This file is indexed.

/usr/lib/python2.7/dist-packages/manuel/capture.py is in python-manuel 1.8.0-4.

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
import manuel
import re
import string
import textwrap

CAPTURE_DIRECTIVE = re.compile(
    r'^(?P<indent>(\t| )*)\.\.\s*->\s*(?P<name>\S+).*$',
    re.MULTILINE)


class Capture(object):
    def __init__(self, name, block):
        self.name = name
        self.block = block

def normalize_whitespace(s):
    return s.replace('\t', ' '*8) # turn tabs into spaces


@manuel.timing(manuel.EARLY)
def find_captures(document):
    while True:
        regions = document.find_regions(CAPTURE_DIRECTIVE)
        if not regions:
            break
        region = regions[-1]
        # note that start and end have different bases, "start" is the offset
        # from the begining of the region, "end" is a document line number
        end = region.lineno - 2

        indent = region.start_match.group('indent')
        indent = normalize_whitespace(indent)

        def indent_matches(line):
            """Is the indentation of a line match what we're looking for?"""
            line = normalize_whitespace(line)

            if not line.strip():
                # the line consists entirely of whitespace (or nothing at all),
                # so is not considered to be of the appropriate indentation
                return False

            if line.startswith(indent):
                if line[len(indent)] not in string.whitespace:
                    return True

            # if none of the above found the indentation to be a match, it is
            # not a match
            return False

        # now that we've extracted the information we need, lets slice up the
        # document's regions to match

        for candidate in document:
            if candidate.lineno >= region.lineno:
                break
            found_region = candidate

        lines = found_region.source.splitlines()
        if found_region.lineno + len(lines) < end:
            raise RuntimeError('both start and end lines must be in the '
                'same region')

        start = None
        for offset, line in reversed(list(enumerate(lines))):
            if offset > end - found_region.lineno:
                continue
            if indent_matches(line):
                break
            start = offset + 1

        if start is None:
            raise RuntimeError("couldn't find the start of the block; "
                "improper indentation of capture directive?")

        _, temp_region = document.split_region(found_region,
            found_region.lineno+start)

        # there are some extra lines in the new region, trim them off
        final_region, _ = document.split_region(temp_region, end+1)
        document.remove_region(final_region)

        name = region.start_match.group('name')
        block = textwrap.dedent(final_region.source)
        document.claim_region(region)
        region.parsed = Capture(name, block)


def store_capture(region, document, globs):
    if not isinstance(region.parsed, Capture):
        return

    globs[region.parsed.name] = region.parsed.block


class Manuel(manuel.Manuel):
    def __init__(self):
        manuel.Manuel.__init__(self, [find_captures], [store_capture])