WHERE ca.id IS NULL)""" );
cursor.execute( "COMMIT" )
+
def bootstrap(self):
"""
scan the existing debs in the pool to populate the contents database tables
"Split queue/done into a date-based hierarchy"),
("stats",
"Generate statistics"),
+ ("calculate-shasums",
+ "Calculate missing sha1sums and sha256sums"),
("bts-categorize",
"Categorize uncategorized bugs filed against ftp.debian.org"),
]
filename int4 not null references content_file_names(id) on delete cascade
);""")
+ c.execute("""CREATE TABLE temp_content_associations (
+ id serial not null,
+ package text not null,
+ version debversion not null,
+ filepath int4 not null references content_file_paths(id) on delete cascade,
+ filename int4 not null references content_file_names(id) on delete cascade
+ );""")
+
c.execute("""CREATE FUNCTION comma_concat(text, text) RETURNS text
AS $_$select case
WHEN $2 is null or $2 = '' THEN $1
suite_id = database.get_suite_id(suite)
projectB.query("INSERT INTO bin_associations (suite, bin) VALUES (%d, currval('binaries_id_seq'))" % (suite_id))
- # insert contents into the database
- contents = utils.generate_contents_information(file)
- q = projectB.query("SELECT currval('binaries_id_seq')")
- bin_id = int(q.getresult()[0][0])
- for file in contents:
- database.insert_content_path(bin_id, file)
+
+ if not database.copy_temporary_contents(package, version, files[newfile]):
+ reject("Missing contents for package")
# If the .orig.tar.gz is in a legacy directory we need to poolify
# it, so that apt-get source (and anything else that goes by the
################################################################################
-import commands, errno, fcntl, os, re, shutil, stat, sys, time, tempfile, traceback
+import commands, errno, fcntl, os, re, shutil, stat, sys, time, tempfile, traceback, tarfile
import apt_inst, apt_pkg
-from daklib import database
+from debian_bundle import deb822
+from daklib.dbconn import DBConn
+from daklib.binary import Binary
from daklib import logging
from daklib import queue
from daklib import utils
################################################################################
+def create_tmpdir():
+ """
+ Create a temporary directory that can be used for unpacking files into for
+ checking
+ """
+ tmpdir = tempfile.mkdtemp()
+ return tmpdir
+
+################################################################################
+
def copy_to_holding(filename):
global in_holding
################################################################################
-def check_deb_ar(filename):
- """
- Sanity check the ar of a .deb, i.e. that there is:
-
- 1. debian-binary
- 2. control.tar.gz
- 3. data.tar.gz or data.tar.bz2
-
- in that order, and nothing else.
- """
- cmd = "ar t %s" % (filename)
- (result, output) = commands.getstatusoutput(cmd)
- if result != 0:
- reject("%s: 'ar t' invocation failed." % (filename))
- reject(utils.prefix_multi_line_string(output, " [ar output:] "), "")
- chunks = output.split('\n')
- if len(chunks) != 3:
- reject("%s: found %d chunks, expected 3." % (filename, len(chunks)))
- if chunks[0] != "debian-binary":
- reject("%s: first chunk is '%s', expected 'debian-binary'." % (filename, chunks[0]))
- if chunks[1] != "control.tar.gz":
- reject("%s: second chunk is '%s', expected 'control.tar.gz'." % (filename, chunks[1]))
- if chunks[2] not in [ "data.tar.bz2", "data.tar.gz" ]:
- reject("%s: third chunk is '%s', expected 'data.tar.gz' or 'data.tar.bz2'." % (filename, chunks[2]))
-
-################################################################################
-
def check_files():
global reprocess
has_binaries = 0
has_source = 0
+ cursor = DBConn().cursor()
+ # Check for packages that have moved from one component to another
+ # STU: this should probably be changed to not join on architecture, suite tables but instead to used their cached name->id mappings from DBConn
+ cursor.execute("""PREPARE moved_pkg_q AS
+ SELECT c.name FROM binaries b, bin_associations ba, suite s, location l,
+ component c, architecture a, files f
+ WHERE b.package = $1 AND s.suite_name = $2
+ AND (a.arch_string = $3 OR a.arch_string = 'all')
+ AND ba.bin = b.id AND ba.suite = s.id AND b.architecture = a.id
+ AND f.location = l.id
+ AND l.component = c.id
+ AND b.file = f.id""")
+
for f in file_keys:
# Ensure the file does not already exist in one of the accepted directories
for d in [ "Accepted", "Byhand", "New", "ProposedUpdates", "OldProposedUpdates", "Embargoed", "Unembargoed" ]:
# Check the version and for file overwrites
reject(Upload.check_binary_against_db(f),"")
- check_deb_ar(f)
+ Binary(f).scan_package()
# Checks for a source package...
else:
# Validate the component
component = files[f]["component"]
- component_id = database.get_component_id(component)
+ component_id = DBConn().get_component_id(component)
if component_id == -1:
reject("file '%s' has unknown component '%s'." % (f, component))
continue
# Determine the location
location = Cnf["Dir::Pool"]
- location_id = database.get_location_id (location, component, archive)
+ location_id = DBConn().get_location_id(location, component, archive)
if location_id == -1:
reject("[INTERNAL ERROR] couldn't determine location (Component: %s, Archive: %s)" % (component, archive))
files[f]["location id"] = location_id
# Check the md5sum & size against existing files (if any)
files[f]["pool name"] = utils.poolify (changes["source"], files[f]["component"])
- files_id = database.get_files_id(files[f]["pool name"] + f, files[f]["size"], files[f]["md5sum"], files[f]["location id"])
+ files_id = DBConn().get_files_id(files[f]["pool name"] + f, files[f]["size"], files[f]["md5sum"], files[f]["location id"])
if files_id == -1:
reject("INTERNAL ERROR, get_files_id() returned multiple matches for %s." % (f))
elif files_id == -2:
files[f]["files id"] = files_id
# Check for packages that have moved from one component to another
- q = Upload.projectB.query("""
-SELECT c.name FROM binaries b, bin_associations ba, suite s, location l,
- component c, architecture a, files f
- WHERE b.package = '%s' AND s.suite_name = '%s'
- AND (a.arch_string = '%s' OR a.arch_string = 'all')
- AND ba.bin = b.id AND ba.suite = s.id AND b.architecture = a.id
- AND f.location = l.id AND l.component = c.id AND b.file = f.id"""
- % (files[f]["package"], suite,
- files[f]["architecture"]))
- ql = q.getresult()
+ files[f]['suite'] = suite
+ cursor.execute("""EXECUTE moved_pkg_q( %(package)s, %(suite)s, %(architecture)s )""", ( files[f] ) )
+ ql = cursor.fetchone()
if ql:
files[f]["othercomponents"] = ql[0][0]
or pkg.orig_tar_gz == -1:
return
- # Create a temporary directory to extract the source into
- if Options["No-Action"]:
- tmpdir = tempfile.mkdtemp()
- else:
- # We're in queue/holding and can create a random directory.
