# Copyright 2016-2017 The Meson development team
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
# http://www.apache.org/licenses/LICENSE-2.0
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# This class contains the basic functionality needed to run any interpreter
# or an interpreter-based tool.
from . import mparser , mesonlib , mlog
from . import environment , dependencies
import abc
import os , copy , re
import collections . abc
from functools import wraps
import typing as T
TV_fw_var = T . Union [ str , int , float , bool , list , dict , ' InterpreterObject ' , ' ObjectHolder ' ]
TV_fw_args = T . List [ T . Union [ mparser . BaseNode , TV_fw_var ] ]
TV_fw_kwargs = T . Dict [ str , T . Union [ mparser . BaseNode , TV_fw_var ] ]
TV_func = T . TypeVar ( ' TV_func ' , bound = T . Callable [ . . . , T . Any ] )
TYPE_elementary = T . Union [ str , int , float , bool ]
TYPE_var = T . Union [ TYPE_elementary , T . List [ T . Any ] , T . Dict [ str , T . Any ] , ' InterpreterObject ' , ' ObjectHolder ' ]
TYPE_nvar = T . Union [ TYPE_var , mparser . BaseNode ]
TYPE_nkwargs = T . Dict [ str , TYPE_nvar ]
TYPE_key_resolver = T . Callable [ [ mparser . BaseNode ] , str ]
class InterpreterObject :
def __init__ ( self ) - > None :
self . methods = { } # type: T.Dict[str, T.Callable[[T.List[TYPE_nvar], TYPE_nkwargs], TYPE_var]]
# Current node set during a method call. This can be used as location
# when printing a warning message during a method call.
self . current_node = None # type: mparser.BaseNode
def method_call (
self ,
method_name : str ,
args : TV_fw_args ,
kwargs : TV_fw_kwargs
) - > TYPE_var :
if method_name in self . methods :
method = self . methods [ method_name ]
if not getattr ( method , ' no-args-flattening ' , False ) :
args = flatten ( args )
return method ( args , kwargs )
raise InvalidCode ( ' Unknown method " %s " in object. ' % method_name )
TV_InterpreterObject = T . TypeVar ( ' TV_InterpreterObject ' )
class ObjectHolder ( T . Generic [ TV_InterpreterObject ] ) :
def __init__ ( self , obj : InterpreterObject , subproject : T . Optional [ str ] = None ) - > None :
self . held_object = obj # type: InterpreterObject
self . subproject = subproject # type: str
def __repr__ ( self ) - > str :
return ' <Holder: {!r} > ' . format ( self . held_object )
class MesonVersionString ( str ) :
pass
# Decorators for method calls.
def check_stringlist ( a : T . Any , msg : str = ' Arguments must be strings. ' ) - > None :
if not isinstance ( a , list ) :
mlog . debug ( ' Not a list: ' , str ( a ) )
raise InvalidArguments ( ' Argument not a list. ' )
if not all ( isinstance ( s , str ) for s in a ) :
mlog . debug ( ' Element not a string: ' , str ( a ) )
raise InvalidArguments ( msg )
def _get_callee_args ( wrapped_args : T . Sequence [ T . Any ] , want_subproject : bool = False ) - > T . Tuple [ T . Any , mparser . BaseNode , TV_fw_args , TV_fw_kwargs , T . Optional [ str ] ] :
s = wrapped_args [ 0 ]
n = len ( wrapped_args )
# Raise an error if the codepaths are not there
subproject = None # type: T.Optional[str]
if want_subproject and n == 2 :
if hasattr ( s , ' subproject ' ) :
# Interpreter base types have 2 args: self, node
node = wrapped_args [ 1 ]
# args and kwargs are inside the node
args = None
kwargs = None
subproject = s . subproject
elif hasattr ( wrapped_args [ 1 ] , ' subproject ' ) :
# Module objects have 2 args: self, interpreter
node = wrapped_args [ 1 ] . current_node
# args and kwargs are inside the node
args = None
kwargs = None
subproject = wrapped_args [ 1 ] . subproject
else :
raise AssertionError ( ' Unknown args: {!r} ' . format ( wrapped_args ) )
elif n == 3 :
# Methods on objects (*Holder, MesonMain, etc) have 3 args: self, args, kwargs
node = s . current_node
args = wrapped_args [ 1 ]
kwargs = wrapped_args [ 2 ]
if want_subproject :
if hasattr ( s , ' subproject ' ) :
subproject = s . subproject
elif hasattr ( s , ' interpreter ' ) :
subproject = s . interpreter . subproject
elif n == 4 :
# Meson functions have 4 args: self, node, args, kwargs
# Module functions have 4 args: self, state, args, kwargs
if isinstance ( s , InterpreterBase ) :
node = wrapped_args [ 1 ]
else :
node = wrapped_args [ 1 ] . current_node
args = wrapped_args [ 2 ]
kwargs = wrapped_args [ 3 ]
if want_subproject :
if isinstance ( s , InterpreterBase ) :
subproject = s . subproject
else :
subproject = wrapped_args [ 1 ] . subproject
elif n == 5 :
# Module snippets have 5 args: self, interpreter, state, args, kwargs
node = wrapped_args [ 2 ] . current_node
args = wrapped_args [ 3 ]
kwargs = wrapped_args [ 4 ]
if want_subproject :
subproject = wrapped_args [ 2 ] . subproject
else :
raise AssertionError ( ' Unknown args: {!r} ' . format ( wrapped_args ) )
# Sometimes interpreter methods are called internally with None instead of
# empty list/dict
args = args if args is not None else [ ]
kwargs = kwargs if kwargs is not None else { }
return s , node , args , kwargs , subproject
def flatten ( args : T . Union [ TYPE_nvar , T . List [ TYPE_nvar ] ] ) - > T . List [ TYPE_nvar ] :
if isinstance ( args , mparser . StringNode ) :
assert isinstance ( args . value , str )
return [ args . value ]
if not isinstance ( args , collections . abc . Sequence ) :
return [ args ]
result = [ ] # type: T.List[TYPE_nvar]
for a in args :
if isinstance ( a , list ) :
rest = flatten ( a )
result = result + rest
elif isinstance ( a , mparser . StringNode ) :
result . append ( a . value )
else :
result . append ( a )
return result
def noPosargs ( f : TV_func ) - > TV_func :
@wraps ( f )
def wrapped ( * wrapped_args : T . Any , * * wrapped_kwargs : T . Any ) - > T . Any :
args = _get_callee_args ( wrapped_args ) [ 2 ]
Don't use len() to test emptiness vs not emptiness
Meson has a common pattern of using 'if len(foo) == 0:' or
'if len(foo) != 0:', however, this is a common anti-pattern in python.
Instead tests for emptiness/non-emptiness should be done with a simple
'if foo:' or 'if not foo:'
Consider the following:
>>> import timeit
>>> timeit.timeit('if len([]) == 0: pass')
0.10730923599840025
>>> timeit.timeit('if not []: pass')
0.030033907998586074
>>> timeit.timeit('if len(['a', 'b', 'c', 'd']) == 0: pass')
0.1154778649979562
>>> timeit.timeit("if not ['a', 'b', 'c', 'd']: pass")
0.08259823200205574
>>> timeit.timeit('if len("") == 0: pass')
0.089759664999292
>>> timeit.timeit('if not "": pass')
0.02340641999762738
>>> timeit.timeit('if len("foo") == 0: pass')
0.08848102600313723
>>> timeit.timeit('if not "foo": pass')
0.04032287199879647
And for the one additional case of 'if len(foo.strip()) == 0', which can
be replaced with 'if not foo.isspace()'
>>> timeit.timeit('if len(" ".strip()) == 0: pass')
0.15294511600222904
>>> timeit.timeit('if " ".isspace(): pass')
0.09413968399894657
>>> timeit.timeit('if len(" abc".strip()) == 0: pass')
0.2023209120015963
>>> timeit.timeit('if " abc".isspace(): pass')
0.09571301700270851
In other words, it's always a win to not use len(), when you don't
actually want to check the length.
