952 lines
30 KiB
Python
952 lines
30 KiB
Python
"""SCons.Util
|
|
|
|
Various utility functions go here.
|
|
|
|
"""
|
|
|
|
#
|
|
# Copyright (c) 2001, 2002, 2003, 2004, 2005, 2006, 2007 The SCons Foundation
|
|
#
|
|
# Permission is hereby granted, free of charge, to any person obtaining
|
|
# a copy of this software and associated documentation files (the
|
|
# "Software"), to deal in the Software without restriction, including
|
|
# without limitation the rights to use, copy, modify, merge, publish,
|
|
# distribute, sublicense, and/or sell copies of the Software, and to
|
|
# permit persons to whom the Software is furnished to do so, subject to
|
|
# the following conditions:
|
|
#
|
|
# The above copyright notice and this permission notice shall be included
|
|
# in all copies or substantial portions of the Software.
|
|
#
|
|
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY
|
|
# KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE
|
|
# WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
|
|
# NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
|
|
# LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
|
|
# OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
|
|
# WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
|
|
#
|
|
|
|
__revision__ = "/home/scons/scons/branch.0/baseline/src/engine/SCons/Util.py 0.97.D001 2007/05/17 11:35:19 knight"
|
|
|
|
import SCons.compat
|
|
|
|
import copy
|
|
import os
|
|
import os.path
|
|
import re
|
|
import string
|
|
import sys
|
|
import types
|
|
|
|
from UserDict import UserDict
|
|
from UserList import UserList
|
|
from UserString import UserString
|
|
|
|
# Don't "from types import ..." these because we need to get at the
|
|
# types module later to look for UnicodeType.
|
|
DictType = types.DictType
|
|
InstanceType = types.InstanceType
|
|
ListType = types.ListType
|
|
StringType = types.StringType
|
|
TupleType = types.TupleType
|
|
|
|
_altsep = os.altsep
|
|
if _altsep is None and sys.platform == 'win32':
|
|
# My ActivePython 2.0.1 doesn't set os.altsep! What gives?
|
|
_altsep = '/'
|
|
if _altsep:
|
|
def rightmost_separator(path, sep, _altsep=_altsep):
|
|
rfind = string.rfind
|
|
return max(rfind(path, sep), rfind(path, _altsep))
|
|
else:
|
|
rightmost_separator = string.rfind
|
|
|
|
# First two from the Python Cookbook, just for completeness.
|
|
# (Yeah, yeah, YAGNI...)
|
|
def containsAny(str, set):
|
|
"""Check whether sequence str contains ANY of the items in set."""
|
|
for c in set:
|
|
if c in str: return 1
|
|
return 0
|
|
|
|
def containsAll(str, set):
|
|
"""Check whether sequence str contains ALL of the items in set."""
|
|
for c in set:
|
|
if c not in str: return 0
|
|
return 1
|
|
|
|
def containsOnly(str, set):
|
|
"""Check whether sequence str contains ONLY items in set."""
|
|
for c in str:
|
|
if c not in set: return 0
|
|
return 1
|
|
|
|
def splitext(path):
|
|
"Same as os.path.splitext() but faster."
|
|
sep = rightmost_separator(path, os.sep)
|
|
dot = string.rfind(path, '.')
|
|
# An ext is only real if it has at least one non-digit char
|
|
if dot > sep and not containsOnly(path[dot:], "0123456789."):
|
|
return path[:dot],path[dot:]
|
|
else:
|
|
return path,""
|
|
|
|
def updrive(path):
|
|
"""
|
|
Make the drive letter (if any) upper case.
|
|
This is useful because Windows is inconsitent on the case
|
|
of the drive letter, which can cause inconsistencies when
|
|
calculating command signatures.
|
|
"""
|
|
drive, rest = os.path.splitdrive(path)
|
|
if drive:
|
|
path = string.upper(drive) + rest
|
|
return path
|
|
|
|
#
|
|
# Generic convert-to-string functions that abstract away whether or
|
|
# not the Python we're executing has Unicode support. The wrapper
|
|
# to_String_for_signature() will use a for_signature() method if the
|
|
# specified object has one.
|
|
#
|
|
if hasattr(types, 'UnicodeType'):
|
|
UnicodeType = types.UnicodeType
|
|
def to_String(s):
|
|
if isinstance(s, UserString):
|
|
t = type(s.data)
|
|
else:
|
|
t = type(s)
|
|
if t is UnicodeType:
|
|
return unicode(s)
|
|
else:
|
|
return str(s)
|
|
else:
|
|
to_String = str
|
|
|
|
def to_String_for_signature(obj):
|
|
try:
|
|
f = obj.for_signature
|
|
except AttributeError:
|
|
return to_String(obj)
|
|
else:
|
|
return f()
|
|
|
|
class CallableComposite(UserList):
|
|
"""A simple composite callable class that, when called, will invoke all
|
|
of its contained callables with the same arguments."""