- tmpdir = "%s" % (os.getpid())
- os.mkdir(tmpdir)
+ tmpdir = create_tmpdir()
# Move into the temporary directory
cwd = os.getcwd()
################################################################################
def lookup_uid_from_fingerprint(fpr):
- q = Upload.projectB.query("SELECT u.uid, u.name, k.debian_maintainer FROM fingerprint f JOIN keyrings k ON (f.keyring=k.id), uid u WHERE f.uid = u.id AND f.fingerprint = '%s'" % (fpr))
- qs = q.getresult()
- if len(qs) == 0:
- return (None, None, None)
+ """
+ Return the uid,name,isdm for a given gpg fingerprint
+
+ @ptype fpr: string
+ @param fpr: a 40 byte GPG fingerprint
+
+ @return (uid, name, isdm)
+ """
+ cursor = DBConn().cursor()
+ cursor.execute( "SELECT u.uid, u.name, k.debian_maintainer FROM fingerprint f JOIN keyrings k ON (f.keyring=k.id), uid u WHERE f.uid = u.id AND f.fingerprint = '%s'" % (fpr))
+ qs = cursor.fetchone()
+ if qs:
+ return qs
else:
- return qs[0]
+ return (None, None, None)
def check_signed_by_key():
"""Ensure the .changes is signed by an authorized uploader."""
if not sponsored and not may_nmu:
source_ids = []
- q = Upload.projectB.query("SELECT s.id, s.version FROM source s JOIN src_associations sa ON (s.id = sa.source) WHERE s.source = '%s' AND s.dm_upload_allowed = 'yes'" % (changes["source"]))
+ cursor.execute( "SELECT s.id, s.version FROM source s JOIN src_associations sa ON (s.id = sa.source) WHERE s.source = %(source)s AND s.dm_upload_allowed = 'yes'", changes )
highest_sid, highest_version = None, None
should_reject = True
- for si in q.getresult():
+ while True:
+ si = cursor.fetchone()
+ if not si:
+ break
+
if highest_version == None or apt_pkg.VersionCompare(si[1], highest_version) == 1:
highest_sid = si[0]
highest_version = si[1]
if highest_sid == None:
reject("Source package %s does not have 'DM-Upload-Allowed: yes' in its most recent version" % changes["source"])
else:
- q = Upload.projectB.query("SELECT m.name FROM maintainer m WHERE m.id IN (SELECT su.maintainer FROM src_uploaders su JOIN source s ON (s.id = su.source) WHERE su.source = %s)" % (highest_sid))
- for m in q.getresult():
+
+ cursor.execute("SELECT m.name FROM maintainer m WHERE m.id IN (SELECT su.maintainer FROM src_uploaders su JOIN source s ON (s.id = su.source) WHERE su.source = %s)" % (highest_sid))
+
+ while True:
+ m = cursor.fetchone()
+ if not m:
+ break
+
(rfc822, rfc2047, name, email) = utils.fix_maintainer(m[0])
if email == uid_email or name == uid_name:
should_reject=False
for b in changes["binary"].keys():
for suite in changes["distribution"].keys():
- suite_id = database.get_suite_id(suite)
- q = Upload.projectB.query("SELECT DISTINCT s.source FROM source s JOIN binaries b ON (s.id = b.source) JOIN bin_associations ba On (b.id = ba.bin) WHERE b.package = '%s' AND ba.suite = %s" % (b, suite_id))
- for s in q.getresult():
+ suite_id = DBConn().get_suite_id(suite)
+
+ cursor.execute("SELECT DISTINCT s.source FROM source s JOIN binaries b ON (s.id = b.source) JOIN bin_associations ba On (b.id = ba.bin) WHERE b.package = %(package)s AND ba.suite = %(suite)s" , {'package':b, 'suite':suite_id} )
+ while True:
+ s = cursor.fetchone()
+ if not s:
+ break
+
if s[0] != changes["source"]:
reject("%s may not hijack %s from source package %s in suite %s" % (uid, b, s, suite))
################################################################################
def is_unembargo ():
- q = Upload.projectB.query(
- "SELECT package FROM disembargo WHERE package = '%s' AND version = '%s'" %
- (changes["source"], changes["version"]))
- ql = q.getresult()
- if ql:
+ cursor = DBConn().cursor()
+ cursor.execute( "SELECT package FROM disembargo WHERE package = %(source)s AND version = %(version)s", changes )
+ if cursor.fetchone():
return 1
oldcwd = os.getcwd()
if changes["architecture"].has_key("source"):
if Options["No-Action"]: return 1
- Upload.projectB.query(
- "INSERT INTO disembargo (package, version) VALUES ('%s', '%s')" %
- (changes["source"], changes["version"]))
+ cursor.execute( "INSERT INTO disembargo (package, version) VALUES ('%(package)s', '%(version)s')",
+ changes )
+ cursor.execute( "COMMIT" )
return 1
return 0
return 0
if not changes["architecture"].has_key("source"):
- pusuite = database.get_suite_id("proposed-updates")
- q = Upload.projectB.query(
- "SELECT S.source FROM source s JOIN src_associations sa ON (s.id = sa.source) WHERE s.source = '%s' AND s.version = '%s' AND sa.suite = %d" %
- (changes["source"], changes["version"], pusuite))
- ql = q.getresult()
- if ql:
+ pusuite = DBConn().get_suite_id("proposed-updates")
+ cursor = DBConn().cursor()
+ cursor.execute( """SELECT 1 FROM source s
+ JOIN src_associations sa ON (s.id = sa.source)
+ WHERE s.source = %(source)s
+ AND s.version = '%(version)s'
+ AND sa.suite = %(suite)d""",
+ {'source' : changes['source'],
+ 'version' : changes['version'],
+ 'suite' : pasuite})
+
+ if cursor.fetchone():
# source is already in proposed-updates so no need to hold
return 0
return 0
if not changes["architecture"].has_key("source"):
- pusuite = database.get_suite_id("oldstable-proposed-updates")
- q = Upload.projectB.query(
- "SELECT S.source FROM source s JOIN src_associations sa ON (s.id = sa.source) WHERE s.source = '%s' AND s.version = '%s' AND sa.suite = %d" %
- (changes["source"], changes["version"], pusuite))
- ql = q.getresult()
- if ql:
- # source is already in oldstable-proposed-updates so no need to hold
+ pusuite = DBConn().get_suite_id("oldstable-proposed-updates")
+ cursor = DBConn().cursor()
+ cursor.execute( """"SELECT 1 FROM source s
+ JOIN src_associations sa ON (s.id = sa.source)
+ WHERE s.source = %(source)s
+ AND s.version = %(version)s
+ AND sa.suite = %d""",
+ {'source' : changes['source'],
+ 'version' : changes['version'],
+ 'suite' : pasuite})
+ if cursor.fetchone():
return 0
return 1
+++ /dev/null
-#!/usr/bin/env python
-
-"""
-Config access class
-
-@contact: Debian FTPMaster <ftpmaster@debian.org>
-@copyright: 2008 Mark Hymers <mhy@debian.org>
-@license: GNU General Public License version 2 or later
-"""
-
-# This program is free software; you can redistribute it and/or modify
-# it under the terms of the GNU General Public License as published by
-# the Free Software Foundation; either version 2 of the License, or
-# (at your option) any later version.
-
-# This program is distributed in the hope that it will be useful,
-# but WITHOUT ANY WARRANTY; without even the implied warranty of
-# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-# GNU General Public License for more details.