8 years ago
if args :
raise InvalidArguments ( ' Function does not take positional arguments. ' )
return f ( * wrapped_args , * * wrapped_kwargs )
return T . cast ( TV_func , wrapped )
def builtinMethodNoKwargs ( f : TV_func ) - > TV_func :
@wraps ( f )
def wrapped ( * wrapped_args : T . Any , * * wrapped_kwargs : T . Any ) - > T . Any :
node = wrapped_args [ 0 ] . current_node
method_name = wrapped_args [ 2 ]
kwargs = wrapped_args [ 4 ]
if kwargs :
mlog . warning ( ' Method {!r} does not take keyword arguments. ' . format ( method_name ) ,
' This will become a hard error in the future ' ,
location = node )
return f ( * wrapped_args , * * wrapped_kwargs )
return T . cast ( TV_func , wrapped )
def noKwargs ( f : TV_func ) - > TV_func :
@wraps ( f )
def wrapped ( * wrapped_args : T . Any , * * wrapped_kwargs : T . Any ) - > T . Any :
kwargs = _get_callee_args ( wrapped_args ) [ 3 ]
Don't use len() to test emptiness vs not emptiness
Meson has a common pattern of using 'if len(foo) == 0:' or
'if len(foo) != 0:', however, this is a common anti-pattern in python.
Instead tests for emptiness/non-emptiness should be done with a simple
'if foo:' or 'if not foo:'
Consider the following:
>>> import timeit
>>> timeit.timeit('if len([]) == 0: pass')
0.10730923599840025
>>> timeit.timeit('if not []: pass')
0.030033907998586074
>>> timeit.timeit('if len(['a', 'b', 'c', 'd']) == 0: pass')
0.1154778649979562
>>> timeit.timeit("if not ['a', 'b', 'c', 'd']: pass")
0.08259823200205574
>>> timeit.timeit('if len("") == 0: pass')
0.089759664999292
>>> timeit.timeit('if not "": pass')
0.02340641999762738
>>> timeit.timeit('if len("foo") == 0: pass')
0.08848102600313723
>>> timeit.timeit('if not "foo": pass')
0.04032287199879647
And for the one additional case of 'if len(foo.strip()) == 0', which can
be replaced with 'if not foo.isspace()'
>>> timeit.timeit('if len(" ".strip()) == 0: pass')
0.15294511600222904
>>> timeit.timeit('if " ".isspace(): pass')
0.09413968399894657
>>> timeit.timeit('if len(" abc".strip()) == 0: pass')
0.2023209120015963
>>> timeit.timeit('if " abc".isspace(): pass')
0.09571301700270851
In other words, it's always a win to not use len(), when you don't
actually want to check the length.
8 years ago
if kwargs :
raise InvalidArguments ( ' Function does not take keyword arguments. ' )
return f ( * wrapped_args , * * wrapped_kwargs )
return T . cast ( TV_func , wrapped )
def stringArgs ( f : TV_func ) - > TV_func :
@wraps ( f )
def wrapped ( * wrapped_args : T . Any , * * wrapped_kwargs : T . Any ) - > T . Any :
args = _get_callee_args ( wrapped_args ) [ 2 ]
assert ( isinstance ( args , list ) )
check_stringlist ( args )
return f ( * wrapped_args , * * wrapped_kwargs )
return T . cast ( TV_func , wrapped )
def noArgsFlattening ( f : TV_func ) - > TV_func :
setattr ( f , ' no-args-flattening ' , True ) # noqa: B010
return f
def disablerIfNotFound ( f : TV_func ) - > TV_func :
@wraps ( f )
def wrapped ( * wrapped_args : T . Any , * * wrapped_kwargs : T . Any ) - > T . Any :
kwargs = _get_callee_args ( wrapped_args ) [ 3 ]
disabler = kwargs . pop ( ' disabler ' , False )
ret = f ( * wrapped_args , * * wrapped_kwargs )
if disabler and not ret . held_object . found ( ) :
return Disabler ( )
return ret
return T . cast ( TV_func , wrapped )
class permittedKwargs :
def __init__ ( self , permitted : T . Set [ str ] ) :
self . permitted = permitted # type: T.Set[str]
def __call__ ( self , f : TV_func ) - > TV_func :
@wraps ( f )
def wrapped ( * wrapped_args : T . Any , * * wrapped_kwargs : T . Any ) - > T . Any :
s , node , args , kwargs , _ = _get_callee_args ( wrapped_args )
for k in kwargs :
if k not in self . permitted :
mlog . warning ( ''' Passed invalid keyword argument " {} " . ''' . format ( k ) , location = node )
mlog . warning ( ' This will become a hard error in the future. ' )
return f ( * wrapped_args , * * wrapped_kwargs )
return T . cast ( TV_func , wrapped )
class FeatureCheckBase ( metaclass = abc . ABCMeta ) :
" Base class for feature version checks "
# In python 3.6 we can just forward declare this, but in 3.5 we can't
# This will be overwritten by the subclasses by necessity
feature_registry = { } # type: T.ClassVar[T.Dict[str, T.Dict[str, T.Set[str]]]]
def __init__ ( self , feature_name : str , version : str , extra_message : T . Optional [ str ] = None ) :
self . feature_name = feature_name # type: str
self . feature_version = version # type: str
self . extra_message = extra_message or ' ' # type: str
@staticmethod
def get_target_version ( subproject : str ) - > str :
# Don't do any checks if project() has not been parsed yet
if subproject not in mesonlib . project_meson_versions :
return ' '
return mesonlib . project_meson_versions [ subproject ]
@staticmethod
@abc . abstractmethod
def check_version ( target_version : str , feature_Version : str ) - > bool :
pass
def use ( self , subproject : str ) - > None :
tv = self . get_target_version ( subproject )
# No target version
if tv == ' ' :
return
# Target version is new enough
if self . check_version ( tv , self . feature_version ) :
return
# Feature is too new for target version, register it
if subproject not in self . feature_registry :
self . feature_registry [ subproject ] = { self . feature_version : set ( ) }
register = self . feature_registry [ subproject ]
if self . feature_version not in register :
register [ self . feature_version ] = set ( )
if self . feature_name in register [ self . feature_version ] :