|
|
def __call__(self, *args, **kwargs):
|
|
retvals = map(lambda x, args=args, kwargs=kwargs: apply(x,
|
|
args,
|
|
kwargs),
|
|
self.data)
|
|
if self.data and (len(self.data) == len(filter(callable, retvals))):
|
|
return self.__class__(retvals)
|
|
return NodeList(retvals)
|
|
|
|
class NodeList(UserList):
|
|
"""This class is almost exactly like a regular list of Nodes
|
|
(actually it can hold any object), with one important difference.
|
|
If you try to get an attribute from this list, it will return that
|
|
attribute from every item in the list. For example:
|
|
|
|
>>> someList = NodeList([ ' foo ', ' bar ' ])
|
|
>>> someList.strip()
|
|
[ 'foo', 'bar' ]
|
|
"""
|
|
def __nonzero__(self):
|
|
return len(self.data) != 0
|
|
|
|
def __str__(self):
|
|
return string.join(map(str, self.data))
|
|
|
|
def __getattr__(self, name):
|
|
if not self.data:
|
|
# If there is nothing in the list, then we have no attributes to
|
|
# pass through, so raise AttributeError for everything.
|
|
raise AttributeError, "NodeList has no attribute: %s" % name
|
|
|
|
# Return a list of the attribute, gotten from every element
|
|
# in the list
|
|
attrList = map(lambda x, n=name: getattr(x, n), self.data)
|
|
|
|
# Special case. If the attribute is callable, we do not want
|
|
# to return a list of callables. Rather, we want to return a
|
|
# single callable that, when called, will invoke the function on
|
|
# all elements of this list.
|
|
if self.data and (len(self.data) == len(filter(callable, attrList))):
|
|
return CallableComposite(attrList)
|
|
return self.__class__(attrList)
|
|
|
|
_valid_var = re.compile(r'[_a-zA-Z]\w*$')
|
|
_get_env_var = re.compile(r'^\$([_a-zA-Z]\w*|{[_a-zA-Z]\w*})$')
|
|
|
|
def is_valid_construction_var(varstr):
|
|
"""Return if the specified string is a legitimate construction
|
|
variable.
|
|
"""
|
|
return _valid_var.match(varstr)
|
|
|
|
def get_environment_var(varstr):
|
|
"""Given a string, first determine if it looks like a reference
|
|
to a single environment variable, like "$FOO" or "${FOO}".
|
|
If so, return that variable with no decorations ("FOO").
|
|
If not, return None."""
|
|
mo=_get_env_var.match(to_String(varstr))
|
|
if mo:
|
|
var = mo.group(1)
|
|
if var[0] == '{':
|
|
return var[1:-1]
|
|
else:
|
|
return var
|
|
else:
|
|
return None
|
|
|
|
class DisplayEngine:
|
|
def __init__(self):
|
|
self.__call__ = self.print_it
|
|
|
|
def print_it(self, text, append_newline=1):
|
|
if append_newline: text = text + '\n'
|
|
sys.stdout.write(text)
|
|
|
|
def dont_print(self, text, append_newline=1):
|
|
pass
|
|
|
|
def set_mode(self, mode):
|
|
if mode:
|
|
self.__call__ = self.print_it
|
|
else:
|
|
self.__call__ = self.dont_print
|
|
|
|
def render_tree(root, child_func, prune=0, margin=[0], visited={}):
|
|
"""
|
|
Render a tree of nodes into an ASCII tree view.
|
|
root - the root node of the tree
|
|
child_func - the function called to get the children of a node
|
|
prune - don't visit the same node twice
|
|
margin - the format of the left margin to use for children of root.
|
|
1 results in a pipe, and 0 results in no pipe.
|
|
visited - a dictionary of visited nodes in the current branch if not prune,
|
|
or in the whole tree if prune.