-
-# You should have received a copy of the GNU General Public License
-# along with this program; if not, write to the Free Software
-# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
-
-################################################################################
-
-# <NCommander> mhy, how about "Now with 20% more monty python references"
-
-################################################################################
-
-import apt_pkg
-import socket
-
-from Singleton import Singleton
-
-################################################################################
-
-default_config = "/etc/dak/dak.conf"
-
-def which_conf_file(Cnf):
- res = socket.gethostbyaddr(socket.gethostname())
- if Cnf.get("Config::" + res[0] + "::DakConfig"):
- return Cnf["Config::" + res[0] + "::DakConfig"]
- else:
- return default_config
-
-class Config(Singleton):
- """
- A Config object is a singleton containing
- information about the DAK configuration
- """
- def __init__(self, *args, **kwargs):
- super(Config, self).__init__(*args, **kwargs)
-
- def _readconf(self):
- apt_pkg.init()
-
- self.Cnf = apt_pkg.newConfiguration()
-
- apt_pkg.ReadConfigFileISC(self.Cnf, default_config)
-
- # Check whether our dak.conf was the real one or
- # just a pointer to our main one
- res = socket.gethostbyaddr(socket.gethostname())
- conffile = self.Cnf.get("Config::" + res[0] + "::DakConfig")
- if conffile:
- apt_pkg.ReadConfigFileISC(self.Cnf, conffile)
-
- # Rebind some functions
- # TODO: Clean this up
- self.get = self.Cnf.get
- self.SubTree = self.Cnf.SubTree
- self.ValueList = self.Cnf.ValueList
-
- def _startup(self, *args, **kwargs):
- self._readconf()
-
- def has_key(self, name):
- return self.Cnf.has_key(name)
-
- def __getitem__(self, name):
- return self.Cnf[name]
-
+++ /dev/null
-#!/usr/bin/env python
-
-""" DB access class
-
-@contact: Debian FTPMaster <ftpmaster@debian.org>
-@copyright: 2000, 2001, 2002, 2003, 2004, 2006 James Troup <james@nocrew.org>
-@copyright: 2008-2009 Mark Hymers <mhy@debian.org>
-@copyright: 2009 Joerg Jaspert <joerg@debian.org>
-@copyright: 2009 Mike O'Connor <stew@debian.org>
-@license: GNU General Public License version 2 or later
-"""
-
-# This program is free software; you can redistribute it and/or modify
-# it under the terms of the GNU General Public License as published by
-# the Free Software Foundation; either version 2 of the License, or
-# (at your option) any later version.
-
-# This program is distributed in the hope that it will be useful,
-# but WITHOUT ANY WARRANTY; without even the implied warranty of
-# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-# GNU General Public License for more details.
-
-# You should have received a copy of the GNU General Public License
-# along with this program; if not, write to the Free Software
-# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
-
-################################################################################
-
-# < mhy> I need a funny comment
-# < sgran> two peanuts were walking down a dark street
-# < sgran> one was a-salted
-# * mhy looks up the definition of "funny"
-
-################################################################################
-
-import os
-import psycopg2
-
-from Singleton import Singleton
-from Config import Config
-
-################################################################################
-
-class Cache(object):
- def __init__(self, hashfunc=None):
- if hashfunc:
- self.hashfunc = hashfunc
- else:
- self.hashfunc = lambda x: x['value']
-
- self.data = {}
-
- def SetValue(self, keys, value):
- self.data[self.hashfunc(keys)] = value
-
- def GetValue(self, keys):
- return self.data.get(self.hashfunc(keys))
-
-################################################################################
-
-class DBConn(Singleton):
- """
- database module init.
- """
- def __init__(self, *args, **kwargs):
- super(DBConn, self).__init__(*args, **kwargs)
-
- def _startup(self, *args, **kwargs):
- self.__createconn()
- self.__init_caches()
-
- ## Connection functions
- def __createconn(self):
- cnf = Config()
- connstr = "dbname=%s" % cnf["DB::Name"]
- if cnf["DB::Host"]:
- connstr += " host=%s" % cnf["DB::Host"]
- if cnf["DB::Port"] and cnf["DB::Port"] != "-1":
- connstr += " port=%s" % cnf["DB::Port"]
-
- self.db_con = psycopg2.connect(connstr)
-
- def reconnect(self):
- try:
- self.db_con.close()
- except psycopg2.InterfaceError:
- pass
-
- self.db_con = None
- self.__createconn()
-
- ## Cache functions
- def __init_caches(self):
- self.caches = {'suite': Cache(),
- 'section': Cache(),
- 'priority': Cache(),
- 'override_type': Cache(),
- 'architecture': Cache(),
- 'archive': Cache(),
- 'component': Cache(),
- 'content_path_names': Cache(),
- 'content_file_names': Cache(),
- 'location': Cache(lambda x: '%s_%s_%s' % (x['location'], x['component'], x['location'])),
- 'maintainer': {}, # TODO
- 'keyring': {}, # TODO
- 'source': Cache(lambda x: '%s_%s_' % (x['source'], x['version'])),
- 'files': {}, # TODO
- 'maintainer': {}, # TODO
- 'fingerprint': {}, # TODO
- 'queue': {}, # TODO
- 'uid': {}, # TODO
- 'suite_version': Cache(lambda x: '%s_%s' % (x['source'], x['suite'])),
- }
-
- def clear_caches(self):
- self.__init_caches()
-
- ## Functions to pass through to the database connector
- def cursor(self):
- return self.db_con.cursor()
-
- def commit(self):
- return self.db_con.commit()
-
- ## Get functions
- def __get_single_id(self, query, values, cachename=None):
- # This is a bit of a hack but it's an internal function only
- if cachename is not None:
- res = self.caches[cachename].GetValue(values)
- if res:
- return res
-
- c = self.db_con.cursor()
- c.execute(query, values)
-
- if c.rowcount != 1:
- return None
-
- res = c.fetchone()[0]
-
- if cachename is not None:
- self.caches[cachename].SetValue(values, res)
-
- return res
-
- def __get_id(self, retfield, table, qfield, value):
- query = "SELECT %s FROM %s WHERE %s = %%(value)s" % (retfield, table, qfield)
- return self.__get_single_id(query, {'value': value}, cachename=table)
-
- def get_suite_id(self, suite):
- """
- Returns database id for given C{suite}.
- Results are kept in a cache during runtime to minimize database queries.
-
- @type suite: string
- @param suite: The name of the suite
-
- @rtype: int
- @return: the database id for the given suite
-
- """
- return self.__get_id('id', 'suite', 'suite_name', suite)
-
- def get_section_id(self, section):
- """
- Returns database id for given C{section}.
- Results are kept in a cache during runtime to minimize database queries.
-
- @type section: string
- @param section: The name of the section
-
- @rtype: int
- @return: the database id for the given section
-
- """
- return self.__get_id('id', 'section', 'section', section)
-
- def get_priority_id(self, priority):
- """
- Returns database id for given C{priority}.
- Results are kept in a cache during runtime to minimize database queries.
-
- @type priority: string
- @param priority: The name of the priority
-
- @rtype: int
- @return: the database id for the given priority
-
- """
- return self.__get_id('id', 'priority', 'priority', priority)
-
- def get_override_type_id(self, override_type):
- """
- Returns database id for given override C{type}.
- Results are kept in a cache during runtime to minimize database queries.