# Don't warn about the same feature multiple times
# FIXME: This is needed to prevent duplicate warnings, but also
# means we won't warn about a feature used in multiple places.
return
register [ self . feature_version ] . add ( self . feature_name )
self . log_usage_warning ( tv )
@classmethod
def report ( cls , subproject : str ) - > None :
if subproject not in cls . feature_registry :
return
warning_str = cls . get_warning_str_prefix ( cls . get_target_version ( subproject ) )
fv = cls . feature_registry [ subproject ]
for version in sorted ( fv . keys ( ) ) :
warning_str + = ' \n * {} : {} ' . format ( version , fv [ version ] )
mlog . warning ( warning_str )
def log_usage_warning ( self , tv : str ) - > None :
raise InterpreterException ( ' log_usage_warning not implemented ' )
@staticmethod
def get_warning_str_prefix ( tv : str ) - > str :
raise InterpreterException ( ' get_warning_str_prefix not implemented ' )
def __call__ ( self , f : TV_func ) - > TV_func :
@wraps ( f )
def wrapped ( * wrapped_args : T . Any , * * wrapped_kwargs : T . Any ) - > T . Any :
subproject = _get_callee_args ( wrapped_args , want_subproject = True ) [ 4 ]
if subproject is None :
raise AssertionError ( ' {!r} ' . format ( wrapped_args ) )
self . use ( subproject )
return f ( * wrapped_args , * * wrapped_kwargs )
return T . cast ( TV_func , wrapped )
@classmethod
def single_use ( cls , feature_name : str , version : str , subproject : str ,
extra_message : T . Optional [ str ] = None ) - > None :
""" Oneline version that instantiates and calls use(). """
cls ( feature_name , version , extra_message ) . use ( subproject )
class FeatureNew ( FeatureCheckBase ) :
""" Checks for new features """
# Class variable, shared across all instances
#
# Format: {subproject: {feature_version: set(feature_names)}}
feature_registry = { } # type: T.ClassVar[T.Dict[str, T.Dict[str, T.Set[str]]]]
@staticmethod
def check_version ( target_version : str , feature_version : str ) - > bool :
return mesonlib . version_compare_condition_with_min ( target_version , feature_version )
@staticmethod
def get_warning_str_prefix ( tv : str ) - > str :
return ' Project specifies a minimum meson_version \' {} \' but uses features which were added in newer versions: ' . format ( tv )
def log_usage_warning ( self , tv : str ) - > None :
args = [
' Project targeting ' , " ' {} ' " . format ( tv ) ,
' but tried to use feature introduced in ' ,
" ' {} ' : " . format ( self . feature_version ) ,
' {} . ' . format ( self . feature_name ) ,
]
if self . extra_message :
args . append ( self . extra_message )
mlog . warning ( * args )
class FeatureDeprecated ( FeatureCheckBase ) :
""" Checks for deprecated features """
# Class variable, shared across all instances
#
# Format: {subproject: {feature_version: set(feature_names)}}
feature_registry = { } # type: T.ClassVar[T.Dict[str, T.Dict[str, T.Set[str]]]]
@staticmethod
def check_version ( target_version : str , feature_version : str ) - > bool :
# For deprecation checks we need to return the inverse of FeatureNew checks
return not mesonlib . version_compare_condition_with_min ( target_version , feature_version )
@staticmethod
def get_warning_str_prefix ( tv : str ) - > str :
return ' Deprecated features used: '
def log_usage_warning ( self , tv : str ) - > None :
args = [
' Project targeting ' , " ' {} ' " . format ( tv ) ,
' but tried to use feature deprecated since ' ,
" ' {} ' : " . format ( self . feature_version ) ,
' {} . ' . format ( self . feature_name ) ,
]
if self . extra_message :
args . append ( self . extra_message )
mlog . warning ( * args )
class FeatureCheckKwargsBase ( metaclass = abc . ABCMeta ) :
@property
@abc . abstractmethod
def feature_check_class ( self ) - > T . Type [ FeatureCheckBase ] :
pass
def __init__ ( self , feature_name : str , feature_version : str ,
kwargs : T . List [ str ] , extra_message : T . Optional [ str ] = None ) :
self . feature_name = feature_name
self . feature_version = feature_version
self . kwargs = kwargs
self . extra_message = extra_message
def __call__ ( self , f : TV_func ) - > TV_func :
@wraps ( f )
def wrapped ( * wrapped_args : T . Any , * * wrapped_kwargs : T . Any ) - > T . Any :
kwargs , subproject = _get_callee_args ( wrapped_args , want_subproject = True ) [ 3 : 5 ]
if subproject is None :
raise AssertionError ( ' {!r} ' . format ( wrapped_args ) )
for arg in self . kwargs :
if arg not in kwargs :
continue
name = arg + ' arg in ' + self . feature_name
self . feature_check_class . single_use (
name , self . feature_version , subproject , self . extra_message )
return f ( * wrapped_args , * * wrapped_kwargs )
return T . cast ( TV_func , wrapped )
class FeatureNewKwargs ( FeatureCheckKwargsBase ) :
feature_check_class = FeatureNew
class FeatureDeprecatedKwargs ( FeatureCheckKwargsBase ) :
feature_check_class = FeatureDeprecated
class InterpreterException ( mesonlib . MesonException ) :
pass
class InvalidCode ( InterpreterException ) :
pass
class InvalidArguments ( InterpreterException ) :
pass
class SubdirDoneRequest ( BaseException ) :
pass
class ContinueRequest ( BaseException ) :
pass
class BreakRequest ( BaseException ) :
pass
class MutableInterpreterObject ( InterpreterObject ) :
def __init__ ( self ) - > None :
super ( ) . __init__ ( )
class Disabler ( InterpreterObject ) :
def __init__ ( self ) - > None :
super ( ) . __init__ ( )
self . methods . update ( { ' found ' : self . found_method } )
def found_method ( self , args : T . Sequence [ T . Any ] , kwargs : T . Dict [ str , T . Any ] ) - > bool :
return False
def is_disabler ( i : T . Any ) - > bool :
return isinstance ( i , Disabler )
def is_arg_disabled ( arg : T . Any ) - > bool :
if is_disabler ( arg ) :
return True
if isinstance ( arg , list ) :
for i in arg :
if is_arg_disabled ( i ) :
return True
return False
def is_disabled ( args : T . Sequence [ T . Any ] , kwargs : T . Dict [ str , T . Any ] ) - > bool :
for i in args :
if is_arg_disabled ( i ) :
return True
for i in kwargs . values ( ) :
if is_arg_disabled ( i ) :
return True
return False
def default_resolve_key ( key : mparser . BaseNode ) - > str :
if not isinstance ( key , mparser . IdNode ) :
raise InterpreterException ( ' Invalid kwargs format. ' )
return key . value
class InterpreterBase :
elementary_types = ( int , float , str , bool , list )
def __init__ ( self , source_root : str , subdir : str , subproject : str ) :
self . source_root = source_root
self . funcs = { } # type: T.Dict[str, T.Callable[[mparser.BaseNode, T.List[TYPE_nvar], T.Dict[str, TYPE_nvar]], TYPE_var]]
self . builtin = { } # type: T.Dict[str, InterpreterObject]
self . subdir = subdir
self . root_subdir = subdir
self . subproject = subproject
self . variables = { } # type: T.Dict[str, TYPE_var]
self . argument_depth = 0
self . current_lineno = - 1
# Current node set during a function call. This can be used as location
# when printing a warning message during a method call.
self . current_node = None # type: mparser.BaseNode
# This is set to `version_string` when this statement is evaluated:
# meson.version().compare_version(version_string)
# If it was part of a if-clause, it is used to temporally override the
# current meson version target within that if-block.
self . tmp_meson_version = None # type: T.Optional[str]
def load_root_meson_file ( self ) - > None :
mesonfile = os . path . join ( self . source_root , self . subdir , environment . build_filename )
if not os . path . isfile ( mesonfile ) :
raise InvalidArguments ( ' Missing Meson file in %s ' % mesonfile )
with open ( mesonfile , encoding = ' utf8 ' ) as mf :
code = mf . read ( )
Don't use len() to test emptiness vs not emptiness
Meson has a common pattern of using 'if len(foo) == 0:' or
'if len(foo) != 0:', however, this is a common anti-pattern in python.
Instead tests for emptiness/non-emptiness should be done with a simple
'if foo:' or 'if not foo:'
Consider the following:
>>> import timeit
>>> timeit.timeit('if len([]) == 0: pass')
0.10730923599840025
>>> timeit.timeit('if not []: pass')
0.030033907998586074
>>> timeit.timeit('if len(['a', 'b', 'c', 'd']) == 0: pass')
0.1154778649979562
>>> timeit.timeit("if not ['a', 'b', 'c', 'd']: pass")
0.08259823200205574
>>> timeit.timeit('if len("") == 0: pass')
0.089759664999292
>>> timeit.timeit('if not "": pass')
0.02340641999762738
>>> timeit.timeit('if len("foo") == 0: pass')
0.08848102600313723
>>> timeit.timeit('if not "foo": pass')
0.04032287199879647
And for the one additional case of 'if len(foo.strip()) == 0', which can
be replaced with 'if not foo.isspace()'
>>> timeit.timeit('if len(" ".strip()) == 0: pass')
0.15294511600222904
>>> timeit.timeit('if " ".isspace(): pass')
0.09413968399894657
>>> timeit.timeit('if len(" abc".strip()) == 0: pass')
0.2023209120015963
>>> timeit.timeit('if " abc".isspace(): pass')
0.09571301700270851
In other words, it's always a win to not use len(), when you don't
actually want to check the length.
