3 Create all the contents files
5 @contact: Debian FTPMaster <ftpmaster@debian.org>
6 @copyright: 2008, 2009 Michael Casadevall <mcasadevall@debian.org>
7 @copyright: 2009 Mike O'Connor <stew@debian.org>
8 @license: GNU General Public License version 2 or later
11 ################################################################################
13 # This program is free software; you can redistribute it and/or modify
14 # it under the terms of the GNU General Public License as published by
15 # the Free Software Foundation; either version 2 of the License, or
16 # (at your option) any later version.
18 # This program is distributed in the hope that it will be useful,
19 # but WITHOUT ANY WARRANTY; without even the implied warranty of
20 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
21 # GNU General Public License for more details.
23 # You should have received a copy of the GNU General Public License
24 # along with this program; if not, write to the Free Software
25 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
27 ################################################################################
29 # <Ganneff> there is the idea to slowly replace contents files
30 # <Ganneff> with a new generation of such files.
31 # <Ganneff> having more info.
33 # <Ganneff> of course that wont help for now where we need to generate them :)
35 ################################################################################
46 from daklib import utils
47 from daklib.binary import Binary
48 from daklib.config import Config
49 from daklib.dbconn import *
51 ################################################################################
53 def usage (exit_code=0):
54 print """Usage: dak contents [options] command [arguments]
58 generate Contents-$arch.gz files
61 scan the debs in the existing pool and load contents into the bin_contents table
64 copy data from the bin_contents table into the deb_contents / udeb_contents tables
67 remove files/paths which are no longer referenced by a binary
71 show this help and exit
74 show verbose information messages
77 supress all output but errors
79 -s, --suite={stable,testing,unstable,...}
80 only operate on a single suite
84 ################################################################################
86 # where in dak.conf all of our configuration will be stowed
88 options_prefix = "Contents"
89 options_prefix = "%s::Options" % options_prefix
91 log = logging.getLogger()
93 ################################################################################
96 class EndOfContents(object):
98 A sentry object for the end of the filename stream
102 class OneAtATime(object):
104 a one space queue which sits between multiple possible producers
105 and multiple possible consumers
108 self.next_in_line = None
109 self.read_lock = threading.Condition()
110 self.write_lock = threading.Condition()
113 def enqueue(self, next):
114 self.write_lock.acquire()
115 while self.next_in_line:
118 self.write_lock.wait()
120 assert( not self.next_in_line )
121 self.next_in_line = next
122 self.write_lock.release()
123 self.read_lock.acquire()
124 self.read_lock.notify()
125 self.read_lock.release()
128 self.read_lock.acquire()
129 while not self.next_in_line:
132 self.read_lock.wait()
134 result = self.next_in_line
136 self.next_in_line = None
137 self.read_lock.release()
138 self.write_lock.acquire()
139 self.write_lock.notify()
140 self.write_lock.release()
145 class ContentsWorkThread(threading.Thread):
148 def __init__(self, upstream, downstream):
149 threading.Thread.__init__(self)
150 self.upstream = upstream
151 self.downstream = downstream
156 contents_file = self.upstream.dequeue()
157 if isinstance(contents_file,EndOfContents):
159 self.downstream.enqueue(contents_file)
162 s = datetime.datetime.now()
163 print("%s start: %s" % (self,contents_file) )
164 self._run(contents_file)
165 print("%s finished: %s in %d seconds" % (self, contents_file, (datetime.datetime.now()-s).seconds ))
167 self.downstream.enqueue(contents_file)
169 traceback.print_exc()
171 class QueryThread(ContentsWorkThread):
172 def __init__(self, upstream, downstream):
173 ContentsWorkThread.__init__(self, upstream, downstream)
179 def _run(self, contents_file):
180 contents_file.query()
182 class IngestThread(ContentsWorkThread):
183 def __init__(self, upstream, downstream):
184 ContentsWorkThread.__init__(self, upstream, downstream)
187 return "IngestThread"
190 def _run(self, contents_file):
191 contents_file.ingest()
193 class SortThread(ContentsWorkThread):
194 def __init__(self, upstream, downstream):
195 ContentsWorkThread.__init__(self, upstream, downstream)
201 def _run(self, contents_file):
