xref: /linux/Documentation/sphinx/kernel_include.py (revision b4db9f840283caca0d904436f187ef56a9126eaa)
1#!/usr/bin/env python3
2# -*- coding: utf-8; mode: python -*-
3# pylint: disable=R0903, C0330, R0914, R0912, E0401
4
5u"""
6    kernel-include
7    ~~~~~~~~~~~~~~
8
9    Implementation of the ``kernel-include`` reST-directive.
10
11    :copyright:  Copyright (C) 2016  Markus Heiser
12    :license:    GPL Version 2, June 1991 see linux/COPYING for details.
13
14    The ``kernel-include`` reST-directive is a replacement for the ``include``
15    directive. The ``kernel-include`` directive expand environment variables in
16    the path name and allows to include files from arbitrary locations.
17
18    .. hint::
19
20      Including files from arbitrary locations (e.g. from ``/etc``) is a
21      security risk for builders. This is why the ``include`` directive from
22      docutils *prohibit* pathnames pointing to locations *above* the filesystem
23      tree where the reST document with the include directive is placed.
24
25    Substrings of the form $name or ${name} are replaced by the value of
26    environment variable name. Malformed variable names and references to
27    non-existing variables are left unchanged.
28"""
29
30# ==============================================================================
31# imports
32# ==============================================================================
33
34import os.path
35
36from docutils import io, nodes, statemachine
37from docutils.utils.error_reporting import SafeString, ErrorString
38from docutils.parsers.rst import directives
39from docutils.parsers.rst.directives.body import CodeBlock, NumberLines
40from docutils.parsers.rst.directives.misc import Include
41
42__version__  = '1.0'
43
44# ==============================================================================
45def setup(app):
46# ==============================================================================
47
48    app.add_directive("kernel-include", KernelInclude)
49    return dict(
50        version = __version__,
51        parallel_read_safe = True,
52        parallel_write_safe = True
53    )
54
55# ==============================================================================
56class KernelInclude(Include):
57# ==============================================================================
58
59    u"""KernelInclude (``kernel-include``) directive"""
60
61    def run(self):
62        env = self.state.document.settings.env
63        path = os.path.realpath(
64            os.path.expandvars(self.arguments[0]))
65
66        # to get a bit security back, prohibit /etc:
67        if path.startswith(os.sep + "etc"):
68            raise self.severe(
69                'Problems with "%s" directive, prohibited path: %s'
70                % (self.name, path))
71
72        self.arguments[0] = path
73
74        env.note_dependency(os.path.abspath(path))
75
76        #return super(KernelInclude, self).run() # won't work, see HINTs in _run()
77        return self._run()
78
79    def _run(self):
80        """Include a file as part of the content of this reST file."""
81
82        # HINT: I had to copy&paste the whole Include.run method. I'am not happy
83        # with this, but due to security reasons, the Include.run method does
84        # not allow absolute or relative pathnames pointing to locations *above*
85        # the filesystem tree where the reST document is placed.
86
87        if not self.state.document.settings.file_insertion_enabled:
88            raise self.warning('"%s" directive disabled.' % self.name)
89        source = self.state_machine.input_lines.source(
90            self.lineno - self.state_machine.input_offset - 1)
91        source_dir = os.path.dirname(os.path.abspath(source))
92        path = directives.path(self.arguments[0])
93        if path.startswith('<') and path.endswith('>'):
94            path = os.path.join(self.standard_include_path, path[1:-1])
95        path = os.path.normpath(os.path.join(source_dir, path))
96
97        # HINT: this is the only line I had to change / commented out:
98        #path = utils.relative_path(None, path)
99
100        path = nodes.reprunicode(path)
101        encoding = self.options.get(
102            'encoding', self.state.document.settings.input_encoding)
103        e_handler=self.state.document.settings.input_encoding_error_handler
104        tab_width = self.options.get(
105            'tab-width', self.state.document.settings.tab_width)
106        try:
107            self.state.document.settings.record_dependencies.add(path)
108            include_file = io.FileInput(source_path=path,
109                                        encoding=encoding,
110                                        error_handler=e_handler)
111        except UnicodeEncodeError as error:
112            raise self.severe('Problems with "%s" directive path:\n'
113                              'Cannot encode input file path "%s" '
114                              '(wrong locale?).' %
115                              (self.name, SafeString(path)))
116        except IOError as error:
117            raise self.severe('Problems with "%s" directive path:\n%s.' %
118                      (self.name, ErrorString(error)))
119        startline = self.options.get('start-line', None)
120        endline = self.options.get('end-line', None)
121        try:
122            if startline or (endline is not None):
123                lines = include_file.readlines()
124                rawtext = ''.join(lines[startline:endline])
125            else:
126                rawtext = include_file.read()
127        except UnicodeError as error:
128            raise self.severe('Problem with "%s" directive:\n%s' %
129                              (self.name, ErrorString(error)))
130        # start-after/end-before: no restrictions on newlines in match-text,
131        # and no restrictions on matching inside lines vs. line boundaries
132        after_text = self.options.get('start-after', None)
133        if after_text:
134            # skip content in rawtext before *and incl.* a matching text
135            after_index = rawtext.find(after_text)
136            if after_index < 0:
137                raise self.severe('Problem with "start-after" option of "%s" '
138                                  'directive:\nText not found.' % self.name)
139            rawtext = rawtext[after_index + len(after_text):]
140        before_text = self.options.get('end-before', None)
141        if before_text:
142            # skip content in rawtext after *and incl.* a matching text
143            before_index = rawtext.find(before_text)
144            if before_index < 0:
145                raise self.severe('Problem with "end-before" option of "%s" '
146                                  'directive:\nText not found.' % self.name)
147            rawtext = rawtext[:before_index]
148
149        include_lines = statemachine.string2lines(rawtext, tab_width,
150                                                  convert_whitespace=True)
151        if 'literal' in self.options:
152            # Convert tabs to spaces, if `tab_width` is positive.
153            if tab_width >= 0:
154                text = rawtext.expandtabs(tab_width)
155            else:
156                text = rawtext
157            literal_block = nodes.literal_block(rawtext, source=path,
158                                    classes=self.options.get('class', []))
159            literal_block.line = 1
160            self.add_name(literal_block)
161            if 'number-lines' in self.options:
162                try:
163                    startline = int(self.options['number-lines'] or 1)
164                except ValueError:
165                    raise self.error(':number-lines: with non-integer '
166                                     'start value')
167                endline = startline + len(include_lines)
168                if text.endswith('\n'):
169                    text = text[:-1]
170                tokens = NumberLines([([], text)], startline, endline)
171                for classes, value in tokens:
172                    if classes:
173                        literal_block += nodes.inline(value, value,
174                                                      classes=classes)
175                    else:
176                        literal_block += nodes.Text(value, value)
177            else:
178                literal_block += nodes.Text(text, text)
179            return [literal_block]
180        if 'code' in self.options:
181            self.options['source'] = path
182            codeblock = CodeBlock(self.name,
183                                  [self.options.pop('code')], # arguments
184                                  self.options,
185                                  include_lines, # content
186                                  self.lineno,
187                                  self.content_offset,
188                                  self.block_text,
189                                  self.state,
190                                  self.state_machine)
191            return codeblock.run()
192        self.state_machine.insert_input(include_lines, path)
193        return []
194