/usr/share/pyshared/genshi/template/interpolation.py is in python-genshi 0.6-2.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 | # -*- coding: utf-8 -*-
#
# Copyright (C) 2007-2009 Edgewall Software
# All rights reserved.
#
# This software is licensed as described in the file COPYING, which
# you should have received as part of this distribution. The terms
# are also available at http://genshi.edgewall.org/wiki/License.
#
# This software consists of voluntary contributions made by many
# individuals. For the exact contribution history, see the revision
# history and logs, available at http://genshi.edgewall.org/log/.
"""String interpolation routines, i.e. the splitting up a given text into some
parts that are literal strings, and others that are Python expressions.
"""
from itertools import chain
import os
import re
from tokenize import PseudoToken
from genshi.core import TEXT
from genshi.template.base import TemplateSyntaxError, EXPR
from genshi.template.eval import Expression
__all__ = ['interpolate']
__docformat__ = 'restructuredtext en'
NAMESTART = 'abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ_'
NAMECHARS = NAMESTART + '.0123456789'
PREFIX = '$'
token_re = re.compile('%s|%s(?s)' % (
r'[uU]?[rR]?("""|\'\'\')((?<!\\)\\\1|.)*?\1',
PseudoToken
))
def interpolate(text, filepath=None, lineno=-1, offset=0, lookup='strict'):
"""Parse the given string and extract expressions.
This function is a generator that yields `TEXT` events for literal strings,
and `EXPR` events for expressions, depending on the results of parsing the
string.
>>> for kind, data, pos in interpolate("hey ${foo}bar"):
... print('%s %r' % (kind, data))
TEXT 'hey '
EXPR Expression('foo')
TEXT 'bar'
:param text: the text to parse
:param filepath: absolute path to the file in which the text was found
(optional)
:param lineno: the line number at which the text was found (optional)
:param offset: the column number at which the text starts in the source
(optional)
:param lookup: the variable lookup mechanism; either "lenient" (the
default), "strict", or a custom lookup class
:return: a list of `TEXT` and `EXPR` events
:raise TemplateSyntaxError: when a syntax error in an expression is
encountered
"""
pos = [filepath, lineno, offset]
textbuf = []
textpos = None
for is_expr, chunk in chain(lex(text, pos, filepath), [(True, '')]):
if is_expr:
if textbuf:
yield TEXT, ''.join(textbuf), textpos
del textbuf[:]
textpos = None
if chunk:
try:
expr = Expression(chunk.strip(), pos[0], pos[1],
lookup=lookup)
yield EXPR, expr, tuple(pos)
except SyntaxError, err:
raise TemplateSyntaxError(err, filepath, pos[1],
pos[2] + (err.offset or 0))
else:
textbuf.append(chunk)
if textpos is None:
textpos = tuple(pos)
if '\n' in chunk:
lines = chunk.splitlines()
pos[1] += len(lines) - 1
pos[2] += len(lines[-1])
else:
pos[2] += len(chunk)
def lex(text, textpos, filepath):
offset = pos = 0
end = len(text)
escaped = False
while 1:
if escaped:
offset = text.find(PREFIX, offset + 2)
escaped = False
else:
offset = text.find(PREFIX, pos)
if offset < 0 or offset == end - 1:
break
next = text[offset + 1]
if next == '{':
if offset > pos:
yield False, text[pos:offset]
pos = offset + 2
level = 1
while level:
match = token_re.match(text, pos)
if match is None:
raise TemplateSyntaxError('invalid syntax', filepath,
*textpos[1:])
pos = match.end()
tstart, tend = match.regs[3]
token = text[tstart:tend]
if token == '{':
level += 1
elif token == '}':
level -= 1
yield True, text[offset + 2:pos - 1]
elif next in NAMESTART:
if offset > pos:
yield False, text[pos:offset]
pos = offset
pos += 1
while pos < end:
char = text[pos]
if char not in NAMECHARS:
break
pos += 1
yield True, text[offset + 1:pos].strip()
elif not escaped and next == PREFIX:
if offset > pos:
yield False, text[pos:offset]
escaped = True
pos = offset + 1
else:
yield False, text[pos:offset + 1]
pos = offset + 1
if pos < end:
yield False, text[pos:]
|