202 contents_file.sorted_keys = sorted(contents_file.filenames.keys())
204 class OutputThread(ContentsWorkThread):
205 def __init__(self, upstream, downstream):
206 ContentsWorkThread.__init__(self, upstream, downstream)
209 return "OutputThread"
212 def _run(self, contents_file):
213 contents_file.open_file()
214 for fname in contents_file.sorted_keys:
215 contents_file.filehandle.write("%s\t%s\n" % (fname,contents_file.filenames[fname]))
216 contents_file.sorted_keys = None
217 contents_file.filenames.clear()
219 class GzipThread(ContentsWorkThread):
220 def __init__(self, upstream, downstream):
221 ContentsWorkThread.__init__(self, upstream, downstream)
227 def _run(self, contents_file):
228 os.system("gzip -f %s" % contents_file.filename)
230 class ContentFile(object):
236 self.filename = filename
238 self.sorted_keys = None
239 self.suite_str = suite_str
240 self.suite_id = suite_id
242 self.filehandle = None
251 self.filenames = None
252 self.sortedkeys = None
253 self.filehandle.close()
258 r = self.results.fetchone()
261 filename, package = r
262 self.filenames[filename]=package
268 # r = self.results.fetchone()
271 # filename, package = r
272 # if self.filenames.has_key(filename):
273 # self.filenames[filename] += ",%s" % (package)
275 # self.filenames[filename] = "%s" % (package)
276 # self.session.close()
280 opens a gzip stream to the contents file
282 # filepath = Config()["Contents::Root"] + self.filename
283 self.filename = "/home/stew/contents/" + self.filename
284 filedir = os.path.dirname(self.filename)
285 if not os.path.isdir(filedir):
287 # self.filehandle = gzip.open(self.filename, "w")
288 self.filehandle = open(self.filename, "w")
291 def _write_header(self):
293 self.filehandle.write(ContentFile.header)
298 def _get_header(self):
300 Internal method to return the header for Contents.gz files
302 This is boilerplate which explains the contents of the file and how
305 if not ContentFile.header:
306 if Config().has_key("Contents::Header"):
308 h = open(os.path.join( Config()["Dir::Templates"],
309 Config()["Contents::Header"] ), "r")
310 ContentFile.header = h.read()
313 log.error( "error opening header file: %d\n%s" % (Config()["Contents::Header"],
314 traceback.format_exc() ))
315 ContentFile.header = None
317 ContentFile.header = None
319 return ContentFile.header
322 class DebContentFile(ContentFile):
329 ContentFile.__init__(self,
333 self.arch_str = arch_str
334 self.arch_id = arch_id
337 self.session = DBConn().session();
339 self.results = self.session.execute("""SELECT filename, comma_separated_list(section || '/' || package)
341 WHERE ( arch=2 or arch = :arch) AND suite = :suite
342 """, { 'arch':self.arch_id, 'suite':self.suite_id } )
344 class UdebContentFile(ContentFile):
351 ContentFile.__init__(self,
357 self.session = DBConn().session();
359 self.results = self.session.execute("""SELECT filename, comma_separated_list(section || '/' || package)
363 """ , { 'suite': self.suite_id } )
365 class Contents(object):
367 Class capable of generating Contents-$arch.gz files
372 def reject(self, message):
373 log.error("E: %s" % message)
377 remove files/paths from the DB which are no longer referenced
378 by binaries and clean the temporary table
380 s = DBConn().session()
382 # clear out all of the temporarily stored content associations
383 # this should be run only after p-a has run. after a p-a
384 # run we should have either accepted or rejected every package
385 # so there should no longer be anything in the queue
386 s.query(PendingContentAssociation).delete()
388 # delete any filenames we are storing which have no binary associated
390 cafq = s.query(ContentAssociation.filename_id).distinct()
391 cfq = s.query(ContentFilename)
392 cfq = cfq.filter(~ContentFilename.cafilename_id.in_(cafq))
395 # delete any paths we are storing which have no binary associated with
397 capq = s.query(ContentAssociation.filepath_id).distinct()
398 cpq = s.query(ContentFilepath)
399 cpq = cpq.filter(~ContentFilepath.cafilepath_id.in_(capq))
405 def bootstrap_bin(self):
407 scan the existing debs in the pool to populate the bin_contents table
409 pooldir = Config()[ 'Dir::Pool' ]
411 s = DBConn().session()
413 for binary in s.query(DBBinary).yield_per(100):
414 print( "binary: %s" % binary.package )
415 filename = binary.poolfile.filename
416 # Check for existing contents
417 existingq = s.execute( "select 1 from bin_contents where binary_id=:id", {'id':binary.binary_id} );
418 if existingq.fetchone():
419 log.debug( "already imported: %s" % (filename))
421 # We don't have existing contents so import them
422 log.debug( "scanning: %s" % (filename) )
424 debfile = os.path.join(pooldir, filename)