8 years ago
if code . isspace ( ) :
raise InvalidCode ( ' Builder file is empty. ' )
assert ( isinstance ( code , str ) )
try :
self . ast = mparser . Parser ( code , mesonfile ) . parse ( )
except mesonlib . MesonException as me :
me . file = mesonfile
raise me
def join_path_strings ( self , args : T . Sequence [ str ] ) - > str :
return os . path . join ( * args ) . replace ( ' \\ ' , ' / ' )
def parse_project ( self ) - > None :
"""
Parses project ( ) and initializes languages , compilers etc . Do this
early because we need this before we parse the rest of the AST .
"""
self . evaluate_codeblock ( self . ast , end = 1 )
def sanity_check_ast ( self ) - > None :
if not isinstance ( self . ast , mparser . CodeBlockNode ) :
raise InvalidCode ( ' AST is of invalid type. Possibly a bug in the parser. ' )
Don't use len() to test emptiness vs not emptiness
Meson has a common pattern of using 'if len(foo) == 0:' or
'if len(foo) != 0:', however, this is a common anti-pattern in python.
Instead tests for emptiness/non-emptiness should be done with a simple
'if foo:' or 'if not foo:'
Consider the following:
>>> import timeit
>>> timeit.timeit('if len([]) == 0: pass')
0.10730923599840025
>>> timeit.timeit('if not []: pass')
0.030033907998586074
>>> timeit.timeit('if len(['a', 'b', 'c', 'd']) == 0: pass')
0.1154778649979562
>>> timeit.timeit("if not ['a', 'b', 'c', 'd']: pass")
0.08259823200205574
>>> timeit.timeit('if len("") == 0: pass')
0.089759664999292
>>> timeit.timeit('if not "": pass')
0.02340641999762738
>>> timeit.timeit('if len("foo") == 0: pass')
0.08848102600313723
>>> timeit.timeit('if not "foo": pass')
0.04032287199879647
And for the one additional case of 'if len(foo.strip()) == 0', which can
be replaced with 'if not foo.isspace()'
>>> timeit.timeit('if len(" ".strip()) == 0: pass')
0.15294511600222904
>>> timeit.timeit('if " ".isspace(): pass')
0.09413968399894657
>>> timeit.timeit('if len(" abc".strip()) == 0: pass')
0.2023209120015963
>>> timeit.timeit('if " abc".isspace(): pass')
0.09571301700270851
In other words, it's always a win to not use len(), when you don't
actually want to check the length.
8 years ago
if not self . ast . lines :
raise InvalidCode ( ' No statements in code. ' )
first = self . ast . lines [ 0 ]
if not isinstance ( first , mparser . FunctionNode ) or first . func_name != ' project ' :
raise InvalidCode ( ' First statement must be a call to project ' )
def run ( self ) - > None :
# Evaluate everything after the first line, which is project() because
# we already parsed that in self.parse_project()
try :
self . evaluate_codeblock ( self . ast , start = 1 )
except SubdirDoneRequest :
pass
def evaluate_codeblock ( self , node : mparser . CodeBlockNode , start : int = 0 , end : T . Optional [ int ] = None ) - > None :
if node is None :
return
if not isinstance ( node , mparser . CodeBlockNode ) :
e = InvalidCode ( ' Tried to execute a non-codeblock. Possibly a bug in the parser. ' )
e . lineno = node . lineno
e . colno = node . colno
raise e
statements = node . lines [ start : end ]
i = 0
while i < len ( statements ) :
cur = statements [ i ]
try :
self . current_lineno = cur . lineno
self . evaluate_statement ( cur )
except Exception as e :
if getattr ( e , ' lineno ' , None ) is None :
# We are doing the equivalent to setattr here and mypy does not like it
e . lineno = cur . lineno # type: ignore
e . colno = cur . colno # type: ignore
e . file = os . path . join ( self . source_root , self . subdir , environment . build_filename ) # type: ignore
raise e
i + = 1 # In THE FUTURE jump over blocks and stuff.
def evaluate_statement ( self , cur : mparser . BaseNode ) - > T . Optional [ TYPE_var ] :
self . current_node = cur
if isinstance ( cur , mparser . FunctionNode ) :
return self . function_call ( cur )
elif isinstance ( cur , mparser . AssignmentNode ) :
self . assignment ( cur )
elif isinstance ( cur , mparser . MethodNode ) :
return self . method_call ( cur )
elif isinstance ( cur , mparser . StringNode ) :
return cur . value
elif isinstance ( cur , mparser . BooleanNode ) :
return cur . value
elif isinstance ( cur , mparser . IfClauseNode ) :
return self . evaluate_if ( cur )
elif isinstance ( cur , mparser . IdNode ) :
return self . get_variable ( cur . value )
elif isinstance ( cur , mparser . ComparisonNode ) :
return self . evaluate_comparison ( cur )
elif isinstance ( cur , mparser . ArrayNode ) :
return self . evaluate_arraystatement ( cur )
elif isinstance ( cur , mparser . DictNode ) :
return self . evaluate_dictstatement ( cur )
elif isinstance ( cur , mparser . NumberNode ) :
return cur . value
elif isinstance ( cur , mparser . AndNode ) :
return self . evaluate_andstatement ( cur )
elif isinstance ( cur , mparser . OrNode ) :
return self . evaluate_orstatement ( cur )
elif isinstance ( cur , mparser . NotNode ) :
return self . evaluate_notstatement ( cur )
elif isinstance ( cur , mparser . UMinusNode ) :
return self . evaluate_uminusstatement ( cur )
elif isinstance ( cur , mparser . ArithmeticNode ) :
return self . evaluate_arithmeticstatement ( cur )
elif isinstance ( cur , mparser . ForeachClauseNode ) :
self . evaluate_foreach ( cur )
elif isinstance ( cur , mparser . PlusAssignmentNode ) :
self . evaluate_plusassign ( cur )
elif isinstance ( cur , mparser . IndexNode ) :
return self . evaluate_indexing ( cur )
elif isinstance ( cur , mparser . TernaryNode ) :
return self . evaluate_ternary ( cur )
elif isinstance ( cur , mparser . ContinueNode ) :
raise ContinueRequest ( )
elif isinstance ( cur , mparser . BreakNode ) :
raise BreakRequest ( )
elif isinstance ( cur , self . elementary_types ) :
return cur
else :
raise InvalidCode ( " Unknown statement. " )
return None
def evaluate_arraystatement ( self , cur : mparser . ArrayNode ) - > list :
( arguments , kwargs ) = self . reduce_arguments ( cur . args )
if len ( kwargs ) > 0 :
raise InvalidCode ( ' Keyword arguments are invalid in array construction. ' )
return arguments
@FeatureNew ( ' dict ' , ' 0.47.0 ' )
def evaluate_dictstatement ( self , cur : mparser . DictNode ) - > TYPE_nkwargs :
def resolve_key ( key : mparser . BaseNode ) - > str :
if not isinstance ( key , mparser . StringNode ) :
FeatureNew . single_use ( ' Dictionary entry using non literal key ' , ' 0.53.0 ' , self . subproject )
str_key = self . evaluate_statement ( key )
if not isinstance ( str_key , str ) :
raise InvalidArguments ( ' Key must be a string ' )
return str_key
arguments , kwargs = self . reduce_arguments ( cur . args , key_resolver = resolve_key , duplicate_key_error = ' Duplicate dictionary key: {} ' )
assert not arguments
return kwargs
def evaluate_notstatement ( self , cur : mparser . NotNode ) - > T . Union [ bool , Disabler ] :
v = self . evaluate_statement ( cur . value )
if isinstance ( v , Disabler ) :
return v
if not isinstance ( v , bool ) :
raise InterpreterException ( ' Argument to " not " is not a boolean. ' )
return not v
def evaluate_if ( self , node : mparser . IfClauseNode ) - > T . Optional [ Disabler ] :
assert ( isinstance ( node , mparser . IfClauseNode ) )
for i in node . ifs :
