#!/usr/bin/env python
-# Various different sanity checks
-# Copyright (C) 2000, 2001, 2002, 2003, 2004 James Troup <james@nocrew.org>
-# $Id: tea,v 1.31 2004-11-27 18:03:11 troup Exp $
+""" Various different sanity checks
+
+@contact: Debian FTP Master <ftpmaster@debian.org>
+@copyright: (C) 2000, 2001, 2002, 2003, 2004, 2006 James Troup <james@nocrew.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
################################################################################
-import commands, os, pg, stat, string, sys, time
-import db_access, utils
-import apt_pkg, apt_inst
+import commands
+import os
+import pg
+import stat
+import sys
+import time
+import apt_pkg
+import apt_inst
+from daklib import database
+from daklib import utils
+from daklib.regexes import re_issource
################################################################################
-Cnf = None
-projectB = None
-db_files = {}
-waste = 0.0
-excluded = {}
+Cnf = None #: Configuration, apt_pkg.Configuration
+projectB = None #: database connection, pgobject
+db_files = {} #: Cache of filenames as known by the database
+waste = 0.0 #: How many bytes are "wasted" by files not referenced in database
+excluded = {} #: List of files which are excluded from files check
current_file = None
future_files = {}
-current_time = time.time()
+current_time = time.time() #: now()
################################################################################
def usage(exit_code=0):
- print """Usage: tea MODE
+ print """Usage: dak check-archive MODE
Run various sanity checks of the archive and/or database.
-h, --help show this help and exit.
The following MODEs are available:
- md5sums - validate the md5sums stored in the database
+ checksums - validate the checksums stored in the database
files - check files in the database against what's in the archive
dsc-syntax - validate the syntax of .dsc files in the archive
missing-overrides - check for missing overrides
################################################################################
def process_dir (unused, dirname, filenames):
+ """
+ Process a directory and output every files name which is not listed already
+ in the C{filenames} or global C{excluded} dictionaries.
+
+ @type dirname: string
+ @param dirname: the directory to look at
+
+ @type filenames: dict
+ @param filenames: Known filenames to ignore
+ """
global waste, db_files, excluded
if dirname.find('/disks-') != -1 or dirname.find('upgrade-') != -1:
filename = filename.replace('potato-proposed-updates', 'proposed-updates')
if os.path.isfile(filename) and not os.path.islink(filename) and not db_files.has_key(filename) and not excluded.has_key(filename):
waste += os.stat(filename)[stat.ST_SIZE]
- print filename
+ print "%s" % (filename)
################################################################################
def check_files():
+ """
+ Prepare the dictionary of existing filenames, then walk through the archive
+ pool/ directory to compare it.
+ """
global db_files
print "Building list of database files..."
- q = projectB.query("SELECT l.path, f.filename FROM files f, location l WHERE f.location = l.id")
+ q = projectB.query("SELECT l.path, f.filename, f.last_used FROM files f, location l WHERE f.location = l.id ORDER BY l.path, f.filename")
ql = q.getresult()
+ print "Missing files:"
db_files.clear()
for i in ql:
- filename = os.path.abspath(i[0] + i[1])
+ filename = os.path.abspath(i[0] + i[1])
db_files[filename] = ""
if os.access(filename, os.R_OK) == 0:
- utils.warn("'%s' doesn't exist." % (filename))
+ if i[2]:
+ print "(last used: %s) %s" % (i[2], filename)
+ else:
+ print "%s" % (filename)
+
filename = Cnf["Dir::Override"]+'override.unreferenced'
if os.path.exists(filename):
- file = utils.open_file(filename)
- for filename in file.readlines():
+ f = utils.open_file(filename)
+ for filename in f.readlines():
filename = filename[:-1]
excluded[filename] = ""
- print "Checking against existent files..."
+ print "Existent files not in db:"
os.path.walk(Cnf["Dir::Root"]+'pool/', process_dir, None)
################################################################################
def check_dscs():
+ """
+ Parse every .dsc file in the archive and check for it's validity.
