3 Helper code for contents generation.
5 @contact: Debian FTPMaster <ftpmaster@debian.org>
6 @copyright: 2011 Torsten Werner <twerner@debian.org>
7 @license: GNU General Public License version 2 or later
10 ################################################################################
12 # This program is free software; you can redistribute it and/or modify
13 # it under the terms of the GNU General Public License as published by
14 # the Free Software Foundation; either version 2 of the License, or
15 # (at your option) any later version.
17 # This program is distributed in the hope that it will be useful,
18 # but WITHOUT ANY WARRANTY; without even the implied warranty of
19 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
20 # GNU General Public License for more details.
22 # You should have received a copy of the GNU General Public License
23 # along with this program; if not, write to the Free Software
24 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
26 ################################################################################
28 from daklib.dbconn import *
29 from daklib.config import Config
30 from daklib.threadpool import ThreadPool
31 from multiprocessing import Pool
33 from sqlalchemy import desc, or_
34 from sqlalchemy.exc import IntegrityError
35 from subprocess import Popen, PIPE
39 class ContentsWriter(object):
41 ContentsWriter writes the Contents-$arch.gz files.
43 def __init__(self, suite, architecture, overridetype, component = None):
45 The constructor clones its arguments into a new session object to make
46 sure that the new ContentsWriter object can be executed in a different
50 self.architecture = architecture
51 self.overridetype = overridetype
52 self.component = component
53 self.session = suite.session()
57 Returns a query object that is doing most of the work.
59 overridesuite = self.suite
60 if self.suite.overridesuite is not None:
61 overridesuite = get_suite(self.suite.overridesuite, self.session)
63 'suite': self.suite.suite_id,
64 'overridesuite': overridesuite.suite_id,
65 'arch_all': get_architecture('all', self.session).arch_id,
66 'arch': self.architecture.arch_id,
67 'type_id': self.overridetype.overridetype_id,
68 'type': self.overridetype.overridetype,
71 if self.component is not None:
72 params['component'] = self.component.component_id
74 create temp table newest_binaries (
75 id integer primary key,
78 create index newest_binaries_by_package on newest_binaries (package);
80 insert into newest_binaries (id, package)
81 select distinct on (package) id, package from binaries
82 where type = :type and
83 (architecture = :arch_all or architecture = :arch) and
84 id in (select bin from bin_associations where suite = :suite)
85 order by package, version desc;
90 (select o.package, s.section
91 from override o, section s
92 where o.suite = :overridesuite and o.type = :type_id and o.section = s.id and
93 o.component = :component)
95 select bc.file, string_agg(o.section || '/' || b.package, ',' order by b.package) as pkglist
96 from newest_binaries b, bin_contents bc, unique_override o
97 where b.id = bc.binary_id and o.package = b.package
102 create temp table newest_binaries (
103 id integer primary key,
106 create index newest_binaries_by_package on newest_binaries (package);
108 insert into newest_binaries (id, package)
109 select distinct on (package) id, package from binaries
110 where type = :type and
111 (architecture = :arch_all or architecture = :arch) and
112 id in (select bin from bin_associations where suite = :suite)
113 order by package, version desc;
118 (select distinct on (o.package, s.section) o.package, s.section
119 from override o, section s
120 where o.suite = :overridesuite and o.type = :type_id and o.section = s.id
121 order by o.package, s.section, o.modified desc)
123 select bc.file, string_agg(o.section || '/' || b.package, ',' order by b.package) as pkglist
124 from newest_binaries b, bin_contents bc, unique_override o
125 where b.id = bc.binary_id and o.package = b.package
128 return self.session.query("file", "pkglist").from_statement(sql). \
131 def formatline(self, filename, package_list):
133 Returns a formatted string for the filename argument.
135 return "%-55s %s\n" % (filename, package_list)
139 Yields a new line of the Contents-$arch.gz file in filename order.
141 for filename, package_list in self.query().yield_per(100):
142 yield self.formatline(filename, package_list)
143 # end transaction to return connection to pool
144 self.session.rollback()
148 Returns a list of lines for the Contents-$arch.gz file.
150 return [item for item in self.fetch()]
152 def output_filename(self):