|
|
"""
|
|
|
|
rname = str(root)
|
|
|
|
children = child_func(root)
|
|
retval = ""
|
|
for pipe in margin[:-1]:
|
|
if pipe:
|
|
retval = retval + "| "
|
|
else:
|
|
retval = retval + " "
|
|
|
|
if visited.has_key(rname):
|
|
return retval + "+-[" + rname + "]\n"
|
|
|
|
retval = retval + "+-" + rname + "\n"
|
|
if not prune:
|
|
visited = copy.copy(visited)
|
|
visited[rname] = 1
|
|
|
|
for i in range(len(children)):
|
|
margin.append(i<len(children)-1)
|
|
retval = retval + render_tree(children[i], child_func, prune, margin, visited
|
|
)
|
|
margin.pop()
|
|
|
|
return retval
|
|
|
|
IDX = lambda N: N and 1 or 0
|
|
|
|
def print_tree(root, child_func, prune=0, showtags=0, margin=[0], visited={}):
|
|
"""
|
|
Print a tree of nodes. This is like render_tree, except it prints
|
|
lines directly instead of creating a string representation in memory,
|
|
so that huge trees can be printed.
|
|
|
|
root - the root node of the tree
|
|
child_func - the function called to get the children of a node
|
|
prune - don't visit the same node twice
|
|
showtags - print status information to the left of each node line
|
|
margin - the format of the left margin to use for children of root.
|
|
1 results in a pipe, and 0 results in no pipe.
|
|
visited - a dictionary of visited nodes in the current branch if not prune,
|
|
or in the whole tree if prune.
|
|
"""
|
|
|
|
rname = str(root)
|
|
|
|
if showtags:
|
|
|
|
if showtags == 2:
|
|
print ' E = exists'
|
|
print ' R = exists in repository only'
|
|
print ' b = implicit builder'
|
|
print ' B = explicit builder'
|
|
print ' S = side effect'
|
|
print ' P = precious'
|
|
print ' A = always build'
|
|
print ' C = current'
|
|
print ' N = no clean'
|
|
print ' H = no cache'
|
|
print ''
|
|
|
|
tags = ['[']
|
|
tags.append(' E'[IDX(root.exists())])
|
|
tags.append(' R'[IDX(root.rexists() and not root.exists())])
|
|
tags.append(' BbB'[[0,1][IDX(root.has_explicit_builder())] +
|
|
[0,2][IDX(root.has_builder())]])
|
|
tags.append(' S'[IDX(root.side_effect)])
|
|
tags.append(' P'[IDX(root.precious)])
|
|
tags.append(' A'[IDX(root.always_build)])
|
|
tags.append(' C'[IDX(root.current())])
|
|
tags.append(' N'[IDX(root.noclean)])
|
|
tags.append(' H'[IDX(root.nocache)])
|
|
tags.append(']')
|
|
|
|
else:
|
|
tags = []
|
|
|
|
def MMM(m):
|
|
return [" ","| "][m]
|
|
margins = map(MMM, margin[:-1])
|
|
|
|
if visited.has_key(rname):
|
|
print string.join(tags + margins + ['+-[', rname, ']'], '')
|
|
return
|
|
|
|
print string.join(tags + margins + ['+-', rname], '')
|
|
|
|
if prune:
|
|
visited[rname] = 1
|
|
|
|
children = child_func(root)
|
|
if children:
|
|
margin.append(1)
|
|
map(lambda C, cf=child_func, p=prune, i=IDX(showtags), m=margin, v=visited:
|
|
print_tree(C, cf, p, i, m, v),
|
|
children[:-1])
|
|
margin[-1] = 0
|
|
print_tree(children[-1], child_func, prune, IDX(showtags), margin, visited)
|
|
margin.pop()