425 if os.path.exists(debfile):
426 Binary(debfile, self.reject).scan_package(binary.binary_id, True)
428 log.error("missing .deb: %s" % filename)
434 scan the existing debs in the pool to populate the contents database tables
436 s = DBConn().session()
439 # get a mapping of all the override types we care about (right now .deb an .udeb)
440 override_type_map = {};
441 for override_type in s.query(OverrideType).all():
442 if override_type.overridetype.endswith('deb' ):
443 override_type_map[override_type.overridetype_id] = override_type.overridetype;
445 for override in s.query(Override).yield_per(100):
446 if not override_type_map.has_key(override.overridetype_id):
447 #this isn't an override we care about
450 binaries = s.execute("""SELECT b.id, b.architecture
452 JOIN bin_associations ba ON ba.bin=b.id
453 WHERE ba.suite=:suite
454 AND b.package=:package""", {'suite':override.suite_id, 'package':override.package})
456 binary = binaries.fetchone()
460 exists = s.execute("SELECT 1 FROM %s_contents WHERE binary_id=:id limit 1" % override_type_map[override.overridetype_id], {'id':binary.id})
463 if exists.fetchone():
469 s.execute( """INSERT INTO %s_contents (filename,section,package,binary_id,arch,suite)
470 SELECT file, :section, :package, :binary_id, :arch, :suite
472 WHERE binary_id=:binary_id;""" % override_type_map[override.overridetype_id],
473 { 'section' : override.section_id,
474 'package' : override.package,
475 'binary_id' : binary.id,
476 'arch' : binary.architecture,
477 'suite' : override.suite_id } )
482 Generate contents files for both deb and udeb
485 # self.udeb_generate()
487 def deb_generate(self):
489 Generate Contents-$arch.gz files for every available arch in each given suite.
491 session = DBConn().session()
492 debtype_id = get_override_type("deb", session)
493 suites = self._suites()
495 inputtoquery = OneAtATime()
496 querytoingest = OneAtATime()
497 ingesttosort = OneAtATime()
498 sorttooutput = OneAtATime()
499 outputtogzip = OneAtATime()
501 qt = QueryThread(inputtoquery,querytoingest)
502 it = IngestThread(querytoingest,ingesttosort)
503 # these actually make things worse
504 # it2 = IngestThread(querytoingest,ingesttosort)
505 # it3 = IngestThread(querytoingest,ingesttosort)
506 # it4 = IngestThread(querytoingest,ingesttosort)
507 st = SortThread(ingesttosort,sorttooutput)
508 ot = OutputThread(sorttooutput,outputtogzip)
509 gt = GzipThread(outputtogzip, None)
520 # Get our suites, and the architectures
521 for suite in [i.lower() for i in suites]:
522 suite_id = get_suite(suite, session).suite_id
523 print( "got suite_id: %s for suite: %s" % (suite_id, suite ) )
524 arch_list = self._arches(suite_id, session)
526 for (arch_id,arch_str) in arch_list:
527 print( "suite: %s, arch: %s time: %s" %(suite_id, arch_id, datetime.datetime.now().isoformat()) )
529 # filename = "dists/%s/Contents-%s.gz" % (suite, arch_str)
530 filename = "dists/%s/Contents-%s" % (suite, arch_str)
531 cf = DebContentFile(filename, suite, suite_id, arch_str, arch_id)
532 inputtoquery.enqueue( cf )
534 inputtoquery.enqueue( EndOfContents() )
537 def udeb_generate(self):
539 Generate Contents-$arch.gz files for every available arch in each given suite.
541 session = DBConn().session()
542 udebtype_id=DBConn().get_override_type_id("udeb")
543 suites = self._suites()
545 inputtoquery = OneAtATime()
546 querytoingest = OneAtATime()
547 ingesttosort = OneAtATime()
548 sorttooutput = OneAtATime()
549 outputtogzip = OneAtATime()
551 qt = QueryThread(inputtoquery,querytoingest)
552 it = IngestThread(querytoingest,ingesttosort)
553 # these actually make things worse
554 # it2 = IngestThread(querytoingest,ingesttosort)
555 # it3 = IngestThread(querytoingest,ingesttosort)
556 # it4 = IngestThread(querytoingest,ingesttosort)
557 st = SortThread(ingesttosort,sorttooutput)
558 ot = OutputThread(sorttooutput,outputtogzip)
559 gt = GzipThread(outputtogzip, None)
570 # for section, fn_pattern in [("debian-installer","dists/%s/Contents-udeb-%s"),
571 # ("non-free/debian-installer", "dists/%s/Contents-udeb-nf-%s")]:
573 # section_id = DBConn().get_section_id(section) # all udebs should be here)
574 # if section_id != -1:
578 # # Get our suites, and the architectures
579 # for suite in [i.lower() for i in suites]:
580 # suite_id = DBConn().get_suite_id(suite)
581 # arch_list = self._arches(suite_id, session)
583 # for arch_id in arch_list:
585 # writer = GzippedContentWriter(fn_pattern % (suite, arch_id[1]))
588 # results = session.execute("EXECUTE udeb_contents_q(%d,%d,%d)" % (suite_id, udebtype_id, section_id, arch_id))
591 # r = cursor.fetchone()
595 # filename, section, package, arch = r
596 # writer.write(filename, section, package)