-
- @type type: string
- @param type: The name of the override type
-
- @rtype: int
- @return: the database id for the given override type
-
- """
- return self.__get_id('id', 'override_type', 'override_type', override_type)
-
- def get_architecture_id(self, architecture):
- """
- Returns database id for given C{architecture}.
- Results are kept in a cache during runtime to minimize database queries.
-
- @type architecture: string
- @param architecture: The name of the override type
-
- @rtype: int
- @return: the database id for the given architecture
-
- """
- return self.__get_id('id', 'architecture', 'arch_string', architecture)
-
- def get_archive_id(self, archive):
- """
- returns database id for given c{archive}.
- results are kept in a cache during runtime to minimize database queries.
-
- @type archive: string
- @param archive: the name of the override type
-
- @rtype: int
- @return: the database id for the given archive
-
- """
- return self.__get_id('id', 'archive', 'lower(name)', archive)
-
- def get_component_id(self, component):
- """
- Returns database id for given C{component}.
- Results are kept in a cache during runtime to minimize database queries.
-
- @type component: string
- @param component: The name of the override type
-
- @rtype: int
- @return: the database id for the given component
-
- """
- return self.__get_id('id', 'component', 'lower(name)', component)
-
- def get_location_id(self, location, component, archive):
- """
- Returns database id for the location behind the given combination of
- - B{location} - the path of the location, eg. I{/srv/ftp.debian.org/ftp/pool/}
- - B{component} - the id of the component as returned by L{get_component_id}
- - B{archive} - the id of the archive as returned by L{get_archive_id}
- Results are kept in a cache during runtime to minimize database queries.
-
- @type location: string
- @param location: the path of the location
-
- @type component: int
- @param component: the id of the component
-
- @type archive: int
- @param archive: the id of the archive
-
- @rtype: int
- @return: the database id for the location
-
- """
-
- archive_id = self.get_archive_id(archive)
-
- if not archive_id:
- return None
-
- res = None
-
- if component:
- component_id = self.get_component_id(component)
- if component_id:
- res = self.__get_single_id("SELECT id FROM location WHERE path=%(location)s AND component=%(component)d AND archive=%(archive)d",
- {'location': location, 'archive': archive_id, 'component': component_id}, cachename='location')
- else:
- res = self.__get_single_id("SELECT id FROM location WHERE path=%(location)s AND archive=%(archive)d",
- {'location': location, 'archive': archive_id, 'component': ''}, cachename='location')
-
- return res
-
- def get_source_id(self, source, version):
- """
- Returns database id for the combination of C{source} and C{version}
- - B{source} - source package name, eg. I{mailfilter}, I{bbdb}, I{glibc}
- - B{version}
- Results are kept in a cache during runtime to minimize database queries.
-
- @type source: string
- @param source: source package name
-
- @type version: string
- @param version: the source version
-
- @rtype: int
- @return: the database id for the source
-
- """
- return self.__get_single_id("SELECT id FROM source s WHERE s.source=%(source)s AND s.version=%(version)s",
- {'source': source, 'version': version}, cachename='source')
-
- def get_suite_version(self, source, suite):
- """
- Returns database id for a combination of C{source} and C{suite}.
-
- - B{source} - source package name, eg. I{mailfilter}, I{bbdb}, I{glibc}
- - B{suite} - a suite name, eg. I{unstable}
-
- Results are kept in a cache during runtime to minimize database queries.
-
- @type source: string
- @param source: source package name
-
- @type suite: string
- @param suite: the suite name
-
- @rtype: string
- @return: the version for I{source} in I{suite}
-
- """
- return self.__get_single_id("""
- SELECT s.version FROM source s, suite su, src_associations sa
- WHERE sa.source=s.id
- AND sa.suite=su.id
- AND su.suite_name=%(suite)s
- AND s.source=%(source)""", {'suite': suite, 'source': source}, cachename='suite_version')
-
-
- def get_or_set_contents_file_id(self, filename):
- """
- Returns database id for given filename.
-
- Results are kept in a cache during runtime to minimize database queries.
- If no matching file is found, a row is inserted.
-
- @type filename: string
- @param filename: The filename
-
- @rtype: int
- @return: the database id for the given component
- """
- values={'value': filename}
- query = "SELECT id FROM content_file_names WHERE file = %(value)s"
- id = self.__get_single_id(query, values, cachename='content_file_names')
- if not id:
- c = self.db_con.cursor()
- c.execute( "INSERT INTO content_file_names VALUES (DEFAULT, %(value)s) RETURNING id",
- values )
-
- id = c.fetchone()[0]
- self.caches['content_file_names'].SetValue(values, id)
-
- return id
-
- def get_or_set_contents_path_id(self, path):
- """
- Returns database id for given path.
-
- Results are kept in a cache during runtime to minimize database queries.
- If no matching file is found, a row is inserted.
-
- @type path: string
- @param path: The filename
-
- @rtype: int
- @return: the database id for the given component
- """
- values={'value': path}
- query = "SELECT id FROM content_file_paths WHERE path = %(value)s"
- id = self.__get_single_id(query, values, cachename='content_path_names')
- if not id:
- c = self.db_con.cursor()
- c.execute( "INSERT INTO content_file_paths VALUES (DEFAULT, %(value)s) RETURNING id",
- values )
-
- id = c.fetchone()[0]
- self.caches['content_path_names'].SetValue(values, id)
-
- return id
-
- def insert_content_paths(self, bin_id, fullpaths):
- """
- Make sure given path is associated with given binary id
-
- @type bin_id: int
- @param bin_id: the id of the binary
- @type fullpath: string
- @param fullpath: the path of the file being associated with the binary
- """
-
- c = self.db_con.cursor()
-
- for fullpath in fullpaths:
- c.execute( "BEGIN WORK" )
- (path, file) = os.path.split(fullpath)
-
- # Get the necessary IDs ...
- file_id = self.get_or_set_contents_file_id(file)
- path_id = self.get_or_set_contents_path_id(path)
-
- # Determine if we're inserting a duplicate row
-
- c.execute("SELECT 1 FROM content_associations WHERE binary_pkg = '%d' AND filepath = '%d' AND filename = '%d'" % (int(bin_id), path_id, file_id))
- if not c.fetchone():
- # no, we are not, do the insert
-
- c.execute("INSERT INTO content_associations VALUES (DEFAULT, '%d', '%d', '%d')" % (bin_id, path_id, file_id))
- c.execute( "COMMIT" )
+++ /dev/null
-#!/usr/bin/env python
-# vim:set et ts=4 sw=4:
-
-"""
-Singleton pattern code
-
-Inspiration for this very simple ABC was taken from various documents /
-tutorials / mailing lists. This may not be thread safe but given that
-(as I write) large chunks of dak aren't even type-safe, I'll live with
-it for now
-
-@contact: Debian FTPMaster <ftpmaster@debian.org>
-@copyright: 2008 Mark Hymers <mhy@debian.org>
-@license: GNU General Public License version 2 or later
-"""
-
-################################################################################
-
-# This program is free software; you can redistribute it and/or modify
-# it under the terms of the GNU General Public License as published by
-# the Free Software Foundation; either version 2 of the License, or
-# (at your option) any later version.
-
-# This program is distributed in the hope that it will be useful,
-# but WITHOUT ANY WARRANTY; without even the implied warranty of
-# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-# GNU General Public License for more details.