|
|
|
|
|
|
|
|
# Functions for deciding if things are like various types, mainly to
|
|
# handle UserDict, UserList and UserString like their underlying types.
|
|
#
|
|
# Yes, all of this manual testing breaks polymorphism, and the real
|
|
# Pythonic way to do all of this would be to just try it and handle the
|
|
# exception, but handling the exception when it's not the right type is
|
|
# too slow.
|
|
#
|
|
# The actual implementations here have been selected after timings
|
|
# coded up in in bench/is_types.py (from the SCons source tree, see the
|
|
# scons-src distribution). Key results from those timings:
|
|
#
|
|
# -- Storing the type of the object in a variable (t = type(obj))
|
|
# slows down the case where it's a native type and the first
|
|
# comparison will match, but nicely speeds up the case where
|
|
# it's a different native type. Since that's going to be common,
|
|
# it's a good tradeoff.
|
|
#
|
|
# -- The data show that calling isinstance() on an object that's
|
|
# a native type (dict, list or string) is expensive enough that
|
|
# checking up front for whether the object is of type InstanceType
|
|
# is a pretty big win, even though it does slow down the case
|
|
# where it really *is* an object instance a little bit.
|
|
|
|
def is_Dict(obj):
|
|
t = type(obj)
|
|
return t is DictType or \
|
|
(t is InstanceType and isinstance(obj, UserDict))
|
|
|
|
def is_List(obj):
|
|
t = type(obj)
|
|
return t is ListType \
|
|
or (t is InstanceType and isinstance(obj, UserList))
|
|
|
|
def is_Sequence(obj):
|
|
t = type(obj)
|
|
return t is ListType \
|
|
or t is TupleType \
|
|
or (t is InstanceType and isinstance(obj, UserList))
|
|
|
|
def is_Tuple(obj):
|
|
t = type(obj)
|
|
return t is TupleType
|
|
|
|
if hasattr(types, 'UnicodeType'):
|
|
def is_String(obj):
|
|
t = type(obj)
|
|
return t is StringType \
|
|
or t is UnicodeType \
|
|
or (t is InstanceType and isinstance(obj, UserString))
|
|
else:
|
|
def is_String(obj):
|
|
t = type(obj)
|
|
return t is StringType \
|
|
or (t is InstanceType and isinstance(obj, UserString))
|
|
|
|
|
|
|
|
def is_Scalar(e):
|
|
return is_String(e) or (not is_List(e) and not is_Tuple(e))
|
|
|
|
def flatten(sequence, scalarp=is_Scalar, result=None):
|
|
if result is None:
|
|
result = []
|
|
for item in sequence:
|
|
if scalarp(item):
|
|
result.append(item)
|
|
else:
|
|
flatten(item, scalarp, result)
|
|
return result
|
|
|
|
class Proxy:
|
|
"""A simple generic Proxy class, forwarding all calls to
|
|
subject. So, for the benefit of the python newbie, what does
|
|
this really mean? Well, it means that you can take an object, let's
|
|
call it 'objA', and wrap it in this Proxy class, with a statement
|
|
like this
|
|
|
|
proxyObj = Proxy(objA),
|
|
|
|
Then, if in the future, you do something like this
|
|
|
|
x = proxyObj.var1,
|
|
|
|
since Proxy does not have a 'var1' attribute (but presumably objA does),
|
|
the request actually is equivalent to saying
|
|
|
|
x = objA.var1
|
|
|
|
Inherit from this class to create a Proxy."""
|
|
|
|
def __init__(self, subject):
|
|
"""Wrap an object as a Proxy object"""
|
|
self.__subject = subject
|
|
|
|
def __getattr__(self, name):
|
|
"""Retrieve an attribute from the wrapped object. If the named
|
|
attribute doesn't exist, AttributeError is raised"""
|
|
return getattr(self.__subject, name)
|
|
|
|
def get(self):
|
|
"""Retrieve the entire wrapped object"""
|
|
return self.__subject
|
|
|
|
def __cmp__(self, other):
|
|
if issubclass(other.__class__, self.__subject.__class__):
|
|
return cmp(self.__subject, other)
|
|
return cmp(self.__dict__, other.__dict__)
|
|
|
|
# attempt to load the windows registry module:
|
|
can_read_reg = 0
|
|
try:
|
|
import _winreg
|
|
|
|
can_read_reg = 1
|
|
hkey_mod = _winreg
|
|
|
|
RegOpenKeyEx = _winreg.OpenKeyEx
|
|