+ """
count = 0
suite = 'unstable'
for component in Cnf.SubTree("Component").List():
- if component == "mixed":
- continue
component = component.lower()
list_filename = '%s%s_%s_source.list' % (Cnf["Dir::Lists"], suite, component)
list_file = utils.open_file(list_filename)
for line in list_file.readlines():
- file = line[:-1]
+ f = line[:-1]
try:
- utils.parse_changes(file, signing_rules=1)
- except utils.invalid_dsc_format_exc, line:
- utils.warn("syntax error in .dsc file '%s', line %s." % (file, line))
+ utils.parse_changes(f, signing_rules=1)
+ except InvalidDscError, line:
+ utils.warn("syntax error in .dsc file '%s', line %s." % (f, line))
count += 1
if count:
################################################################################
def check_override():
+ """
+ Check for missing overrides in stable and unstable.
+ """
for suite in [ "stable", "unstable" ]:
print suite
print "-"*len(suite)
print
- suite_id = db_access.get_suite_id(suite)
+ suite_id = database.get_suite_id(suite)
q = projectB.query("""
SELECT DISTINCT b.package FROM binaries b, bin_associations ba
WHERE b.id = ba.bin AND ba.suite = %s AND NOT EXISTS
################################################################################
-# Ensure that the source files for any given package is all in one
-# directory so that 'apt-get source' works...
def check_source_in_one_dir():
+ """
+ Ensure that the source files for any given package is all in one
+ directory so that 'apt-get source' works...
+ """
+
# Not the most enterprising method, but hey...
broken_count = 0
q = projectB.query("SELECT id FROM source;")
################################################################################
-def check_md5sums():
+def check_checksums():
+ """
+ Validate all files
+ """
print "Getting file information from database..."
- q = projectB.query("SELECT l.path, f.filename, f.md5sum, f.size FROM files f, location l WHERE f.location = l.id")
+ q = projectB.query("SELECT l.path, f.filename, f.md5sum, f.sha1sum, f.sha256sum, f.size FROM files f, location l WHERE f.location = l.id")
ql = q.getresult()
- print "Checking file md5sums & sizes..."
+ print "Checking file checksums & sizes..."
for i in ql:
- filename = os.path.abspath(i[0] + i[1])
+ filename = os.path.abspath(i[0] + i[1])
db_md5sum = i[2]
- db_size = int(i[3])
+ db_sha1sum = i[3]
+ db_sha256sum = i[4]
+ db_size = int(i[5])
try:
- file = utils.open_file(filename)
+ f = utils.open_file(filename)
except:
utils.warn("can't open '%s'." % (filename))
continue
- md5sum = apt_pkg.md5sum(file)
+ md5sum = apt_pkg.md5sum(f)
size = os.stat(filename)[stat.ST_SIZE]
if md5sum != db_md5sum:
utils.warn("**WARNING** md5sum mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, md5sum, db_md5sum))
if size != db_size:
utils.warn("**WARNING** size mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, size, db_size))
+ f.seek(0)
+ sha1sum = apt_pkg.sha1sum(f)
+ if sha1sum != db_sha1sum:
+ utils.warn("**WARNING** sha1sum mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, sha1sum, db_sha1sum))
+
+ f.seek(0)
+ sha256sum = apt_pkg.sha256sum(f)
+ if sha256sum != db_sha256sum:
+ utils.warn("**WARNING** sha256sum mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, sha256sum, db_sha256sum))
print "Done."
################################################################################
#
-# Check all files for timestamps in the future; common from hardware
-# (e.g. alpha) which have far-future dates as their default dates.
def Ent(Kind,Name,Link,Mode,UID,GID,Size,MTime,Major,Minor):
global future_files
print "%s: %s '%s','%s',%u,%u,%u,%u,%u,%u,%u" % (current_file, Kind,Name,Link,Mode,UID,GID,Size, MTime, Major, Minor)
def check_timestamps():
+ """
+ Check all files for timestamps in the future; common from hardware
+ (e.g. alpha) which have far-future dates as their default dates.
