-# find me all of the contents for a given .deb
-contents_q = """PREPARE contents_q(int,int) as
- SELECT (p.path||'/'||n.file) AS fn,
- s.section,
- b.package,
- b.architecture
- FROM content_associations c join content_file_paths p ON (c.filepath=p.id)
- JOIN content_file_names n ON (c.filename=n.id)
- JOIN binaries b ON (b.id=c.binary_pkg)
- JOIN override o ON (o.package=b.package)
- JOIN section s ON (s.id=o.section)
- WHERE o.suite = $1 AND o.type = $2
- AND b.type='deb'
- ORDER BY fn"""
-
-# find me all of the contents for a given .udeb
-udeb_contents_q = """PREPARE udeb_contents_q(int,int,int) as
- SELECT (p.path||'/'||n.file) AS fn,
- s.section,
- b.package,
- b.architecture
- FROM content_associations c join content_file_paths p ON (c.filepath=p.id)
- JOIN content_file_names n ON (c.filename=n.id)
- JOIN binaries b ON (b.id=c.binary_pkg)
- JOIN override o ON (o.package=b.package)
- JOIN section s ON (s.id=o.section)
- WHERE o.suite = $1 AND o.type = $2
- AND s.id = $3
- AND b.type='udeb'
- ORDER BY fn"""
-
-# FROM content_file_paths p join content_associations c ON (c.filepath=p.id)
-# JOIN content_file_names n ON (c.filename=n.id)
-# JOIN binaries b ON (b.id=c.binary_pkg)
-# JOIN override o ON (o.package=b.package)
-# JOIN section s ON (s.id=o.section)
-# WHERE o.suite = $1 AND o.type = $2
-# AND s.id = $3
-# AND b.id in (SELECT ba.bin from bin_associations ba join binaries b on b.id=ba.bin where (b.architecture=$3 or b.architecture=$4)and ba.suite=$1 and b.type='udeb')
-# GROUP BY fn
-# ORDER BY fn;"""
-
-class EndOfContents(object):
- """
- A sentry object for the end of the filename stream
- """
- pass
-
-class GzippedContentWriter(object):
- """
- An object which will write contents out to a Contents-$arch.gz
- file on a separate thread
- """
-
- header = None # a class object holding the header section of contents file
-
- def __init__(self, filename):
- """
- @ptype filename: string
- @param filename: the name of the file to write to
- """
- self.queue = Queue.Queue()
- self.current_file = None
- self.first_package = True
- self.output = self.open_file(filename)
- self.thread = threading.Thread(target=self.write_thread,
- name='Contents writer')
- self.thread.start()
-
- def open_file(self, filename):
- """
- opens a gzip stream to the contents file
- """
- filepath = Config()["Contents::Root"] + filename
- filedir = os.path.dirname(filepath)
- if not os.path.isdir(filedir):
- os.makedirs(filedir)
- return gzip.open(filepath, "w")
-
- def write(self, filename, section, package):
- """
- enqueue content to be written to the file on a separate thread
- """
- self.queue.put((filename,section,package))
-
- def write_thread(self):
- """
- the target of a Thread which will do the actual writing
- """
- while True:
- next = self.queue.get()
- if isinstance(next, EndOfContents):
- self.output.write('\n')
- self.output.close()
- break
-
- (filename,section,package)=next
- if next != self.current_file:
- # this is the first file, so write the header first
- if not self.current_file:
- self.output.write(self._getHeader())
-
- self.output.write('\n%s\t' % filename)
- self.first_package = True
-
- self.current_file=filename
-
- if not self.first_package:
- self.output.write(',')
- else:
- self.first_package=False
- self.output.write('%s/%s' % (section,package))
-
- def finish(self):
- """
- enqueue the sentry object so that writers will know to terminate
- """
- self.queue.put(EndOfContents())
-
- @classmethod
- def _getHeader(self):
- """
- Internal method to return the header for Contents.gz files
-
- This is boilerplate which explains the contents of the file and how
- it can be used.