RegEnumKey = _winreg.EnumKey
|
|
RegEnumValue = _winreg.EnumValue
|
|
RegQueryValueEx = _winreg.QueryValueEx
|
|
RegError = _winreg.error
|
|
|
|
except ImportError:
|
|
try:
|
|
import win32api
|
|
import win32con
|
|
can_read_reg = 1
|
|
hkey_mod = win32con
|
|
|
|
RegOpenKeyEx = win32api.RegOpenKeyEx
|
|
RegEnumKey = win32api.RegEnumKey
|
|
RegEnumValue = win32api.RegEnumValue
|
|
RegQueryValueEx = win32api.RegQueryValueEx
|
|
RegError = win32api.error
|
|
|
|
except ImportError:
|
|
class _NoError(Exception):
|
|
pass
|
|
RegError = _NoError
|
|
|
|
if can_read_reg:
|
|
HKEY_CLASSES_ROOT = hkey_mod.HKEY_CLASSES_ROOT
|
|
HKEY_LOCAL_MACHINE = hkey_mod.HKEY_LOCAL_MACHINE
|
|
HKEY_CURRENT_USER = hkey_mod.HKEY_CURRENT_USER
|
|
HKEY_USERS = hkey_mod.HKEY_USERS
|
|
|
|
def RegGetValue(root, key):
|
|
"""This utility function returns a value in the registry
|
|
without having to open the key first. Only available on
|
|
Windows platforms with a version of Python that can read the
|
|
registry. Returns the same thing as
|
|
SCons.Util.RegQueryValueEx, except you just specify the entire
|
|
path to the value, and don't have to bother opening the key
|
|
first. So:
|
|
|
|
Instead of:
|
|
k = SCons.Util.RegOpenKeyEx(SCons.Util.HKEY_LOCAL_MACHINE,
|
|
r'SOFTWARE\Microsoft\Windows\CurrentVersion')
|
|
out = SCons.Util.RegQueryValueEx(k,
|
|
'ProgramFilesDir')
|
|
|
|
You can write:
|
|
out = SCons.Util.RegGetValue(SCons.Util.HKEY_LOCAL_MACHINE,
|
|
r'SOFTWARE\Microsoft\Windows\CurrentVersion\ProgramFilesDir')
|
|
"""
|
|
# I would use os.path.split here, but it's not a filesystem
|
|
# path...
|
|
p = key.rfind('\\') + 1
|
|
keyp = key[:p]
|
|
val = key[p:]
|
|
k = RegOpenKeyEx(root, keyp)
|
|
return RegQueryValueEx(k,val)
|
|
|
|
if sys.platform == 'win32':
|
|
|
|
def WhereIs(file, path=None, pathext=None, reject=[]):
|
|
if path is None:
|
|
try:
|
|
path = os.environ['PATH']
|
|
except KeyError:
|
|
return None
|
|
if is_String(path):
|
|
path = string.split(path, os.pathsep)
|
|
if pathext is None:
|
|
try:
|
|
pathext = os.environ['PATHEXT']
|
|
except KeyError:
|
|
pathext = '.COM;.EXE;.BAT;.CMD'
|
|
if is_String(pathext):
|
|
pathext = string.split(pathext, os.pathsep)
|
|
for ext in pathext:
|
|
if string.lower(ext) == string.lower(file[-len(ext):]):
|
|
pathext = ['']
|
|
break
|
|
if not is_List(reject) and not is_Tuple(reject):
|
|
reject = [reject]
|
|
for dir in path:
|
|
f = os.path.join(dir, file)
|
|
for ext in pathext:
|
|
fext = f + ext
|
|
if os.path.isfile(fext):
|
|
try:
|
|
reject.index(fext)
|
|
except ValueError:
|
|
return os.path.normpath(fext)
|
|
continue
|
|
return None
|
|
|
|
elif os.name == 'os2':
|
|
|
|
def WhereIs(file, path=None, pathext=None, reject=[]):
|
|
if path is None:
|
|
try:
|
|
path = os.environ['PATH']
|
|
except KeyError:
|
|
return None
|
|
if is_String(path):
|
|
path = string.split(path, os.pathsep)
|
|
if pathext is None:
|
|
pathext = ['.exe', '.cmd']
|
|
for ext in pathext:
|
|
if string.lower(ext) == string.lower(file[-len(ext):]):
|
|
pathext = ['']
|
|
break
|
|
if not is_List(reject) and not is_Tuple(reject):
|
|
reject = [reject]
|
|
for dir in path:
|
|
f = os.path.join(dir, file)
|
|
for ext in pathext:
|
|
fext = f + ext
|
|
if os.path.isfile(fext):
|
|
try:
|
|
reject.index(fext)
|
|
except ValueError:
|
|
return os.path.normpath(fext)
|
|
continue
|
|
return None
|
|
|
|
else:
|
|
|
|
def WhereIs(file, path=None, pathext=None, reject=[]):
|
|
import stat
|
|
if path is None:
|
|
try:
|
|
path = os.environ['PATH']
|
|
except KeyError:
|
|
return None
|
|
if is_String(path):
|
|
path = string.split(path, os.pathsep)
|
|
if not is_List(reject) and not is_Tuple(reject):
|
|
reject = [reject]
|
|
for d in path:
|
|
f = os.path.join(d, file)
|
|
if os.path.isfile(f):
|
|
try:
|
|
st = os.stat(f)
|
|
except OSError:
