print information while running gatherers.
for node in self.ActiveDescendants():
if hasattr(node, 'RunPreSubstitutionGatherer'):
with node:
assert self.output_language
for node in self.ActiveDescendants():
if hasattr(node, 'RunPostSubstitutionGatherer'):
with node:
class IdentifierNode(base.Node):
"""A node for specifying identifiers that should appear in the resource
header file, and be unique amongst all other resource identifiers, but don't
have any other attributes or reference any resources.
def MandatoryAttributes(self):
return ['name']
def DefaultAttributes(self):
return { 'comment' : '', 'id' : '', 'systemid': 'false' }
def GetId(self):
"""Returns the id of this identifier if it has one, None otherwise
if 'id' in self.attrs:
return self.attrs['id']
return None
def EndParsing(self):
"""Handles system identifiers."""
if self.attrs['systemid'] == 'true':
def Construct(parent, name, id, comment, systemid='false'):
"""Creates a new node which is a child of 'parent', with attributes set
by parameters of the same name.
node = IdentifierNode()
node.StartParsing('identifier', parent)
node.HandleAttribute('name', name)
node.HandleAttribute('id', id)
node.HandleAttribute('comment', comment)
node.HandleAttribute('systemid', systemid)
return node