# Reset self.tmp_meson_version to know if it gets set during this
# statement evaluation.
self . tmp_meson_version = None
result = self . evaluate_statement ( i . condition )
if isinstance ( result , Disabler ) :
return result
if not ( isinstance ( result , bool ) ) :
raise InvalidCode ( ' If clause {!r} does not evaluate to true or false. ' . format ( result ) )
if result :
prev_meson_version = mesonlib . project_meson_versions [ self . subproject ]
if self . tmp_meson_version :
mesonlib . project_meson_versions [ self . subproject ] = self . tmp_meson_version
try :
self . evaluate_codeblock ( i . block )
finally :
mesonlib . project_meson_versions [ self . subproject ] = prev_meson_version
return None
if not isinstance ( node . elseblock , mparser . EmptyNode ) :
self . evaluate_codeblock ( node . elseblock )
return None
def validate_comparison_types ( self , val1 : T . Any , val2 : T . Any ) - > bool :
if type ( val1 ) != type ( val2 ) :
return False
return True
def evaluate_in ( self , val1 : T . Any , val2 : T . Any ) - > bool :
if not isinstance ( val1 , ( str , int , float , ObjectHolder ) ) :
raise InvalidArguments ( ' lvalue of " in " operator must be a string, integer, float, or object ' )
if not isinstance ( val2 , ( list , dict ) ) :
raise InvalidArguments ( ' rvalue of " in " operator must be an array or a dict ' )
return val1 in val2
def evaluate_comparison ( self , node : mparser . ComparisonNode ) - > T . Union [ bool , Disabler ] :
val1 = self . evaluate_statement ( node . left )
if isinstance ( val1 , Disabler ) :
return val1
val2 = self . evaluate_statement ( node . right )
if isinstance ( val2 , Disabler ) :
return val2
if node . ctype == ' in ' :
return self . evaluate_in ( val1 , val2 )
elif node . ctype == ' notin ' :
return not self . evaluate_in ( val1 , val2 )
valid = self . validate_comparison_types ( val1 , val2 )
# Ordering comparisons of different types isn't allowed since PR #1810
# (0.41.0). Since PR #2884 we also warn about equality comparisons of
# different types, which will one day become an error.
if not valid and ( node . ctype == ' == ' or node . ctype == ' != ' ) :
mlog . warning ( ''' Trying to compare values of different types ( {} , {} ) using {} .
The result of this is undefined and will become a hard error in a future Meson release . '''
. format ( type ( val1 ) . __name__ , type ( val2 ) . __name__ , node . ctype ) , location = node )
if node . ctype == ' == ' :
return val1 == val2
elif node . ctype == ' != ' :
return val1 != val2
elif not valid :
raise InterpreterException (
' Values of different types ( {} , {} ) cannot be compared using {} . ' . format ( type ( val1 ) . __name__ ,
type ( val2 ) . __name__ ,
node . ctype ) )
elif not isinstance ( val1 , self . elementary_types ) :
raise InterpreterException ( ' {} can only be compared for equality. ' . format ( getattr ( node . left , ' value ' , ' <ERROR> ' ) ) )
elif not isinstance ( val2 , self . elementary_types ) :
raise InterpreterException ( ' {} can only be compared for equality. ' . format ( getattr ( node . right , ' value ' , ' <ERROR> ' ) ) )
# Use type: ignore because mypy will complain that we are comparing two Unions,
# but we actually guarantee earlier that both types are the same
elif node . ctype == ' < ' :
return val1 < val2 # type: ignore
elif node . ctype == ' <= ' :
return val1 < = val2 # type: ignore
elif node . ctype == ' > ' :
return val1 > val2 # type: ignore
elif node . ctype == ' >= ' :
return val1 > = val2 # type: ignore
else :
raise InvalidCode ( ' You broke my compare eval. ' )
def evaluate_andstatement ( self , cur : mparser . AndNode ) - > T . Union [ bool , Disabler ] :
l = self . evaluate_statement ( cur . left )
if isinstance ( l , Disabler ) :
return l
if not isinstance ( l , bool ) :
raise InterpreterException ( ' First argument to " and " is not a boolean. ' )
if not l :
return False
r = self . evaluate_statement ( cur . right )
if isinstance ( r , Disabler ) :
return r
if not isinstance ( r , bool ) :
raise InterpreterException ( ' Second argument to " and " is not a boolean. ' )
return r
def evaluate_orstatement ( self , cur : mparser . OrNode ) - > T . Union [ bool , Disabler ] :
l = self . evaluate_statement ( cur . left )
if isinstance ( l , Disabler ) :
return l
if not isinstance ( l , bool ) :
raise InterpreterException ( ' First argument to " or " is not a boolean. ' )
if l :
return True
r = self . evaluate_statement ( cur . right )
if isinstance ( r , Disabler ) :
return r
if not isinstance ( r , bool ) :
raise InterpreterException ( ' Second argument to " or " is not a boolean. ' )
return r
def evaluate_uminusstatement ( self , cur : mparser . UMinusNode ) - > T . Union [ int , Disabler ] :
v = self . evaluate_statement ( cur . value )
if isinstance ( v , Disabler ) :
return v
if not isinstance ( v , int ) :
raise InterpreterException ( ' Argument to negation is not an integer. ' )
return - v
@FeatureNew ( ' / with string arguments ' , ' 0.49.0 ' )
def evaluate_path_join ( self , l : str , r : str ) - > str :
if not isinstance ( l , str ) :
raise InvalidCode ( ' The division operator can only append to a string. ' )
if not isinstance ( r , str ) :
raise InvalidCode ( ' The division operator can only append a string. ' )
return self . join_path_strings ( ( l , r ) )
def evaluate_division ( self , l : T . Any , r : T . Any ) - > T . Union [ int , str ] :
if isinstance ( l , str ) or isinstance ( r , str ) :
return self . evaluate_path_join ( l , r )
if isinstance ( l , int ) and isinstance ( r , int ) :
if r == 0 :
raise InvalidCode ( ' Division by zero. ' )
return l / / r
raise InvalidCode ( ' Division works only with strings or integers. ' )
def evaluate_arithmeticstatement ( self , cur : mparser . ArithmeticNode ) - > T . Union [ int , str , dict , list , Disabler ] :
l = self . evaluate_statement ( cur . left )
if isinstance ( l , Disabler ) :
return l
r = self . evaluate_statement ( cur . right )
if isinstance ( r , Disabler ) :
return r
if cur . operation == ' add ' :
if isinstance ( l , dict ) and isinstance ( r , dict ) :
return { * * l , * * r }
try :
# MyPy error due to handling two Unions (we are catching all exceptions anyway)
return l + r # type: ignore
except Exception as e :
raise InvalidCode ( ' Invalid use of addition: ' + str ( e ) )
elif cur . operation == ' sub ' :
if not isinstance ( l , int ) or not isinstance ( r , int ) :
raise InvalidCode ( ' Subtraction works only with integers. ' )
return l - r
elif cur . operation == ' mul ' :
if not isinstance ( l , int ) or not isinstance ( r , int ) :
raise InvalidCode ( ' Multiplication works only with integers. ' )
return l * r
elif cur . operation == ' div ' :
return self . evaluate_division ( l , r )
elif cur . operation == ' mod ' :
if not isinstance ( l , int ) or not isinstance ( r , int ) :
raise InvalidCode ( ' Modulo works only with integers. ' )
return l % r
else :
raise InvalidCode ( ' You broke me. ' )
def evaluate_ternary ( self , node : mparser . TernaryNode ) - > TYPE_var :
assert ( isinstance ( node , mparser . TernaryNode ) )
result = self . evaluate_statement ( node . condition )
if isinstance ( result , Disabler ) :
return result
if not isinstance ( result , bool ) :
raise InterpreterException ( ' Ternary condition is not boolean. ' )
if result :
return self . evaluate_statement ( node . trueblock )
else :
return self . evaluate_statement ( node . falseblock )
def evaluate_foreach ( self , node : mparser . ForeachClauseNode ) - > None :
assert ( isinstance ( node , mparser . ForeachClauseNode ) )
items = self . evaluate_statement ( node . items )
if isinstance ( items , list ) :
if len ( node . varnames ) != 1 :
raise InvalidArguments ( ' Foreach on array does not unpack ' )
varname = node . varnames [ 0 ]
for item in items :
self . set_variable ( varname , item )
try :
self . evaluate_codeblock ( node . block )
except ContinueRequest :
continue
except BreakRequest :
break
elif isinstance ( items , dict ) :
if len ( node . varnames ) != 2 :
raise InvalidArguments ( ' Foreach on dict unpacks key and value ' )
for key , value in sorted ( items . items ( ) ) :
self . set_variable ( node . varnames [ 0 ] , key )
self . set_variable ( node . varnames [ 1 ] , value )
try :
self . evaluate_codeblock ( node . block )
except ContinueRequest :
continue
except BreakRequest :
break
else :
raise InvalidArguments ( ' Items of foreach loop must be an array or a dict ' )
def evaluate_plusassign ( self , node : mparser . PlusAssignmentNode ) - > None :
assert ( isinstance ( node , mparser . PlusAssignmentNode ) )
varname = node . var_name
addition = self . evaluate_statement ( node . value )