605 Generate Contents-$arch.gz files for every available arch in each given suite.
607 session = DBConn().session()
609 arch_all_id = get_architecture("all", session).arch_id
611 # The MORE fun part. Ok, udebs need their own contents files, udeb, and udeb-nf (not-free)
612 # This is HORRIBLY debian specific :-/
613 for dtype, section, fn_pattern in \
614 [('deb', None, "dists/%s/Contents-%s.gz"),
615 ('udeb', "debian-installer", "dists/%s/Contents-udeb-%s.gz"),
616 ('udeb', "non-free/debian-installer", "dists/%s/Contents-udeb-nf-%s.gz")]:
618 overridetype = get_override_type(dtype, session)
620 # For udebs, we only look in certain sections (see the for loop above)
621 if section is not None:
622 section = get_section(section, session)
625 for suite in which_suites(session):
626 # Which architectures do we need to work on
627 arch_list = get_suite_architectures(suite.suite_name, skipsrc=True, skipall=True, session=session)
629 # Set up our file writer dictionary
632 # One file writer per arch
633 for arch in arch_list:
634 file_writers[arch.arch_id] = GzippedContentWriter(fn_pattern % (suite, arch.arch_string))
636 for r in get_suite_contents(suite, overridetype, section, session=session).fetchall():
637 filename, section, package, arch_id = r
639 if arch_id == arch_all_id:
640 # It's arch all, so all contents files get it
641 for writer in file_writers.values():
642 writer.write(filename, section, package)
644 if file_writers.has_key(arch_id):
645 file_writers[arch_id].write(filename, section, package)
648 # close all the files
649 for writer in file_writers.values():
653 return a list of suites to operate on
655 if Config().has_key( "%s::%s" %(options_prefix,"Suite")):
656 suites = utils.split_args(Config()[ "%s::%s" %(options_prefix,"Suite")])
658 suites = [ 'unstable', 'testing' ]
659 # suites = Config().SubTree("Suite").List()
663 def _arches(self, suite, session):
665 return a list of archs to operate on
668 arches = session.execute(
669 """SELECT s.architecture, a.arch_string
670 FROM suite_architectures s
671 JOIN architecture a ON (s.architecture=a.id)
672 WHERE suite = :suite_id""",
673 {'suite_id':suite } )
676 r = arches.fetchone()
680 if r[1] != "source" and r[1] != "all":
681 arch_list.append((r[0], r[1]))
686 ################################################################################
691 arguments = [('h',"help", "%s::%s" % (options_prefix,"Help")),
692 ('s',"suite", "%s::%s" % (options_prefix,"Suite"),"HasArg"),
693 ('q',"quiet", "%s::%s" % (options_prefix,"Quiet")),
694 ('v',"verbose", "%s::%s" % (options_prefix,"Verbose")),
697 commands = {'generate' : Contents.deb_generate,
698 'bootstrap_bin' : Contents.bootstrap_bin,
699 'bootstrap' : Contents.bootstrap,
700 'cruft' : Contents.cruft,
703 args = apt_pkg.ParseCommandLine(cnf.Cnf, arguments,sys.argv)
705 if (len(args) < 1) or not commands.has_key(args[0]):
708 if cnf.has_key("%s::%s" % (options_prefix,"Help")):
712 if cnf.has_key("%s::%s" % (options_prefix,"Quiet")):
715 elif cnf.has_key("%s::%s" % (options_prefix,"Verbose")):
719 logging.basicConfig( level=level,
720 format='%(asctime)s %(levelname)s %(message)s',
721 stream = sys.stderr )
723 commands[args[0]](Contents())
725 def which_suites(session):
727 return a list of suites to operate on
729 if Config().has_key( "%s::%s" %(options_prefix,"Suite")):
730 suites = utils.split_args(Config()[ "%s::%s" %(options_prefix,"Suite")])
732 suites = Config().SubTree("Suite").List()
734 return [get_suite(s.lower(), session) for s in suites]
737 if __name__ == '__main__':