-
-# You should have received a copy of the GNU General Public License
-# along with this program; if not, write to the Free Software
-# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
-
-################################################################################
-
-# < sgran> NCommander: in SQL, it's better to join than to repeat information
-# < tomv_w> that makes SQL the opposite to Debian mailing lists!
-
-################################################################################
-
-"""
-This class set implements objects that may need to be instantiated multiple
-times, but we don't want the overhead of actually creating and init'ing
-them more than once. It also saves us using globals all over the place
-"""
-
-class Singleton(object):
- """This is the ABC for other dak Singleton classes"""
- __single = None
- def __new__(cls, *args, **kwargs):
- # Check to see if a __single exists already for this class
- # Compare class types instead of just looking for None so
- # that subclasses will create their own __single objects
- if cls != type(cls.__single):
- cls.__single = object.__new__(cls, *args, **kwargs)
- cls.__single._startup(*args, **kwargs)
- return cls.__single
-
- def __init__(self, *args, **kwargs):
- if type(self) == "Singleton":
- raise NotImplementedError("Singleton is an ABC")
-
- def _startup(self):
- """
- _startup is a private method used instead of __init__ due to the way
- we instantiate this object
- """
- raise NotImplementedError("Singleton is an ABC")
-
--- /dev/null
+#!/usr/bin/python
+
+"""
+Functions related debian binary packages
+
+@contact: Debian FTPMaster <ftpmaster@debian.org>
+@copyright: 2009 Mike O'Connor <stew@debian.org>
+@license: GNU General Public License version 2 or later
+"""
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, write to the Free Software
+# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
+
+################################################################################
+
+import os
+import shutil
+import tempfile
+import tarfile
+import commands
+import traceback
+from debian_bundle import deb822
+from dbconn import DBConn
+
+class Binary(object):
+ def __init__(self, filename):
+ self.filename = filename
+ self.tmpdir = None
+ self.chunks = None
+
+ def __del__(self):
+ # we need to remove the temporary directory, if we created one
+ if self.tmpdir and os.path.exists(self.tmpdir):
+ shutil.rmtree(self.tmpdir)
+
+ def __scan_ar(self):
+ # get a list of the ar contents
+ if not self.chunks:
+
+ cmd = "ar t %s" % (self.filename)
+
+ (result, output) = commands.getstatusoutput(cmd)
+ if result != 0:
+ rejected = True
+ reject("%s: 'ar t' invocation failed." % (self.filename))
+ reject(utils.prefix_multi_line_string(output, " [ar output:] "), "")
+ self.chunks = output.split('\n')
+
+
+
+ def __unpack(self):
+ # Internal function which extracts the contents of the .ar to
+ # a temporary directory
+
+ if not self.tmpdir:
+ tmpdir = tempfile.mkdtemp()
+ cwd = os.getcwd()
+ try:
+ os.chdir( tmpdir )
+ cmd = "ar x %s %s %s" % (os.path.join(cwd,self.filename), self.chunks[1], self.chunks[2])
+ (result, output) = commands.getstatusoutput(cmd)
+ if result != 0:
+ reject("%s: '%s' invocation failed." % (filename, cmd))
+ reject(utils.prefix_multi_line_string(output, " [ar output:] "), "")
+ else:
+ self.tmpdir = tmpdir
+
+ finally:
+ os.chdir( cwd )
+
+ def valid_deb(self):
+ """
+ Check deb contents making sure the .deb contains:
+ 1. debian-binary
+ 2. control.tar.gz
+ 3. data.tar.gz or data.tar.bz2
+ in that order, and nothing else.
+ """
+ self.__scan_ar()
+ rejected = not self.chunks
+ if len(self.chunks) != 3:
+ rejected = True
+ reject("%s: found %d chunks, expected 3." % (self.filename, len(self.chunks)))
+ if self.chunks[0] != "debian-binary":
+ rejected = True
+ reject("%s: first chunk is '%s', expected 'debian-binary'." % (self.filename, self.chunks[0]))
+ if self.chunks[1] != "control.tar.gz":
+ rejected = True
+ reject("%s: second chunk is '%s', expected 'control.tar.gz'." % (self.filename, self.chunks[1]))
+ if self.chunks[2] not in [ "data.tar.bz2", "data.tar.gz" ]:
+ rejected = True
+ reject("%s: third chunk is '%s', expected 'data.tar.gz' or 'data.tar.bz2'." % (self.filename, self.chunks[2]))
+
+ return not rejected
+
+ def scan_package(self):
+ """
+ Unpack the .deb, do sanity checking, and gather info from it.
+
+ Currently information gathering consists of getting the contents list. In
+ the hopefully near future, it should also include gathering info from the
+ control file.
+
+ @return True if the deb is valid and contents were imported
+ """
+ rejected = not self.valid_deb()
+ self.__unpack()
+
+ if not rejected and self.tmpdir:
+ cwd = os.getcwd()
+ try:
+ os.chdir(self.tmpdir)
+ if self.chunks[1] == "control.tar.gz":
+ control = tarfile.open(os.path.join(self.tmpdir, "control.tar.gz" ), "r:gz")
+ elif self.chunks[1] == "control.tar.bz2":
+ control = tarfile.open(os.path.join(self.tmpdir, "control.tar.bz2" ), "r:bz2")
+
+ pkg = deb822.Packages.iter_paragraphs( control.extractfile('./control') ).next()
+
+ if self.chunks[2] == "data.tar.gz":
+ data = tarfile.open(os.path.join(self.tmpdir, "data.tar.gz"), "r:gz")
+ elif self.chunks[2] == "data.tar.bz2":
+ data = tarfile.open(os.path.join(self.tmpdir, "data.tar.bz2" ), "r:bz2")
+
+ return DBConn().insert_content_paths(pkg, [ tarinfo.name for tarinfo in data if tarinfo.isdir()])
+
+ except:
+ traceback.print_exc()
+
+ return False
+
+ finally:
+ os.chdir( cwd )
+
+
+
+
+if __name__ == "__main__":
+ Binary( "/srv/ftp.debian.org/queue/accepted/halevt_0.1.3-2_amd64.deb" ).scan_package()
+
--- /dev/null
+#!/usr/bin/env python
+
+"""
+Config access class
+
+@contact: Debian FTPMaster <ftpmaster@debian.org>
+@copyright: 2008 Mark Hymers <mhy@debian.org>
+@license: GNU General Public License version 2 or later
+"""
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, write to the Free Software
+# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
+
+################################################################################
+
+# <NCommander> mhy, how about "Now with 20% more monty python references"
+
+################################################################################
+
+import apt_pkg
+import socket
+
+from singleton import Singleton
+
+################################################################################
+
+default_config = "/etc/dak/dak.conf"
+
+def which_conf_file(Cnf):
+ res = socket.gethostbyaddr(socket.gethostname())