# Remember that all variables are immutable. We must always create a
# full new variable and then assign it.
old_variable = self . get_variable ( varname )
new_value = None # type: T.Union[str, int, float, bool, dict, list]
if isinstance ( old_variable , str ) :
if not isinstance ( addition , str ) :
raise InvalidArguments ( ' The += operator requires a string on the right hand side if the variable on the left is a string ' )
new_value = old_variable + addition
elif isinstance ( old_variable , int ) :
if not isinstance ( addition , int ) :
raise InvalidArguments ( ' The += operator requires an int on the right hand side if the variable on the left is an int ' )
new_value = old_variable + addition
elif isinstance ( old_variable , list ) :
if isinstance ( addition , list ) :
new_value = old_variable + addition
else :
new_value = old_variable + [ addition ]
elif isinstance ( old_variable , dict ) :
if not isinstance ( addition , dict ) :
raise InvalidArguments ( ' The += operator requires a dict on the right hand side if the variable on the left is a dict ' )
new_value = { * * old_variable , * * addition }
# Add other data types here.
else :
raise InvalidArguments ( ' The += operator currently only works with arrays, dicts, strings or ints ' )
self . set_variable ( varname , new_value )
def evaluate_indexing ( self , node : mparser . IndexNode ) - > TYPE_var :
assert ( isinstance ( node , mparser . IndexNode ) )
iobject = self . evaluate_statement ( node . iobject )
if isinstance ( iobject , Disabler ) :
return iobject
if not hasattr ( iobject , ' __getitem__ ' ) :
raise InterpreterException (
' Tried to index an object that doesn \' t support indexing. ' )
index = self . evaluate_statement ( node . index )
if isinstance ( iobject , dict ) :
if not isinstance ( index , str ) :
raise InterpreterException ( ' Key is not a string ' )
try :
# The cast is required because we don't have recursive types...
return T . cast ( TYPE_var , iobject [ index ] )
except KeyError :
raise InterpreterException ( ' Key %s is not in dict ' % index )
else :
if not isinstance ( index , int ) :
raise InterpreterException ( ' Index value is not an integer. ' )
try :
# Ignore the MyPy error, since we don't know all indexable types here
# and we handle non indexable types with an exception
# TODO maybe find a better solution
return iobject [ index ] # type: ignore
except IndexError :
# We are already checking for the existance of __getitem__, so this should be save
raise InterpreterException ( ' Index %d out of bounds of array of size %d . ' % ( index , len ( iobject ) ) ) # type: ignore
def function_call ( self , node : mparser . FunctionNode ) - > T . Optional [ TYPE_var ] :
func_name = node . func_name
( posargs , kwargs ) = self . reduce_arguments ( node . args )
if is_disabled ( posargs , kwargs ) and func_name not in { ' get_variable ' , ' set_variable ' , ' is_disabler ' } :
return Disabler ( )
if func_name in self . funcs :
func = self . funcs [ func_name ]
func_args = posargs # type: T.Any
if not getattr ( func , ' no-args-flattening ' , False ) :
func_args = flatten ( posargs )
return func ( node , func_args , kwargs )
else :
self . unknown_function_called ( func_name )
return None
def method_call ( self , node : mparser . MethodNode ) - > TYPE_var :
invokable = node . source_object
if isinstance ( invokable , mparser . IdNode ) :
object_name = invokable . value
obj = self . get_variable ( object_name )
else :
obj = self . evaluate_statement ( invokable )
method_name = node . name
( args , kwargs ) = self . reduce_arguments ( node . args )
if is_disabled ( args , kwargs ) :
return Disabler ( )
if isinstance ( obj , str ) :
return self . string_method_call ( obj , method_name , args , kwargs )
if isinstance ( obj , bool ) :
return self . bool_method_call ( obj , method_name , args , kwargs )
if isinstance ( obj , int ) :
return self . int_method_call ( obj , method_name , args , kwargs )
if isinstance ( obj , list ) :
return self . array_method_call ( obj , method_name , args , kwargs )
if isinstance ( obj , dict ) :
return self . dict_method_call ( obj , method_name , args , kwargs )
if isinstance ( obj , mesonlib . File ) :
raise InvalidArguments ( ' File object " %s " is not callable. ' % obj )
if not isinstance ( obj , InterpreterObject ) :
raise InvalidArguments ( ' Variable " %s " is not callable. ' % object_name )
# Special case. This is the only thing you can do with a disabler
# object. Every other use immediately returns the disabler object.
if isinstance ( obj , Disabler ) :
if method_name == ' found ' :
return False
else :
return Disabler ( )
if method_name == ' extract_objects ' :
if not isinstance ( obj , ObjectHolder ) :
raise InvalidArguments ( ' Invalid operation " extract_objects " on variable " {} " ' . format ( object_name ) )
self . validate_extraction ( obj . held_object )
obj . current_node = node
return obj . method_call ( method_name , args , kwargs )
@builtinMethodNoKwargs
def bool_method_call ( self , obj : bool , method_name : str , posargs : T . List [ TYPE_nvar ] , kwargs : T . Dict [ str , T . Any ] ) - > T . Union [ str , int ] :
if method_name == ' to_string ' :
Don't use len() to test emptiness vs not emptiness
Meson has a common pattern of using 'if len(foo) == 0:' or
'if len(foo) != 0:', however, this is a common anti-pattern in python.
Instead tests for emptiness/non-emptiness should be done with a simple
'if foo:' or 'if not foo:'
Consider the following:
>>> import timeit
>>> timeit.timeit('if len([]) == 0: pass')
0.10730923599840025
>>> timeit.timeit('if not []: pass')
0.030033907998586074
>>> timeit.timeit('if len(['a', 'b', 'c', 'd']) == 0: pass')
0.1154778649979562
>>> timeit.timeit("if not ['a', 'b', 'c', 'd']: pass")
0.08259823200205574
>>> timeit.timeit('if len("") == 0: pass')
0.089759664999292
>>> timeit.timeit('if not "": pass')
0.02340641999762738
>>> timeit.timeit('if len("foo") == 0: pass')
0.08848102600313723
>>> timeit.timeit('if not "foo": pass')
0.04032287199879647
And for the one additional case of 'if len(foo.strip()) == 0', which can
be replaced with 'if not foo.isspace()'
>>> timeit.timeit('if len(" ".strip()) == 0: pass')
0.15294511600222904
>>> timeit.timeit('if " ".isspace(): pass')
0.09413968399894657
>>> timeit.timeit('if len(" abc".strip()) == 0: pass')
0.2023209120015963
>>> timeit.timeit('if " abc".isspace(): pass')
0.09571301700270851
In other words, it's always a win to not use len(), when you don't
actually want to check the length.