154 Returns the name of the output file.
157 'root': Config()['Dir::Root'],
158 'suite': self.suite.suite_name,
159 'architecture': self.architecture.arch_string
161 if self.component is None:
162 return "%(root)s/dists/%(suite)s/Contents-%(architecture)s.gz" % values
163 values['component'] = self.component.component_name
164 return "%(root)s/dists/%(suite)s/%(component)s/Contents-%(architecture)s.gz" % values
166 def get_header(self):
168 Returns the header for the Contents files as a string.
172 filename = os.path.join(Config()['Dir::Templates'], 'contents')
173 header_file = open(filename)
174 return header_file.read()
179 def write_file(self):
181 Write the output file.
183 command = ['gzip', '--rsyncable']
184 final_filename = self.output_filename()
185 temp_filename = final_filename + '.new'
186 output_file = open(temp_filename, 'w')
187 gzip = Popen(command, stdin = PIPE, stdout = output_file)
188 gzip.stdin.write(self.get_header())
189 for item in self.fetch():
190 gzip.stdin.write(item)
195 os.remove(final_filename)
198 os.rename(temp_filename, final_filename)
199 os.chmod(final_filename, 0664)
202 def write_all(class_, suite_names = [], force = False):
204 Writes all Contents files for suites in list suite_names which defaults
205 to all 'touchable' suites if not specified explicitely. Untouchable
206 suites will be included if the force argument is set to True.
208 session = DBConn().session()
209 suite_query = session.query(Suite)
210 if len(suite_names) > 0:
211 suite_query = suite_query.filter(Suite.suite_name.in_(suite_names))
213 suite_query = suite_query.filter_by(untouchable = False)
214 deb_id = get_override_type('deb', session).overridetype_id
215 udeb_id = get_override_type('udeb', session).overridetype_id
216 main_id = get_component('main', session).component_id
217 non_free_id = get_component('non-free', session).component_id
219 for suite in suite_query:
220 suite_id = suite.suite_id
221 for architecture in suite.get_architectures(skipsrc = True, skipall = True):
222 arch_id = architecture.arch_id
223 # handle 'deb' packages
224 pool.apply_async(generate_helper, (suite_id, arch_id, deb_id))
225 # handle 'udeb' packages for 'main' and 'non-free'
226 pool.apply_async(generate_helper, (suite_id, arch_id, udeb_id, main_id))
227 pool.apply_async(generate_helper, (suite_id, arch_id, udeb_id, non_free_id))
232 def generate_helper(suite_id, arch_id, overridetype_id, component_id = None):
234 This function is called in a new subprocess.
237 session = DBConn().session()
238 suite = Suite.get(suite_id, session)
239 architecture = Architecture.get(arch_id, session)
240 overridetype = OverrideType.get(overridetype_id, session)
241 if component_id is None:
244 component = Component.get(component_id, session)
245 contents_writer = ContentsWriter(suite, architecture, overridetype, component)
246 contents_writer.write_file()
249 class ContentsScanner(object):
251 ContentsScanner provides a threadsafe method scan() to scan the contents of
254 def __init__(self, binary):
256 The argument binary is the actual DBBinary object that should be
259 self.binary_id = binary.binary_id
261 def scan(self, dummy_arg = None):
263 This method does the actual scan and fills in the associated BinContents
264 property. It commits any changes to the database. The argument dummy_arg
265 is ignored but needed by our threadpool implementation.
267 session = DBConn().session()
268 binary = session.query(DBBinary).get(self.binary_id)
269 fileset = set(binary.scan_contents())
270 if len(fileset) == 0:
271 fileset.add('EMPTY_PACKAGE')
272 for filename in fileset:
273 binary.contents.append(BinContents(file = filename))
278 def scan_all(class_, limit = None):
280 The class method scan_all() scans all binaries using multiple threads.
281 The number of binaries to be scanned can be limited with the limit
282 argument. Returns the number of processed and remaining packages as a
285 session = DBConn().session()
286 query = session.query(DBBinary).filter(DBBinary.contents == None)
287 remaining = query.count
288 if limit is not None:
289 query = query.limit(limit)
290 processed = query.count()
291 threadpool = ThreadPool()
292 for binary in query.yield_per(100):
293 threadpool.queueTask(ContentsScanner(binary).scan)
295 remaining = remaining()
297 return { 'processed': processed, 'remaining': remaining }