+ if Cnf.get("Config::" + res[0] + "::DakConfig"):
+ return Cnf["Config::" + res[0] + "::DakConfig"]
+ else:
+ return default_config
+
+class Config(Singleton):
+ """
+ A Config object is a singleton containing
+ information about the DAK configuration
+ """
+ def __init__(self, *args, **kwargs):
+ super(Config, self).__init__(*args, **kwargs)
+
+ def _readconf(self):
+ apt_pkg.init()
+
+ self.Cnf = apt_pkg.newConfiguration()
+
+ apt_pkg.ReadConfigFileISC(self.Cnf, default_config)
+
+ # Check whether our dak.conf was the real one or
+ # just a pointer to our main one
+ res = socket.gethostbyaddr(socket.gethostname())
+ conffile = self.Cnf.get("Config::" + res[0] + "::DakConfig")
+ if conffile:
+ apt_pkg.ReadConfigFileISC(self.Cnf, conffile)
+
+ # Rebind some functions
+ # TODO: Clean this up
+ self.get = self.Cnf.get
+ self.SubTree = self.Cnf.SubTree
+ self.ValueList = self.Cnf.ValueList
+
+ def _startup(self, *args, **kwargs):
+ self._readconf()
+
+ def has_key(self, name):
+ return self.Cnf.has_key(name)
+
+ def __getitem__(self, name):
+ return self.Cnf[name]
+
uid_id_cache = {} #: cache for uids
suite_version_cache = {} #: cache for suite_versions (packages)
suite_bin_version_cache = {}
-content_path_id_cache = {}
-content_file_id_cache = {}
-insert_contents_file_cache = {}
cache_preloaded = False
################################################################################
q = projectB.query(sql)
return map(lambda x: x[0], q.getresult())
-################################################################################
-
-def get_or_set_contents_file_id(file):
- global content_file_id_cache
-
- if not content_file_id_cache.has_key(file):
- sql_select = "SELECT id FROM content_file_names WHERE file = '%s'" % file
- q = projectB.query(sql_select)
- if not q.getresult():
- # since this can be called within a transaction, we can't use currval
- q = projectB.query("INSERT INTO content_file_names VALUES (DEFAULT, '%s') RETURNING id" % (file))
- content_file_id_cache[file] = int(q.getresult()[0][0])
- return content_file_id_cache[file]
################################################################################
-def get_or_set_contents_path_id(path):
- global content_path_id_cache
+def copy_temporary_contents(package, version, deb):
+ """
+ copy the previously stored contents from the temp table to the permanant one
- if not content_path_id_cache.has_key(path):
- sql_select = "SELECT id FROM content_file_paths WHERE path = '%s'" % path
- q = projectB.query(sql_select)
- if not q.getresult():
- # since this can be called within a transaction, we can't use currval
- q = projectB.query("INSERT INTO content_file_paths VALUES (DEFAULT, '%s') RETURNING id" % (path))
- content_path_id_cache[path] = int(q.getresult()[0][0])
- return content_path_id_cache[path]
+ during process-unchecked, the deb should have been scanned and the
+ contents stored in temp_content_associations
+ """
-################################################################################
+ # first see if contents exist:
-def insert_content_path(bin_id, fullpath):
- global insert_contents_file_cache
- cache_key = "%s_%s" % (bin_id, fullpath)
+ exists = projectB.query("""SELECT 1 FROM temp_content_associations
+ WHERE package='%s' LIMIT 1""" % package ).getresult()
- # have we seen this contents before?
- # probably only revelant during package import
- if insert_contents_file_cache.has_key(cache_key):
- return
+ if not exists:
+ # This should NOT happen. We should have added contents
+ # during process-unchecked. if it did, log an error, and send
+ # an email.
+ subst = {
+ "__PACKAGE__": package,
+ "__VERSION__": version,
+ "__DAK_ADDRESS__": Cnf["Dinstall::MyEmailAddress"]
+ }
- # split the path into basename, and pathname
- (path, file) = os.path.split(fullpath)
+ message = utils.TemplateSubst(Subst, Cnf["Dir::Templates"]+"/bts-categorize")
+ utils.send_mail( message )
- # Get the necessary IDs ...
- file_id = get_or_set_contents_file_id(file)
- path_id = get_or_set_contents_path_id(path)
+ exists = DBConn().insert_content_path(package, version, deb)
- # Determine if we're inserting a duplicate row
- q = projectB.query("SELECT 1 FROM content_associations WHERE binary_pkg = '%d' AND filepath = '%d' AND filename = '%d'" % (int(bin_id), path_id, file_id))
- if q.getresult():
- # Yes we are, return without doing the insert
- return
+ if exists:
+ sql = """INSERT INTO content_associations(binary_pkg,filepath,filename)
+ SELECT currval('binaries_id_seq'), filepath, filename FROM temp_content_associations
+ WHERE package='%s'
+ AND version='%s'""" % (package, version)
+ projectB.query(sql)
+ projectB.query("""DELETE from temp_content_associations
+ WHERE package='%s'
+ AND version='%s'""" % (package, version))
- # Put them into content_assiocations
- projectB.query("INSERT INTO content_associations VALUES (DEFAULT, '%d', '%d', '%d')" % (bin_id, path_id, file_id))
- return
+ return exists
--- /dev/null
+#!/usr/bin/python
+
+""" DB access class
+
+@contact: Debian FTPMaster <ftpmaster@debian.org>
+@copyright: 2000, 2001, 2002, 2003, 2004, 2006 James Troup <james@nocrew.org>
+@copyright: 2008-2009 Mark Hymers <mhy@debian.org>
+@copyright: 2009 Joerg Jaspert <joerg@debian.org>
+@copyright: 2009 Mike O'Connor <stew@debian.org>
+@license: GNU General Public License version 2 or later
+"""
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, write to the Free Software
+# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
+
+################################################################################
+
+# < mhy> I need a funny comment
+# < sgran> two peanuts were walking down a dark street
+# < sgran> one was a-salted
+# * mhy looks up the definition of "funny"
+
+################################################################################
+
+import os
+import psycopg2
+import traceback
+
+from singleton import Singleton
+from config import Config
+
+################################################################################
+
+class Cache(object):
+ def __init__(self, hashfunc=None):
+ if hashfunc:
+ self.hashfunc = hashfunc
+ else:
+ self.hashfunc = lambda x: x['value']
+
+ self.data = {}
+
+ def SetValue(self, keys, value):
+ self.data[self.hashfunc(keys)] = value
+
+ def GetValue(self, keys):
+ return self.data.get(self.hashfunc(keys))
+
+################################################################################
+
+class DBConn(Singleton):
+ """
+ database module init.