8 years ago
if not posargs :
if obj :
return ' true '
else :
return ' false '
elif len ( posargs ) == 2 and isinstance ( posargs [ 0 ] , str ) and isinstance ( posargs [ 1 ] , str ) :
if obj :
return posargs [ 0 ]
else :
return posargs [ 1 ]
else :
raise InterpreterException ( ' bool.to_string() must have either no arguments or exactly two string arguments that signify what values to return for true and false. ' )
elif method_name == ' to_int ' :
if obj :
return 1
else :
return 0
else :
raise InterpreterException ( ' Unknown method " %s " for a boolean. ' % method_name )
@builtinMethodNoKwargs
def int_method_call ( self , obj : int , method_name : str , posargs : T . List [ TYPE_nvar ] , kwargs : T . Dict [ str , T . Any ] ) - > T . Union [ str , bool ] :
if method_name == ' is_even ' :
Don't use len() to test emptiness vs not emptiness
Meson has a common pattern of using 'if len(foo) == 0:' or
'if len(foo) != 0:', however, this is a common anti-pattern in python.
Instead tests for emptiness/non-emptiness should be done with a simple
'if foo:' or 'if not foo:'
Consider the following:
>>> import timeit
>>> timeit.timeit('if len([]) == 0: pass')
0.10730923599840025
>>> timeit.timeit('if not []: pass')
0.030033907998586074
>>> timeit.timeit('if len(['a', 'b', 'c', 'd']) == 0: pass')
0.1154778649979562
>>> timeit.timeit("if not ['a', 'b', 'c', 'd']: pass")
0.08259823200205574
>>> timeit.timeit('if len("") == 0: pass')
0.089759664999292
>>> timeit.timeit('if not "": pass')
0.02340641999762738
>>> timeit.timeit('if len("foo") == 0: pass')
0.08848102600313723
>>> timeit.timeit('if not "foo": pass')
0.04032287199879647
And for the one additional case of 'if len(foo.strip()) == 0', which can
be replaced with 'if not foo.isspace()'
>>> timeit.timeit('if len(" ".strip()) == 0: pass')
0.15294511600222904
>>> timeit.timeit('if " ".isspace(): pass')
0.09413968399894657
>>> timeit.timeit('if len(" abc".strip()) == 0: pass')
0.2023209120015963
>>> timeit.timeit('if " abc".isspace(): pass')
0.09571301700270851
In other words, it's always a win to not use len(), when you don't
actually want to check the length.
8 years ago
if not posargs :
return obj % 2 == 0
else :
raise InterpreterException ( ' int.is_even() must have no arguments. ' )
elif method_name == ' is_odd ' :
Don't use len() to test emptiness vs not emptiness
Meson has a common pattern of using 'if len(foo) == 0:' or
'if len(foo) != 0:', however, this is a common anti-pattern in python.
Instead tests for emptiness/non-emptiness should be done with a simple
'if foo:' or 'if not foo:'
Consider the following:
>>> import timeit
>>> timeit.timeit('if len([]) == 0: pass')
0.10730923599840025
>>> timeit.timeit('if not []: pass')
0.030033907998586074
>>> timeit.timeit('if len(['a', 'b', 'c', 'd']) == 0: pass')
0.1154778649979562
>>> timeit.timeit("if not ['a', 'b', 'c', 'd']: pass")
0.08259823200205574
>>> timeit.timeit('if len("") == 0: pass')
0.089759664999292
>>> timeit.timeit('if not "": pass')
0.02340641999762738
>>> timeit.timeit('if len("foo") == 0: pass')
0.08848102600313723
>>> timeit.timeit('if not "foo": pass')
0.04032287199879647
And for the one additional case of 'if len(foo.strip()) == 0', which can
be replaced with 'if not foo.isspace()'
>>> timeit.timeit('if len(" ".strip()) == 0: pass')
0.15294511600222904
>>> timeit.timeit('if " ".isspace(): pass')
0.09413968399894657
>>> timeit.timeit('if len(" abc".strip()) == 0: pass')
0.2023209120015963
>>> timeit.timeit('if " abc".isspace(): pass')
0.09571301700270851
In other words, it's always a win to not use len(), when you don't
actually want to check the length.
8 years ago
if not posargs :
return obj % 2 != 0
else :
raise InterpreterException ( ' int.is_odd() must have no arguments. ' )
elif method_name == ' to_string ' :
if not posargs :
return str ( obj )
else :
raise InterpreterException ( ' int.to_string() must have no arguments. ' )
else :
raise InterpreterException ( ' Unknown method " %s " for an integer. ' % method_name )
@staticmethod
def _get_one_string_posarg ( posargs : T . List [ TYPE_nvar ] , method_name : str ) - > str :
if len ( posargs ) > 1 :
m = ' {} () must have zero or one arguments '
raise InterpreterException ( m . format ( method_name ) )
elif len ( posargs ) == 1 :
s = posargs [ 0 ]
if not isinstance ( s , str ) :
m = ' {} () argument must be a string '
raise InterpreterException ( m . format ( method_name ) )
return s
return None
@builtinMethodNoKwargs
def string_method_call ( self , obj : str , method_name : str , posargs : T . List [ TYPE_nvar ] , kwargs : T . Dict [ str , T . Any ] ) - > T . Union [ str , int , bool , T . List [ str ] ] :
if method_name == ' strip ' :
s1 = self . _get_one_string_posarg ( posargs , ' strip ' )
if s1 is not None :
return obj . strip ( s1 )
return obj . strip ( )
elif method_name == ' format ' :
return self . format_string ( obj , posargs )
elif method_name == ' to_upper ' :
return obj . upper ( )
elif method_name == ' to_lower ' :
return obj . lower ( )
elif method_name == ' underscorify ' :
return re . sub ( r ' [^a-zA-Z0-9] ' , ' _ ' , obj )
elif method_name == ' split ' :
s2 = self . _get_one_string_posarg ( posargs , ' split ' )
if s2 is not None :
return obj . split ( s2 )
return obj . split ( )
elif method_name == ' startswith ' or method_name == ' contains ' or method_name == ' endswith ' :
s3 = posargs [ 0 ]
if not isinstance ( s3 , str ) :
raise InterpreterException ( ' Argument must be a string. ' )
if method_name == ' startswith ' :
return obj . startswith ( s3 )
elif method_name == ' contains ' :
return obj . find ( s3 ) > = 0
return obj . endswith ( s3 )
elif method_name == ' to_int ' :
try :
return int ( obj )
except Exception :
raise InterpreterException ( ' String {!r} cannot be converted to int ' . format ( obj ) )
elif method_name == ' join ' :
if len ( posargs ) != 1 :
raise InterpreterException ( ' Join() takes exactly one argument. ' )
strlist = posargs [ 0 ]
check_stringlist ( strlist )
assert isinstance ( strlist , list ) # Required for mypy
return obj . join ( strlist )
elif method_name == ' version_compare ' :
if len ( posargs ) != 1 :
raise InterpreterException ( ' Version_compare() takes exactly one argument. ' )
cmpr = posargs [ 0 ]
if not isinstance ( cmpr , str ) :
raise InterpreterException ( ' Version_compare() argument must be a string. ' )
if isinstance ( obj , MesonVersionString ) :
self . tmp_meson_version = cmpr
return mesonlib . version_compare ( obj , cmpr )
elif method_name == ' substring ' :
if len ( posargs ) > 2 :
raise InterpreterException ( ' substring() takes maximum two arguments. ' )
start = 0
end = len ( obj )
if len ( posargs ) > 0 :
if not isinstance ( posargs [ 0 ] , int ) :
raise InterpreterException ( ' substring() argument must be an int ' )
start = posargs [ 0 ]
if len ( posargs ) > 1 :
if not isinstance ( posargs [ 1 ] , int ) :
raise InterpreterException ( ' substring() argument must be an int ' )
end = posargs [ 1 ]
return obj [ start : end ]
raise InterpreterException ( ' Unknown method " %s " for a string. ' % method_name )
def format_string ( self , templ : str , args : T . List [ TYPE_nvar ] ) - > str :
arg_strings = [ ]
for arg in args :
if isinstance ( arg , mparser . BaseNode ) :
arg = self . evaluate_statement ( arg )
if isinstance ( arg , bool ) : # Python boolean is upper case.
