~launchpad-pqm/launchpad/devel

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
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
# Copyright 2009 Canonical Ltd.  This software is licensed under the
# GNU Affero General Public License version 3 (see the file LICENSE).

__metaclass__ = type

from datetime import datetime
import time
from urlparse import urlparse

from twisted.python import log
from twisted.web import resource, static, util, server, proxy
from twisted.internet.threads import deferToThread

from canonical.config import config
from canonical.librarian.client import url_path_quote
from lp.services.database import (
    read_transaction,
    write_transaction,
    )
from canonical.librarian.utils import guess_librarian_encoding


defaultResource = static.Data("""
        <html>
        <body>
        <h1>Launchpad Librarian</h1>
        <p>
        http://librarian.launchpad.net/ is a
        file repository used by <a href="https://launchpad.net/">Launchpad</a>.
        </p>
        <p><small>Copyright 2004-2009 Canonical Ltd.</small></p>
        <!-- kthxbye. -->
        </body></html>
        """, type='text/html')
fourOhFour = resource.NoResource('No such resource')


class NotFound(Exception):
    pass


class LibraryFileResource(resource.Resource):
    def __init__(self, storage, upstreamHost, upstreamPort):
        resource.Resource.__init__(self)
        self.storage = storage
        self.upstreamHost = upstreamHost
        self.upstreamPort = upstreamPort

    def getChild(self, name, request):
        if name == '':
            # Root resource
            return defaultResource
        try:
            aliasID = int(name)
        except ValueError:
            log.msg(
                "404: alias is not an int: %r" % (name,))
            return fourOhFour

        return LibraryFileAliasResource(self.storage, aliasID,
                self.upstreamHost, self.upstreamPort)


class LibraryFileAliasResource(resource.Resource):
    def __init__(self, storage, aliasID, upstreamHost, upstreamPort):
        resource.Resource.__init__(self)
        self.storage = storage
        self.aliasID = aliasID
        self.upstreamHost = upstreamHost
        self.upstreamPort = upstreamPort

    def getChild(self, filename, request):
        # If we still have another component of the path, then we have
        # an old URL that encodes the content ID. We want to keep supporting
        # these, so we just ignore the content id that is currently in
        # self.aliasID and extract the real one from the URL. Note that
        # tokens do not work with the old URL style: they are URL specific.
        if len(request.postpath) == 1:
            try:
                self.aliasID = int(filename)
            except ValueError:
                log.msg(
                    "404 (old URL): alias is not an int: %r" % (filename,))
                return fourOhFour
            filename = request.postpath[0]

        # IFF the request has a .restricted. subdomain, ensure there is a
        # alias id in the right most subdomain, and that it matches
        # self.aliasIDd, And that the host precisely matches what we generate
        # (specifically to stop people putting a good prefix to the left of an
        # attacking one).
        hostname = request.getRequestHostname()
        if '.restricted.' in hostname:
            # Configs can change without warning: evaluate every time.
            download_url = config.librarian.download_url
            parsed = list(urlparse(download_url))
            netloc = parsed[1]
            # Strip port if present
            if netloc.find(':') > -1:
                netloc = netloc[:netloc.find(':')]
            expected_hostname = 'i%d.restricted.%s' % (self.aliasID, netloc)
            if expected_hostname != hostname:
                log.msg(
                    '404: expected_hostname != hostname: %r != %r' %
                    (expected_hostname, hostname))
                return fourOhFour

        token = request.args.get('token', [None])[0]
        path = request.path
        deferred = deferToThread(self._getFileAlias, self.aliasID, token, path)
        deferred.addCallback(
                self._cb_getFileAlias, filename, request
                )
        deferred.addErrback(self._eb_getFileAlias)
        return util.DeferredResource(deferred)

    @write_transaction
    def _getFileAlias(self, aliasID, token, path):
        try:
            alias = self.storage.getFileAlias(aliasID, token, path)
            alias.updateLastAccessed()
            return (alias.contentID, alias.filename,
                alias.mimetype, alias.date_created, alias.restricted)
        except LookupError:
            raise NotFound

    def _eb_getFileAlias(self, failure):
        failure.trap(NotFound)
        return fourOhFour

    def _cb_getFileAlias(
            self, (dbcontentID, dbfilename, mimetype, date_created, restricted),
            filename, request
            ):
        # Return a 404 if the filename in the URL is incorrect. This offers
        # a crude form of access control (stuff we care about can have
        # unguessable names effectively using the filename as a secret).
        if dbfilename.encode('utf-8') != filename:
            log.msg(
                "404: dbfilename.encode('utf-8') != filename: %r != %r"
                % (dbfilename.encode('utf-8'), filename))
            return fourOhFour

        if not restricted:
            # Set our caching headers. Librarian files can be cached forever.
            request.setHeader('Cache-Control', 'max-age=31536000, public')
        else:
            # Restricted files require revalidation every time. For now,
            # until the deployment details are completely reviewed, the
            # simplest, most cautious approach is taken: no caching permited.
            request.setHeader('Cache-Control', 'max-age=0, private')

        if self.storage.hasFile(dbcontentID) or self.upstreamHost is None:
            # XXX: Brad Crittenden 2007-12-05 bug=174204: When encodings are
            # stored as part of a file's metadata this logic will be replaced.
            encoding, mimetype = guess_librarian_encoding(filename, mimetype)
            return File(
                mimetype, encoding, date_created,
                self.storage._fileLocation(dbcontentID))
        else:
            return proxy.ReverseProxyResource(self.upstreamHost,
                                              self.upstreamPort, request.path)

    def render_GET(self, request):
        return defaultResource.render(request)


class File(static.File):
    isLeaf = True
    def __init__(
        self, contentType, encoding, modification_time, *args, **kwargs):
        # Have to convert the UTC datetime to POSIX timestamp (localtime)
        offset = datetime.utcnow() - datetime.now()
        local_modification_time = modification_time - offset
        self._modification_time = time.mktime(
            local_modification_time.timetuple())
        static.File.__init__(self, *args, **kwargs)
        self.type = contentType
        self.encoding = encoding

    def getModificationTime(self):
        """Override the time on disk with the time from the database.

        This is used by twisted to set the Last-Modified: header.
        """
        return self._modification_time


class DigestSearchResource(resource.Resource):
    def __init__(self, storage):
        self.storage = storage
        resource.Resource.__init__(self)

    def render_GET(self, request):
        try:
            digest = request.args['digest'][0]
        except LookupError:
            return static.Data('Bad search', 'text/plain').render(request)

        deferred = deferToThread(self._matchingAliases, digest)
        deferred.addCallback(self._cb_matchingAliases, request)
        deferred.addErrback(_eb, request)
        return server.NOT_DONE_YET

    @read_transaction
    def _matchingAliases(self, digest):
        library = self.storage.library
        matches = ['%s/%s' % (aID, url_path_quote(aName))
                   for fID in library.lookupBySHA1(digest)
                   for aID, aName, aType in library.getAliases(fID)]
        return matches

    def _cb_matchingAliases(self, matches, request):
        text = '\n'.join([str(len(matches))] + matches)
        response = static.Data(text.encode('utf-8'),
                               'text/plain; charset=utf-8').render(request)
        request.write(response)
        request.finish()


# Ask robots not to index or archive anything in the librarian.
robotsTxt = static.Data("""
User-agent: *
Disallow: /
""", type='text/plain')


def _eb(failure, request):
    """Generic errback for failures during a render_GET."""
    request.processingFailed(failure)