+ """
+
global current_file
q = projectB.query("SELECT l.path, f.filename FROM files f, location l WHERE f.location = l.id AND f.filename ~ '.deb$'")
db_files.clear()
count = 0
for i in ql:
- filename = os.path.abspath(i[0] + i[1])
+ filename = os.path.abspath(i[0] + i[1])
if os.access(filename, os.R_OK):
- file = utils.open_file(filename)
+ f = utils.open_file(filename)
current_file = filename
sys.stderr.write("Processing %s.\n" % (filename))
- apt_inst.debExtract(file,Ent,"control.tar.gz")
- file.seek(0)
- apt_inst.debExtract(file,Ent,"data.tar.gz")
+ apt_inst.debExtract(f, Ent, "control.tar.gz")
+ f.seek(0)
+ apt_inst.debExtract(f, Ent, "data.tar.gz")
count += 1
print "Checked %d files (out of %d)." % (count, len(db_files.keys()))
################################################################################
def check_missing_tar_gz_in_dsc():
+ """
+ Ensure each .dsc lists a .tar.gz file
+ """
count = 0
print "Building list of database files..."
utils.fubar("error parsing .dsc file '%s'." % (filename))
dsc_files = utils.build_file_list(dsc, is_a_dsc=1)
has_tar = 0
- for file in dsc_files.keys():
- m = utils.re_issource.match(file)
+ for f in dsc_files.keys():
+ m = re_issource.match(f)
if not m:
- utils.fubar("%s not recognised as source." % (file))
- type = m.group(3)
- if type == "orig.tar.gz" or type == "tar.gz":
+ utils.fubar("%s not recognised as source." % (f))
+ ftype = m.group(3)
+ if ftype == "orig.tar.gz" or ftype == "tar.gz":
has_tar = 1
if not has_tar:
- utils.warn("%s has no .tar.gz in the .dsc file." % (file))
+ utils.warn("%s has no .tar.gz in the .dsc file." % (f))
count += 1
if count:
################################################################################
def validate_sources(suite, component):
+ """
+ Ensure files mentioned in Sources exist
+ """
filename = "%s/dists/%s/%s/source/Sources.gz" % (Cnf["Dir::Root"], suite, component)
print "Processing %s..." % (filename)
# apt_pkg.ParseTagFile needs a real file handle and can't handle a GzipFile instance...
- temp_filename = utils.temp_filename()
+ (fd, temp_filename) = utils.temp_filename()
(result, output) = commands.getstatusoutput("gunzip -c %s > %s" % (filename, temp_filename))
if (result != 0):
sys.stderr.write("Gunzip invocation failed!\n%s\n" % (output))
########################################
def validate_packages(suite, component, architecture):
+ """
+ Ensure files mentioned in Packages exist
+ """
filename = "%s/dists/%s/%s/binary-%s/Packages.gz" \
% (Cnf["Dir::Root"], suite, component, architecture)
print "Processing %s..." % (filename)