arg = str ( arg ) . lower ( )
arg_strings . append ( str ( arg ) )
def arg_replace ( match : T . Match [ str ] ) - > str :
idx = int ( match . group ( 1 ) )
if idx > = len ( arg_strings ) :
raise InterpreterException ( ' Format placeholder @ {} @ out of range. ' . format ( idx ) )
return arg_strings [ idx ]
return re . sub ( r ' @( \ d+)@ ' , arg_replace , templ )
def unknown_function_called ( self , func_name : str ) - > None :
raise InvalidCode ( ' Unknown function " %s " . ' % func_name )
@builtinMethodNoKwargs
def array_method_call ( self , obj : T . List [ TYPE_var ] , method_name : str , posargs : T . List [ TYPE_nvar ] , kwargs : T . Dict [ str , T . Any ] ) - > TYPE_var :
if method_name == ' contains ' :
def check_contains ( el : list ) - > bool :
if len ( posargs ) != 1 :
raise InterpreterException ( ' Contains method takes exactly one argument. ' )
item = posargs [ 0 ]
for element in el :
if isinstance ( element , list ) :
found = check_contains ( element )
if found :
return True
if element == item :
return True
return False
return check_contains ( obj )
elif method_name == ' length ' :
return len ( obj )
elif method_name == ' get ' :
index = posargs [ 0 ]
fallback = None
if len ( posargs ) == 2 :
fallback = posargs [ 1 ]
elif len ( posargs ) > 2 :
m = ' Array method \' get() \' only takes two arguments: the ' \
' index and an optional fallback value if the index is ' \
' out of range. '
raise InvalidArguments ( m )
if not isinstance ( index , int ) :
raise InvalidArguments ( ' Array index must be a number. ' )
if index < - len ( obj ) or index > = len ( obj ) :
if fallback is None :
m = ' Array index {!r} is out of bounds for array of size {!r} . '
raise InvalidArguments ( m . format ( index , len ( obj ) ) )
if isinstance ( fallback , mparser . BaseNode ) :
return self . evaluate_statement ( fallback )
return fallback
return obj [ index ]
m = ' Arrays do not have a method called {!r} . '
raise InterpreterException ( m . format ( method_name ) )
@builtinMethodNoKwargs
def dict_method_call ( self , obj : T . Dict [ str , TYPE_var ] , method_name : str , posargs : T . List [ TYPE_nvar ] , kwargs : T . Dict [ str , T . Any ] ) - > TYPE_var :
if method_name in ( ' has_key ' , ' get ' ) :
if method_name == ' has_key ' :
if len ( posargs ) != 1 :
raise InterpreterException ( ' has_key() takes exactly one argument. ' )
else :
if len ( posargs ) not in ( 1 , 2 ) :
raise InterpreterException ( ' get() takes one or two arguments. ' )
key = posargs [ 0 ]
if not isinstance ( key , ( str ) ) :
raise InvalidArguments ( ' Dictionary key must be a string. ' )
has_key = key in obj
if method_name == ' has_key ' :
return has_key
if has_key :
return obj [ key ]
if len ( posargs ) == 2 :
fallback = posargs [ 1 ]
if isinstance ( fallback , mparser . BaseNode ) :
return self . evaluate_statement ( fallback )
return fallback
raise InterpreterException ( ' Key {!r} is not in the dictionary. ' . format ( key ) )
if method_name == ' keys ' :
if len ( posargs ) != 0 :
raise InterpreterException ( ' keys() takes no arguments. ' )
return sorted ( obj . keys ( ) )
raise InterpreterException ( ' Dictionaries do not have a method called " %s " . ' % method_name )
def reduce_arguments (
self ,
args : mparser . ArgumentNode ,
key_resolver : T . Callable [ [ mparser . BaseNode ] , str ] = default_resolve_key ,
duplicate_key_error : T . Optional [ str ] = None ,
) - > T . Tuple [ T . List [ TYPE_nvar ] , TYPE_nkwargs ] :
assert ( isinstance ( args , mparser . ArgumentNode ) )
if args . incorrect_order ( ) :
raise InvalidArguments ( ' All keyword arguments must be after positional arguments. ' )
self . argument_depth + = 1
reduced_pos = [ self . evaluate_statement ( arg ) for arg in args . arguments ] # type: T.List[TYPE_nvar]
reduced_kw = { } # type: TYPE_nkwargs
for key , val in args . kwargs . items ( ) :
reduced_key = key_resolver ( key )
reduced_val = val # type: TYPE_nvar
if isinstance ( reduced_val , mparser . BaseNode ) :
reduced_val = self . evaluate_statement ( reduced_val )
if duplicate_key_error and reduced_key in reduced_kw :
raise InvalidArguments ( duplicate_key_error . format ( reduced_key ) )
reduced_kw [ reduced_key ] = reduced_val
self . argument_depth - = 1
final_kw = self . expand_default_kwargs ( reduced_kw )
return reduced_pos , final_kw
def expand_default_kwargs ( self , kwargs : TYPE_nkwargs ) - > TYPE_nkwargs :
if ' kwargs ' not in kwargs :
return kwargs
to_expand = kwargs . pop ( ' kwargs ' )
if not isinstance ( to_expand , dict ) :
raise InterpreterException ( ' Value of " kwargs " must be dictionary. ' )
if ' kwargs ' in to_expand :
raise InterpreterException ( ' Kwargs argument must not contain a " kwargs " entry. Points for thinking meta, though. :P ' )
for k , v in to_expand . items ( ) :
if k in kwargs :
raise InterpreterException ( ' Entry " {} " defined both as a keyword argument and in a " kwarg " entry. ' . format ( k ) )
kwargs [ k ] = v
return kwargs
def assignment ( self , node : mparser . AssignmentNode ) - > None :
assert ( isinstance ( node , mparser . AssignmentNode ) )
if self . argument_depth != 0 :
raise InvalidArguments ( ''' Tried to assign values inside an argument list.
To specify a keyword argument , use : instead of = . ''' )
var_name = node . var_name
if not isinstance ( var_name , str ) :
raise InvalidArguments ( ' Tried to assign value to a non-variable. ' )
value = self . evaluate_statement ( node . value )
if not self . is_assignable ( value ) :
raise InvalidCode ( ' Tried to assign an invalid value to variable. ' )
# For mutable objects we need to make a copy on assignment
if isinstance ( value , MutableInterpreterObject ) :
value = copy . deepcopy ( value )
self . set_variable ( var_name , value )
return None
def set_variable ( self , varname : str , variable : TYPE_var ) - > None :
if variable is None :
raise InvalidCode ( ' Can not assign None to variable. ' )
if not isinstance ( varname , str ) :
raise InvalidCode ( ' First argument to set_variable must be a string. ' )
if not self . is_assignable ( variable ) :
raise InvalidCode ( ' Assigned value not of assignable type. ' )
if re . match ( ' [_a-zA-Z][_0-9a-zA-Z]*$ ' , varname ) is None :
raise InvalidCode ( ' Invalid variable name: ' + varname )
if varname in self . builtin :
raise InvalidCode ( ' Tried to overwrite internal variable " %s " ' % varname )
self . variables [ varname ] = variable
def get_variable ( self , varname : str ) - > TYPE_var :
if varname in self . builtin :
return self . builtin [ varname ]
if varname in self . variables :
return self . variables [ varname ]
raise InvalidCode ( ' Unknown variable " %s " . ' % varname )
def is_assignable ( self , value : T . Any ) - > bool :
return isinstance ( value , ( InterpreterObject , dependencies . Dependency ,
str , int , list , dict , mesonlib . File ) )
def validate_extraction ( self , buildtarget : InterpreterObject ) - > None :
raise InterpreterException ( ' validate_extraction is not implemented in this context (please file a bug) ' )