|
|
# os.stat() raises OSError, not IOError if the file
|
|
# doesn't exist, so in this case we let IOError get
|
|
# raised so as to not mask possibly serious disk or
|
|
# network issues.
|
|
continue
|
|
if stat.S_IMODE(st[stat.ST_MODE]) & 0111:
|
|
try:
|
|
reject.index(f)
|
|
except ValueError:
|
|
return os.path.normpath(f)
|
|
continue
|
|
return None
|
|
|
|
def PrependPath(oldpath, newpath, sep = os.pathsep):
|
|
"""This prepends newpath elements to the given oldpath. Will only
|
|
add any particular path once (leaving the first one it encounters
|
|
and ignoring the rest, to preserve path order), and will
|
|
os.path.normpath and os.path.normcase all paths to help assure
|
|
this. This can also handle the case where the given old path
|
|
variable is a list instead of a string, in which case a list will
|
|
be returned instead of a string.
|
|
|
|
Example:
|
|
Old Path: "/foo/bar:/foo"
|
|
New Path: "/biz/boom:/foo"
|
|
Result: "/biz/boom:/foo:/foo/bar"
|
|
"""
|
|
|
|
orig = oldpath
|
|
is_list = 1
|
|
paths = orig
|
|
if not is_List(orig) and not is_Tuple(orig):
|
|
paths = string.split(paths, sep)
|
|
is_list = 0
|
|
|
|
if is_List(newpath) or is_Tuple(newpath):
|
|
newpaths = newpath
|
|
else:
|
|
newpaths = string.split(newpath, sep)
|
|
|
|
newpaths = newpaths + paths # prepend new paths
|
|
|
|
normpaths = []
|
|
paths = []
|
|
# now we add them only if they are unique
|
|
for path in newpaths:
|
|
normpath = os.path.normpath(os.path.normcase(path))
|
|
if path and not normpath in normpaths:
|
|
paths.append(path)
|
|
normpaths.append(normpath)
|
|
|
|
if is_list:
|
|
return paths
|
|
else:
|
|
return string.join(paths, sep)
|
|
|
|
def AppendPath(oldpath, newpath, sep = os.pathsep):
|
|
"""This appends new path elements to the given old path. Will
|
|
only add any particular path once (leaving the last one it
|
|
encounters and ignoring the rest, to preserve path order), and
|
|
will os.path.normpath and os.path.normcase all paths to help
|
|
assure this. This can also handle the case where the given old
|
|
path variable is a list instead of a string, in which case a list
|
|
will be returned instead of a string.
|
|
|
|
Example:
|
|
Old Path: "/foo/bar:/foo"
|
|
New Path: "/biz/boom:/foo"
|
|
Result: "/foo/bar:/biz/boom:/foo"
|
|
"""
|
|
|
|
orig = oldpath
|
|
is_list = 1
|
|
paths = orig
|
|
if not is_List(orig) and not is_Tuple(orig):
|
|
paths = string.split(paths, sep)
|
|
is_list = 0
|
|
|
|
if is_List(newpath) or is_Tuple(newpath):
|
|
newpaths = newpath
|
|
else:
|
|
newpaths = string.split(newpath, sep)
|
|
|
|
newpaths = paths + newpaths # append new paths
|
|
newpaths.reverse()
|
|
|
|
normpaths = []
|
|
paths = []
|
|
# now we add them only of they are unique
|
|
for path in newpaths:
|
|
normpath = os.path.normpath(os.path.normcase(path))
|
|
if path and not normpath in normpaths:
|
|
paths.append(path)
|
|
normpaths.append(normpath)
|
|
|
|
paths.reverse()
|
|
|
|
if is_list:
|
|
return paths
|
|
else:
|
|
return string.join(paths, sep)
|
|
|
|
if sys.platform == 'cygwin':
|
|
def get_native_path(path):
|
|
"""Transforms an absolute path into a native path for the system. In
|
|
Cygwin, this converts from a Cygwin path to a Windows one."""
|
|
return string.replace(os.popen('cygpath -w ' + path).read(), '\n', '')
|
|
else:
|
|
def get_native_path(path):
|
|
"""Transforms an absolute path into a native path for the system.