- """
- if not GzippedContentWriter.header:
- if Config().has_key("Contents::Header"):
- try:
- h = open(os.path.join( Config()["Dir::Templates"],
- Config()["Contents::Header"] ), "r")
- GzippedContentWriter.header = h.read()
- h.close()
- except:
- log.error( "error opening header file: %d\n%s" % (Config()["Contents::Header"],
- traceback.format_exc() ))
- GzippedContentWriter.header = None
- else:
- GzippedContentWriter.header = None
-
- return GzippedContentWriter.header
-
-
-class Contents(object):
- """
- Class capable of generating Contents-$arch.gz files
-
- Usage GenerateContents().generateContents( ["main","contrib","non-free"] )
- """
-
- def __init__(self):
- self.header = None
-
- def reject(self, message):
- log.error("E: %s" % message)
-
- def cruft(self):
- """
- remove files/paths from the DB which are no longer referenced
- by binaries and clean the temporary table
- """
- s = DBConn().session()
-
- # clear out all of the temporarily stored content associations
- # this should be run only after p-a has run. after a p-a
- # run we should have either accepted or rejected every package
- # so there should no longer be anything in the queue
- s.query(PendingContentsAssociation).delete()
-
- # delete any filenames we are storing which have no binary associated
- # with them
- cafq = s.query(ContentAssociation.filename_id).distinct()
- cfq = s.query(ContentFilename)
- cfq = cfq.filter(~ContentFilename.cafilename_id.in_(cafq))
- cfq.delete()
-
- # delete any paths we are storing which have no binary associated with
- # them
- capq = s.query(ContentAssociation.filepath_id).distinct()
- cpq = s.query(ContentFilepath)
- cpq = cpq.filter(~ContentFilepath.cafilepath_id.in_(capq))
- cpq.delete()
-
- s.commit()
-
-
- def bootstrap(self):
- """
- scan the existing debs in the pool to populate the contents database tables
- """
- pooldir = Config()[ 'Dir::Pool' ]
-
- s = DBConn().session()
-
- for suite in s.query(Suite).all():
- for arch in get_suite_architectures(suite.suite_name, skipsrc=True, skipall=True, session=s):
- q = s.query(BinAssociation).join(Suite)
- q = q.join(Suite).filter_by(suite_name=suite.suite_name)
- q = q.join(DBBinary).join(Architecture).filter_by(arch.arch_string)
- for ba in q:
- filename = ba.binary.poolfile.filename
- # Check for existing contents
- existingq = s.query(ContentAssociations).filter_by(binary_pkg=ba.binary_id).limit(1)
- if existingq.count() > 0:
- log.debug( "already imported: %s" % (filename))
- else:
- # We don't have existing contents so import them
- log.debug( "scanning: %s" % (filename) )
- debfile = os.path.join(pooldir, filename)
- if os.path.exists(debfile):
- Binary(debfile, self.reject).scan_package(ba.binary_id, True)
- else:
- log.error("missing .deb: %s" % filename)
-
-
- def generate(self):
- """
- Generate Contents-$arch.gz files for every available arch in each given suite.
- """
- cursor = DBConn().cursor()
-
- DBConn().prepare("contents_q", contents_q)
- DBConn().prepare("udeb_contents_q", udeb_contents_q)
-
- debtype_id=DBConn().get_override_type_id("deb")
- udebtype_id=DBConn().get_override_type_id("udeb")
-
- arch_all_id = DBConn().get_architecture_id("all")
- suites = self._suites()
-
-
- # Get our suites, and the architectures
- for suite in [i.lower() for i in suites]:
- suite_id = DBConn().get_suite_id(suite)
- arch_list = self._arches(cursor, suite_id)
-
- file_writers = {}
-
- try:
- for arch_id in arch_list:
- file_writers[arch_id[0]] = GzippedContentWriter("dists/%s/Contents-%s.gz" % (suite, arch_id[1]))
-
- cursor.execute("EXECUTE contents_q(%d,%d);" % (suite_id, debtype_id))
-
- while True:
- r = cursor.fetchone()
- if not r:
- break
-
- filename, section, package, arch = r
-
- if not file_writers.has_key( arch ):
- continue