+ """
+ def __init__(self, *args, **kwargs):
+ super(DBConn, self).__init__(*args, **kwargs)
+
+ def _startup(self, *args, **kwargs):
+ self.__createconn()
+ self.__init_caches()
+
+ ## Connection functions
+ def __createconn(self):
+ cnf = Config()
+ connstr = "dbname=%s" % cnf["DB::Name"]
+ if cnf["DB::Host"]:
+ connstr += " host=%s" % cnf["DB::Host"]
+ if cnf["DB::Port"] and cnf["DB::Port"] != "-1":
+ connstr += " port=%s" % cnf["DB::Port"]
+
+ self.db_con = psycopg2.connect(connstr)
+
+ def reconnect(self):
+ try:
+ self.db_con.close()
+ except psycopg2.InterfaceError:
+ pass
+
+ self.db_con = None
+ self.__createconn()
+
+ ## Cache functions
+ def __init_caches(self):
+ self.caches = {'suite': Cache(),
+ 'section': Cache(),
+ 'priority': Cache(),
+ 'override_type': Cache(),
+ 'architecture': Cache(),
+ 'archive': Cache(),
+ 'component': Cache(),
+ 'content_path_names': Cache(),
+ 'content_file_names': Cache(),
+ 'location': Cache(lambda x: '%s_%s_%s' % (x['location'], x['component'], x['location'])),
+ 'maintainer': {}, # TODO
+ 'keyring': {}, # TODO
+ 'source': Cache(lambda x: '%s_%s_' % (x['source'], x['version'])),
+ 'files': Cache(lambda x: '%s_%s_' % (x['filename'], x['location'])),
+ 'maintainer': {}, # TODO
+ 'fingerprint': {}, # TODO
+ 'queue': {}, # TODO
+ 'uid': {}, # TODO
+ 'suite_version': Cache(lambda x: '%s_%s' % (x['source'], x['suite'])),
+ }
+
+ def clear_caches(self):
+ self.__init_caches()
+
+ ## Functions to pass through to the database connector
+ def cursor(self):
+ return self.db_con.cursor()
+
+ def commit(self):
+ return self.db_con.commit()
+
+ ## Get functions
+ def __get_single_id(self, query, values, cachename=None):
+ # This is a bit of a hack but it's an internal function only
+ if cachename is not None:
+ res = self.caches[cachename].GetValue(values)
+ if res:
+ return res
+
+ c = self.db_con.cursor()
+ c.execute(query, values)
+
+ if c.rowcount != 1:
+ return None
+
+ res = c.fetchone()[0]
+
+ if cachename is not None:
+ self.caches[cachename].SetValue(values, res)
+
+ return res
+
+ def __get_id(self, retfield, table, qfield, value):
+ query = "SELECT %s FROM %s WHERE %s = %%(value)s" % (retfield, table, qfield)
+ return self.__get_single_id(query, {'value': value}, cachename=table)
+
+ def get_suite_id(self, suite):
+ """
+ Returns database id for given C{suite}.
+ Results are kept in a cache during runtime to minimize database queries.
+
+ @type suite: string
+ @param suite: The name of the suite
+
+ @rtype: int
+ @return: the database id for the given suite
+
+ """
+ return self.__get_id('id', 'suite', 'suite_name', suite)
+
+ def get_section_id(self, section):
+ """
+ Returns database id for given C{section}.
+ Results are kept in a cache during runtime to minimize database queries.
+
+ @type section: string
+ @param section: The name of the section
+
+ @rtype: int
+ @return: the database id for the given section
+
+ """
+ return self.__get_id('id', 'section', 'section', section)
+
+ def get_priority_id(self, priority):
+ """
+ Returns database id for given C{priority}.
+ Results are kept in a cache during runtime to minimize database queries.
+
+ @type priority: string
+ @param priority: The name of the priority
+
+ @rtype: int
+ @return: the database id for the given priority
+
+ """
+ return self.__get_id('id', 'priority', 'priority', priority)
+
+ def get_override_type_id(self, override_type):
+ """
+ Returns database id for given override C{type}.
+ Results are kept in a cache during runtime to minimize database queries.
+
+ @type type: string
+ @param type: The name of the override type
+
+ @rtype: int
+ @return: the database id for the given override type
+
+ """
+ return self.__get_id('id', 'override_type', 'override_type', override_type)
+
+ def get_architecture_id(self, architecture):
+ """
+ Returns database id for given C{architecture}.
+ Results are kept in a cache during runtime to minimize database queries.
+
+ @type architecture: string
+ @param architecture: The name of the override type
+
+ @rtype: int
+ @return: the database id for the given architecture
+
+ """
+ return self.__get_id('id', 'architecture', 'arch_string', architecture)
+
+ def get_archive_id(self, archive):
+ """
+ returns database id for given c{archive}.
+ results are kept in a cache during runtime to minimize database queries.
+
+ @type archive: string
+ @param archive: the name of the override type
+
+ @rtype: int
+ @return: the database id for the given archive
+
+ """
+ return self.__get_id('id', 'archive', 'lower(name)', archive)
+
+ def get_component_id(self, component):
+ """
+ Returns database id for given C{component}.
+ Results are kept in a cache during runtime to minimize database queries.
+
+ @type component: string
+ @param component: The name of the override type
+
+ @rtype: int
+ @return: the database id for the given component
+
+ """
+ return self.__get_id('id', 'component', 'lower(name)', component)
+
+ def get_location_id(self, location, component, archive):
+ """
+ Returns database id for the location behind the given combination of
+ - B{location} - the path of the location, eg. I{/srv/ftp.debian.org/ftp/pool/}
+ - B{component} - the id of the component as returned by L{get_component_id}
+ - B{archive} - the id of the archive as returned by L{get_archive_id}
+ Results are kept in a cache during runtime to minimize database queries.
+
+ @type location: string
+ @param location: the path of the location
+
+ @type component: int
+ @param component: the id of the component
+
+ @type archive: int
+ @param archive: the id of the archive
+
+ @rtype: int
+ @return: the database id for the location
+
+ """
+
+ archive_id = self.get_archive_id(archive)
+
+ if not archive_id:
+ return None
+
+ res = None
+
+ if component:
+ component_id = self.get_component_id(component)
+ if component_id:
+ res = self.__get_single_id("SELECT id FROM location WHERE path=%(location)s AND component=%(component)s AND archive=%(archive)s",
+ {'location': location,
+ 'archive': int(archive_id),
+ 'component': component_id}, cachename='location')
+ else:
+ res = self.__get_single_id("SELECT id FROM location WHERE path=%(location)s AND archive=%(archive)d",
+ {'location': location, 'archive': archive_id, 'component': ''}, cachename='location')
+
+ return res
+
+ def get_source_id(self, source, version):
+ """
+ Returns database id for the combination of C{source} and C{version}
+ - B{source} - source package name, eg. I{mailfilter}, I{bbdb}, I{glibc}
+ - B{version}
+ Results are kept in a cache during runtime to minimize database queries.
+
+ @type source: string
+ @param source: source package name
+
+ @type version: string
+ @param version: the source version
+
+ @rtype: int
+ @return: the database id for the source
+
+ """
+ return self.__get_single_id("SELECT id FROM source s WHERE s.source=%(source)s AND s.version=%(version)s",
+ {'source': source, 'version': version}, cachename='source')
+
+ def get_suite_version(self, source, suite):
+ """
+ Returns database id for a combination of C{source} and C{suite}.
+
+ - B{source} - source package name, eg. I{mailfilter}, I{bbdb}, I{glibc}
+ - B{suite} - a suite name, eg. I{unstable}
+
+ Results are kept in a cache during runtime to minimize database queries.
+
+ @type source: string
+ @param source: source package name
+
+ @type suite: string
+ @param suite: the suite name
+
+ @rtype: string
+ @return: the version for I{source} in I{suite}
+
+ """
+ return self.__get_single_id("""
+ SELECT s.version FROM source s, suite su, src_associations sa
+ WHERE sa.source=s.id
+ AND sa.suite=su.id
+ AND su.suite_name=%(suite)s
+ AND s.source=%(source)""", {'suite': suite, 'source': source}, cachename='suite_version')
+
+
+ def get_files_id (self, filename, size, md5sum, location_id):
+ """
+ Returns -1, -2 or the file_id for filename, if its C{size} and C{md5sum} match an
+ existing copy.