|
|
Non-Cygwin version, just leave the path alone."""
|
|
return path
|
|
|
|
display = DisplayEngine()
|
|
|
|
def Split(arg):
|
|
if is_List(arg) or is_Tuple(arg):
|
|
return arg
|
|
elif is_String(arg):
|
|
return string.split(arg)
|
|
else:
|
|
return [arg]
|
|
|
|
class CLVar(UserList):
|
|
"""A class for command-line construction variables.
|
|
|
|
This is a list that uses Split() to split an initial string along
|
|
white-space arguments, and similarly to split any strings that get
|
|
added. This allows us to Do the Right Thing with Append() and
|
|
Prepend() (as well as straight Python foo = env['VAR'] + 'arg1
|
|
arg2') regardless of whether a user adds a list or a string to a
|
|
command-line construction variable.
|
|
"""
|
|
def __init__(self, seq = []):
|
|
UserList.__init__(self, Split(seq))
|
|
def __coerce__(self, other):
|
|
return (self, CLVar(other))
|
|
def __str__(self):
|
|
return string.join(self.data)
|
|
|
|
# A dictionary that preserves the order in which items are added.
|
|
# Submitted by David Benjamin to ActiveState's Python Cookbook web site:
|
|
# http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/107747
|
|
# Including fixes/enhancements from the follow-on discussions.
|
|
class OrderedDict(UserDict):
|
|
def __init__(self, dict = None):
|
|
self._keys = []
|
|
UserDict.__init__(self, dict)
|
|
|
|
def __delitem__(self, key):
|
|
UserDict.__delitem__(self, key)
|
|
self._keys.remove(key)
|
|
|
|
def __setitem__(self, key, item):
|
|
UserDict.__setitem__(self, key, item)
|
|
if key not in self._keys: self._keys.append(key)
|
|
|
|
def clear(self):
|
|
UserDict.clear(self)
|
|
self._keys = []
|
|
|
|
def copy(self):
|
|
dict = OrderedDict()
|
|
dict.update(self)
|
|
return dict
|
|
|
|
def items(self):
|
|
return zip(self._keys, self.values())
|
|
|
|
def keys(self):
|
|
return self._keys[:]
|
|
|
|
def popitem(self):
|
|
try:
|
|
key = self._keys[-1]
|
|
except IndexError:
|
|
raise KeyError('dictionary is empty')
|
|
|
|
val = self[key]
|
|
del self[key]
|
|
|
|
return (key, val)
|
|
|
|
def setdefault(self, key, failobj = None):
|
|
UserDict.setdefault(self, key, failobj)
|
|
if key not in self._keys: self._keys.append(key)
|
|
|
|
def update(self, dict):
|
|
for (key, val) in dict.items():
|
|
self.__setitem__(key, val)
|
|
|
|
def values(self):
|
|
return map(self.get, self._keys)
|
|
|
|
class Selector(OrderedDict):
|
|
"""A callable ordered dictionary that maps file suffixes to
|
|
dictionary values. We preserve the order in which items are added
|
|
so that get_suffix() calls always return the first suffix added."""
|
|
def __call__(self, env, source):
|
|
try:
|
|
ext = splitext(str(source[0]))[1]
|
|
except IndexError:
|
|
ext = ""
|
|
try:
|
|
return self[ext]
|
|
except KeyError:
|
|
# Try to perform Environment substitution on the keys of
|
|
# the dictionary before giving up.
|
|
s_dict = {}
|
|
for (k,v) in self.items():
|
|
if not k is None:
|
|
s_k = env.subst(k)
|
|
if s_dict.has_key(s_k):
|
|
# We only raise an error when variables point
|
|
# to the same suffix. If one suffix is literal
|
|
# and a variable suffix contains this literal,
|
|
# the literal wins and we don't raise an error.
|
|
raise KeyError, (s_dict[s_k][0], k, s_k)
|
|
s_dict[s_k] = (k,v)
|
|
try:
|
|
return s_dict[ext][1]
|
|
except KeyError:
|
|
try:
|
|
return self[None]
|
|
except KeyError:
|
|
return None
|
|
|
|
|
|
if sys.platform == 'cygwin':