# apt_pkg.ParseTagFile needs a real file handle and can't handle a GzipFile instance...
- temp_filename = utils.temp_filename()
+ (fd, temp_filename) = utils.temp_filename()
(result, output) = commands.getstatusoutput("gunzip -c %s > %s" % (filename, temp_filename))
if (result != 0):
sys.stderr.write("Gunzip invocation failed!\n%s\n" % (output))
########################################
def check_indices_files_exist():
+ """
+ Ensure files mentioned in Packages & Sources exist
+ """
for suite in [ "stable", "testing", "unstable" ]:
for component in Cnf.ValueList("Suite::%s::Components" % (suite)):
- architectures = Cnf.ValueList("Suite::%s::Architectures" % (suite))
- for arch in map(string.lower, architectures):
+ architectures = database.get_suite_architectures(suite)
+ for arch in [ i.lower() for i in architectures ]:
if arch == "source":
validate_sources(suite, component)
elif arch == "all":
################################################################################
def check_files_not_symlinks():
+ """
+ Check files in the database aren't symlinks
+ """
print "Building list of database files... ",
before = time.time()
q = projectB.query("SELECT l.path, f.filename, f.id FROM files f, location l WHERE f.location = l.id")
print "done. (%d seconds)" % (int(time.time()-before))
q_files = q.getresult()
-# locations = {}
-# q = projectB.query("SELECT l.path, c.name, l.id FROM location l, component c WHERE l.component = c.id")
-# for i in q.getresult():
-# path = os.path.normpath(i[0] + i[1])
-# locations[path] = (i[0], i[2])
-
-# q = projectB.query("BEGIN WORK")
for i in q_files:
- filename = os.path.normpath(i[0] + i[1])
-# file_id = i[2]
+ filename = os.path.normpath(i[0] + i[1])
if os.access(filename, os.R_OK) == 0:
utils.warn("%s: doesn't exist." % (filename))
else:
if os.path.islink(filename):
utils.warn("%s: is a symlink." % (filename))
- # You probably don't want to use the rest of this...
-# print "%s: is a symlink." % (filename)
-# dest = os.readlink(filename)
-# if not os.path.isabs(dest):
-# dest = os.path.normpath(os.path.join(os.path.dirname(filename), dest))
-# print "--> %s" % (dest)
-# # Determine suitable location ID
-# # [in what must be the suckiest way possible?]
-# location_id = None
-# for path in locations.keys():
-# if dest.find(path) == 0:
-# (location, location_id) = locations[path]
-# break
-# if not location_id:
-# utils.fubar("Can't find location for %s (%s)." % (dest, filename))
-# new_filename = dest.replace(location, "")
-# q = projectB.query("UPDATE files SET filename = '%s', location = %s WHERE id = %s" % (new_filename, location_id, file_id))
-# q = projectB.query("COMMIT WORK")
################################################################################
################################################################################
def check_build_depends():
+ """ Validate build-dependencies of .dsc files in the archive """
os.path.walk(Cnf["Dir::Root"], chk_bd_process_dir, None)
################################################################################
global Cnf, projectB, db_files, waste, excluded
Cnf = utils.get_conf()
- Arguments = [('h',"help","Tea::Options::Help")]
+ Arguments = [('h',"help","Check-Archive::Options::Help")]
for i in [ "help" ]:
- if not Cnf.has_key("Tea::Options::%s" % (i)):
- Cnf["Tea::Options::%s" % (i)] = ""
+ if not Cnf.has_key("Check-Archive::Options::%s" % (i)):
+ Cnf["Check-Archive::Options::%s" % (i)] = ""
args = apt_pkg.ParseCommandLine(Cnf, Arguments, sys.argv)
- Options = Cnf.SubTree("Tea::Options")
+ Options = Cnf.SubTree("Check-Archive::Options")
if Options["Help"]:
- usage()
+ usage()
if len(args) < 1:
- utils.warn("tea requires at least one argument")
+ utils.warn("dak check-archive requires at least one argument")
usage(1)
elif len(args) > 1:
- utils.warn("tea accepts only one argument")
+ utils.warn("dak check-archive accepts only one argument")
usage(1)
mode = args[0].lower()
projectB = pg.connect(Cnf["DB::Name"], Cnf["DB::Host"], int(Cnf["DB::Port"]))
- db_access.init(Cnf, projectB)
+ database.init(Cnf, projectB)
- if mode == "md5sums":
- check_md5sums()
+ if mode == "checksums":
+ check_checksums()
elif mode == "files":
check_files()
elif mode == "dsc-syntax":
if __name__ == '__main__':
main()
-
projects / dak.git / blobdiff