+
+ The database is queried using the C{filename} and C{location_id}. If a file does exist
+ at that location, the existing size and md5sum are checked against the provided
+ parameters. A size or checksum mismatch returns -2. If more than one entry is
+ found within the database, a -1 is returned, no result returns None, otherwise
+ the file id.
+
+ Results are kept in a cache during runtime to minimize database queries.
+
+ @type filename: string
+ @param filename: the filename of the file to check against the DB
+
+ @type size: int
+ @param size: the size of the file to check against the DB
+
+ @type md5sum: string
+ @param md5sum: the md5sum of the file to check against the DB
+
+ @type location_id: int
+ @param location_id: the id of the location as returned by L{get_location_id}
+
+ @rtype: int / None
+ @return: Various return values are possible:
+ - -2: size/checksum error
+ - -1: more than one file found in database
+ - None: no file found in database
+ - int: file id
+
+ """
+ values = {'filename' : filename,
+ 'location' : location_id}
+
+ res = self.caches['files'].GetValue( values )
+
+ if not res:
+ query = """SELECT id, size, md5sum
+ FROM files
+ WHERE filename = %(filename)s AND location = %(location)s"""
+
+ cursor = self.db_con.cursor()
+ cursor.execute( query, values )
+
+ if cursor.rowcount == 0:
+ res = None
+
+ elif cursor.rowcount != 1:
+ res = -1
+
+ else:
+ row = cursor.fetchone()
+
+ if row[1] != size or row[2] != md5sum:
+ res = -2
+
+ else:
+ self.caches[cachename].SetValue(values, row[0])
+ res = row[0]
+
+ return res
+
+
+ def get_or_set_contents_file_id(self, filename):
+ """
+ Returns database id for given filename.
+
+ Results are kept in a cache during runtime to minimize database queries.
+ If no matching file is found, a row is inserted.
+
+ @type filename: string
+ @param filename: The filename
+
+ @rtype: int
+ @return: the database id for the given component
+ """
+ values={'value': filename}
+ query = "SELECT id FROM content_file_names WHERE file = %(value)s"
+ id = self.__get_single_id(query, values, cachename='content_file_names')
+ if not id:
+ c = self.db_con.cursor()
+ c.execute( "INSERT INTO content_file_names VALUES (DEFAULT, %(value)s) RETURNING id",
+ values )
+
+ id = c.fetchone()[0]
+ self.caches['content_file_names'].SetValue(values, id)
+
+ return id
+
+ def get_or_set_contents_path_id(self, path):
+ """
+ Returns database id for given path.
+
+ Results are kept in a cache during runtime to minimize database queries.
+ If no matching file is found, a row is inserted.
+
+ @type path: string
+ @param path: The filename
+
+ @rtype: int
+ @return: the database id for the given component
+ """
+ values={'value': path}
+ query = "SELECT id FROM content_file_paths WHERE path = %(value)s"
+ id = self.__get_single_id(query, values, cachename='content_path_names')
+ if not id:
+ c = self.db_con.cursor()
+ c.execute( "INSERT INTO content_file_paths VALUES (DEFAULT, %(value)s) RETURNING id",
+ values )
+
+ id = c.fetchone()[0]
+ self.caches['content_path_names'].SetValue(values, id)
+
+ return id
+
+ def insert_content_paths(self, package, fullpaths):
+ """
+ Make sure given path is associated with given binary id
+
+ @type bin_id: int
+ @param bin_id: the id of the binary
+ @type fullpath: string
+ @param fullpath: the path of the file being associated with the binary
+
+ @return True upon success
+ """
+
+ c = self.db_con.cursor()
+
+ c.execute("BEGIN WORK")
+ try:
+
+ # Remove any already existing recorded files for this package
+ c.execute("""DELETE FROM temp_content_associations
+ WHERE package=%(Package)s
+ AND version=%(Version)s""", package )
+
+ for fullpath in fullpaths:
+ (path, file) = os.path.split(fullpath)
+
+ # Get the necessary IDs ...
+ file_id = self.get_or_set_contents_file_id(file)
+ path_id = self.get_or_set_contents_path_id(path)
+
+ c.execute("""INSERT INTO temp_content_associations
+ (package, version, filepath, filename)
+ VALUES (%%(Package)s, %%(Version)s, '%d', '%d')""" % (path_id, file_id),
+ package )
+ c.execute("COMMIT")
+ return True
+ except:
+ traceback.print_exc()
+ c.execute("ROLLBACK")
+ return False
--- /dev/null
+#!/usr/bin/env python
+# vim:set et ts=4 sw=4:
+
+"""
+Singleton pattern code
+
+Inspiration for this very simple ABC was taken from various documents /
+tutorials / mailing lists. This may not be thread safe but given that
+(as I write) large chunks of dak aren't even type-safe, I'll live with
+it for now
+
+@contact: Debian FTPMaster <ftpmaster@debian.org>
+@copyright: 2008 Mark Hymers <mhy@debian.org>
+@license: GNU General Public License version 2 or later
+"""
+
+################################################################################
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, write to the Free Software
+# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
+
+################################################################################
+
+# < sgran> NCommander: in SQL, it's better to join than to repeat information
+# < tomv_w> that makes SQL the opposite to Debian mailing lists!
+
+################################################################################
+
+"""
+This class set implements objects that may need to be instantiated multiple
+times, but we don't want the overhead of actually creating and init'ing
+them more than once. It also saves us using globals all over the place
+"""
+
+class Singleton(object):
+ """This is the ABC for other dak Singleton classes"""
+ __single = None
+ def __new__(cls, *args, **kwargs):
+ # Check to see if a __single exists already for this class
+ # Compare class types instead of just looking for None so
+ # that subclasses will create their own __single objects
+ if cls != type(cls.__single):
+ cls.__single = object.__new__(cls, *args, **kwargs)
+ cls.__single._startup(*args, **kwargs)
+ return cls.__single
+
+ def __init__(self, *args, **kwargs):
+ if type(self) == "Singleton":
+ raise NotImplementedError("Singleton is an ABC")
+
+ def _startup(self):
+ """
+ _startup is a private method used instead of __init__ due to the way
+ we instantiate this object
+ """
+ raise NotImplementedError("Singleton is an ABC")
+
Generate a list of flies contained in a .deb
@type filename: string
- @param filename: the path to a .deb
+ @param filename: the path to a data.tar.gz or data.tar.bz2
@rtype: list
@return: a list of files in the data.tar.* portion of the .deb
finally:
if os.path.exists( chunks[2] ):
+ shutil.rmtree( chunks[2] )
os.remove( chunks[2] )
return contents
--- /dev/null
+From: __DAK_ADDRESS__
+X-Debian: DAK
+X-Debian-Package: __PACKAGE__
+MIME-Version: 1.0
+Content-Type: text/plain; charset="utf-8"
+Content-Transfer-Encoding: 8bit
+Subject: Missing contents for __PACKAGE__ in accepted queue
+
+While processing the accepted queue, I didn't have contents in the
+database for __PACKAGE__ version __VERSION__. These contents should
+have been put into the database by process-unchecked when the package
+first arrived.
+
+This is probably stew's fault.
\ No newline at end of file
To: __MAINTAINER_TO__
__BCC__
X-Debian: DAK
-X-Debian-Package: __SOURCE__
+X-Debian-Package: __PACKAGE__
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
projects / dak.git / commitdiff