|
|
# On Cygwin, os.path.normcase() lies, so just report back the
|
|
# fact that the underlying Windows OS is case-insensitive.
|
|
def case_sensitive_suffixes(s1, s2):
|
|
return 0
|
|
else:
|
|
def case_sensitive_suffixes(s1, s2):
|
|
return (os.path.normcase(s1) != os.path.normcase(s2))
|
|
|
|
def adjustixes(fname, pre, suf):
|
|
if pre:
|
|
path, fn = os.path.split(os.path.normpath(fname))
|
|
if fn[:len(pre)] != pre:
|
|
fname = os.path.join(path, pre + fn)
|
|
# Only append a suffix if the file does not have one.
|
|
if suf and not splitext(fname)[1] and fname[-len(suf):] != suf:
|
|
fname = fname + suf
|
|
return fname
|
|
|
|
|
|
def unique(s):
|
|
"""Return a list of the elements in s, but without duplicates.
|
|
|
|
For example, unique([1,2,3,1,2,3]) is some permutation of [1,2,3],
|
|
unique("abcabc") some permutation of ["a", "b", "c"], and
|
|
unique(([1, 2], [2, 3], [1, 2])) some permutation of
|
|
[[2, 3], [1, 2]].
|
|
|
|
For best speed, all sequence elements should be hashable. Then
|
|
unique() will usually work in linear time.
|
|
|
|
If not possible, the sequence elements should enjoy a total
|
|
ordering, and if list(s).sort() doesn't raise TypeError it's
|
|
assumed that they do enjoy a total ordering. Then unique() will
|
|
usually work in O(N*log2(N)) time.
|
|
|
|
If that's not possible either, the sequence elements must support
|
|
equality-testing. Then unique() will usually work in quadratic
|
|
time.
|
|
"""
|
|
|
|
n = len(s)
|
|
if n == 0:
|
|
return []
|
|
|
|
# Try using a dict first, as that's the fastest and will usually
|
|
# work. If it doesn't work, it will usually fail quickly, so it
|
|
# usually doesn't cost much to *try* it. It requires that all the
|
|
# sequence elements be hashable, and support equality comparison.
|
|
u = {}
|
|
try:
|
|
for x in s:
|
|
u[x] = 1
|
|
except TypeError:
|
|
pass # move on to the next method
|
|
else:
|
|
return u.keys()
|
|
del u
|
|
|
|
# We can't hash all the elements. Second fastest is to sort,
|
|
# which brings the equal elements together; then duplicates are
|
|
# easy to weed out in a single pass.
|
|
# NOTE: Python's list.sort() was designed to be efficient in the
|
|
# presence of many duplicate elements. This isn't true of all
|
|
# sort functions in all languages or libraries, so this approach
|
|
# is more effective in Python than it may be elsewhere.
|
|
try:
|
|
t = list(s)
|
|
t.sort()
|
|
except TypeError:
|
|
pass # move on to the next method
|
|
else:
|
|
assert n > 0
|
|
last = t[0]
|
|
lasti = i = 1
|
|
while i < n:
|
|
if t[i] != last:
|
|
t[lasti] = last = t[i]
|
|
lasti = lasti + 1
|
|
i = i + 1
|
|
return t[:lasti]
|
|
del t
|
|
|
|
# Brute force is all that's left.
|
|
u = []
|
|
for x in s:
|
|
if x not in u:
|
|
u.append(x)
|
|
return u
|
|
|
|
# Much of the logic here was originally based on recipe 4.9 from the
|
|
# Python CookBook, but we had to dumb it way down for Python 1.5.2.
|
|
class LogicalLines:
|
|
|
|
def __init__(self, fileobj):
|
|
self.fileobj = fileobj
|
|
|
|
def readline(self):
|
|
result = []
|
|
while 1:
|
|
line = self.fileobj.readline()
|
|
if not line:
|
|
break
|
|
if line[-2:] == '\\\n':
|
|
result.append(line[:-2])
|
|
else:
|
|
result.append(line)
|
|
break
|
|
return string.join(result, '')
|
|
|
|
def readlines(self):
|
|
result = []
|
|
while 1:
|
|
line = self.readline()
|
|
if not line:
|
|
break
|
|
result.append(line)
|
|
return result
|
|
|
|
class Unbuffered:
|
|
"""
|
|
A proxy class that wraps a file object, flushing after every write,
|
|
and delegating everything else to the wrapped object.
|
|
"""
|
|
def __init__(self, file):
|
|
self.file = file
|
|
def write(self, arg):
|
|
self.file.write(arg)
|
|
self.file.flush()
|
|
def __getattr__(self, attr):
|
|
return getattr(self.file, attr)
|