X-Git-Url: https://git.decadent.org.uk/gitweb/?a=blobdiff_plain;f=dak%2Fcheck_archive.py;h=80782908864f5a72ed52ee5fbd55b76976fc3d1a;hb=fdf3c42445b4f11f4cd71634dd2b57cb7d7a4f36;hp=84bc779988fc1847c0a6d6e95207a9c60819d71a;hpb=e9628d0da14a4a046b04ac6c20675432168dcc4a;p=dak.git diff --git a/dak/check_archive.py b/dak/check_archive.py index 84bc7799..80782908 100755 --- a/dak/check_archive.py +++ b/dak/check_archive.py @@ -1,7 +1,11 @@ #!/usr/bin/env python -# Various different sanity checks -# Copyright (C) 2000, 2001, 2002, 2003, 2004, 2006 James Troup +""" Various different sanity checks + +@contact: Debian FTP Master +@copyright: (C) 2000, 2001, 2002, 2003, 2004, 2006 James Troup +@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 @@ -26,21 +30,28 @@ ################################################################################ -import commands, os, pg, stat, sys, time -import apt_pkg, apt_inst -import daklib.database -import daklib.utils +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() ################################################################################ @@ -52,7 +63,7 @@ Run various sanity checks of the archive and/or database. 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 @@ -68,6 +79,16 @@ The following MODEs are available: ################################################################################ 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: @@ -80,68 +101,81 @@ def process_dir (unused, dirname, filenames): 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: - daklib.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 = daklib.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) print - print "%s wasted..." % (daklib.utils.size_type(waste)) + print "%s wasted..." % (utils.size_type(waste)) ################################################################################ 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 = daklib.utils.open_file(list_filename) + list_file = utils.open_file(list_filename) for line in list_file.readlines(): - file = line[:-1] + f = line[:-1] try: - daklib.utils.parse_changes(file, signing_rules=1) - except daklib.utils.invalid_dsc_format_exc, line: - daklib.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: - daklib.utils.warn("Found %s invalid .dsc files." % (count)) + utils.warn("Found %s invalid .dsc files." % (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 = daklib.database.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 @@ -157,10 +191,13 @@ SELECT DISTINCT s.source FROM source s, src_associations sa ################################################################################ -# 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;") @@ -189,34 +226,46 @@ SELECT l.path, f.filename FROM files f, dsc_files df, location l WHERE df.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 = daklib.utils.open_file(filename) + f = utils.open_file(filename) except: - daklib.utils.warn("can't open '%s'." % (filename)) + 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: - daklib.utils.warn("**WARNING** md5sum mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, md5sum, db_md5sum)) + utils.warn("**WARNING** md5sum mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, md5sum, db_md5sum)) if size != db_size: - daklib.utils.warn("**WARNING** size mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, 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 @@ -226,6 +275,11 @@ def Ent(Kind,Name,Link,Mode,UID,GID,Size,MTime,Major,Minor): 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$'") @@ -233,20 +287,23 @@ def check_timestamps(): 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 = daklib.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..." @@ -260,38 +317,41 @@ def check_missing_tar_gz_in_dsc(): filename = os.path.abspath(i[0] + i[1]) try: # NB: don't enforce .dsc syntax - dsc = daklib.utils.parse_changes(filename) + dsc = utils.parse_changes(filename) except: - daklib.utils.fubar("error parsing .dsc file '%s'." % (filename)) - dsc_files = daklib.utils.build_file_list(dsc, is_a_dsc=1) + 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 = daklib.utils.re_issource.match(file) + for f in dsc_files.keys(): + m = re_issource.match(f) if not m: - daklib.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: - daklib.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: - daklib.utils.warn("Found %s invalid .dsc files." % (count)) + utils.warn("Found %s invalid .dsc files." % (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 = daklib.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)) sys.exit(result) - sources = daklib.utils.open_file(temp_filename) + sources = utils.open_file(temp_filename) Sources = apt_pkg.ParseTagFile(sources) while Sources.Step(): source = Sources.Section.Find('Package') @@ -304,7 +364,7 @@ def validate_sources(suite, component): if directory.find("potato") == -1: print "W: %s missing." % (filename) else: - pool_location = daklib.utils.poolify (source, component) + pool_location = utils.poolify (source, component) pool_filename = "%s/%s/%s" % (Cnf["Dir::Pool"], pool_location, name) if not os.path.exists(pool_filename): print "E: %s missing (%s)." % (filename, pool_filename) @@ -312,7 +372,7 @@ def validate_sources(suite, component): # Create symlink pool_filename = os.path.normpath(pool_filename) filename = os.path.normpath(filename) - src = daklib.utils.clean_symlink(pool_filename, filename, Cnf["Dir::Root"]) + src = utils.clean_symlink(pool_filename, filename, Cnf["Dir::Root"]) print "Symlinking: %s -> %s" % (filename, src) #os.symlink(src, filename) sources.close() @@ -321,16 +381,19 @@ def validate_sources(suite, component): ######################################## 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 = daklib.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)) sys.exit(result) - packages = daklib.utils.open_file(temp_filename) + packages = utils.open_file(temp_filename) Packages = apt_pkg.ParseTagFile(packages) while Packages.Step(): filename = "%s/%s" % (Cnf["Dir::Root"], Packages.Section.Find('Filename')) @@ -342,9 +405,12 @@ def validate_packages(suite, component, architecture): ######################################## 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)) + architectures = database.get_suite_architectures(suite) for arch in [ i.lower() for i in architectures ]: if arch == "source": validate_sources(suite, component) @@ -356,45 +422,22 @@ def check_indices_files_exist(): ################################################################################ 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: - daklib.utils.warn("%s: doesn't exist." % (filename)) + utils.warn("%s: doesn't exist." % (filename)) else: if os.path.islink(filename): - daklib.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: -# daklib.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") + utils.warn("%s: is a symlink." % (filename)) ################################################################################ @@ -403,7 +446,7 @@ def chk_bd_process_dir (unused, dirname, filenames): if not name.endswith(".dsc"): continue filename = os.path.abspath(dirname+'/'+name) - dsc = daklib.utils.parse_changes(filename) + dsc = utils.parse_changes(filename) for field_name in [ "build-depends", "build-depends-indep" ]: field = dsc.get(field_name) if field: @@ -416,6 +459,7 @@ def chk_bd_process_dir (unused, dirname, filenames): ################################################################################ def check_build_depends(): + """ Validate build-dependencies of .dsc files in the archive """ os.path.walk(Cnf["Dir::Root"], chk_bd_process_dir, None) ################################################################################ @@ -423,31 +467,31 @@ def check_build_depends(): def main (): global Cnf, projectB, db_files, waste, excluded - Cnf = daklib.utils.get_conf() + Cnf = utils.get_conf() Arguments = [('h',"help","Check-Archive::Options::Help")] for i in [ "help" ]: - if not Cnf.has_key("Check-Archive::Options::%s" % (i)): - Cnf["Check-Archive::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("Check-Archive::Options") if Options["Help"]: - usage() + usage() if len(args) < 1: - daklib.utils.warn("dak check-archive requires at least one argument") + utils.warn("dak check-archive requires at least one argument") usage(1) elif len(args) > 1: - daklib.utils.warn("dak check-archive 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"])) - daklib.database.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": @@ -467,7 +511,7 @@ def main (): elif mode == "validate-builddeps": check_build_depends() else: - daklib.utils.warn("unknown mode '%s'" % (mode)) + utils.warn("unknown mode '%s'" % (mode)) usage(1) ################################################################################