X-Git-Url: https://git.decadent.org.uk/gitweb/?a=blobdiff_plain;f=dak%2Fcheck_archive.py;h=89c3dcbd82d690141e01bc2c05d981d8f5f99b72;hb=27e00376e81d1c37ff327ee0d39670b266418869;hp=2d9321d68c68f1caf5e7d77799c6eece507cb739;hpb=aa83ebb15882823869b109d370e24ef3efd730f8;p=dak.git diff --git a/dak/check_archive.py b/dak/check_archive.py index 2d9321d6..89c3dcbd 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 +30,27 @@ ################################################################################ -import commands, os, pg, stat, sys, time -import apt_pkg, apt_inst -from daklib import database +import commands +import os +import stat +import sys +import time +import apt_pkg +import apt_inst + +from daklib.dbconn import * from daklib import utils +from daklib.config import Config +from daklib.dak_exceptions import InvalidDscError, ChangesUnicodeError, CantOpenError ################################################################################ -Cnf = None -projectB = None -db_files = {} -waste = 0.0 -excluded = {} +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() ################################################################################ @@ -52,13 +62,13 @@ 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 source-in-one-dir - ensure the source for each package is in one directory timestamps - check for future timestamps in .deb's - tar-gz-in-dsc - ensure each .dsc lists a .tar.gz file + files-in-dsc - ensure each .dsc references appropriate Files validate-indices - ensure files mentioned in Packages & Sources exist files-not-symlinks - check files in the database aren't symlinks validate-builddeps - validate build-dependencies of .dsc files in the archive @@ -68,6 +78,16 @@ 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: @@ -76,8 +96,7 @@ def process_dir (unused, dirname, filenames): if dirname.find('proposed-updates') != -1: return for name in filenames: - filename = os.path.abspath(dirname+'/'+name) - filename = filename.replace('potato-proposed-updates', 'proposed-updates') + filename = os.path.abspath(os.path.join(dirname,name)) 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 "%s" % (filename) @@ -85,25 +104,31 @@ def process_dir (unused, dirname, filenames): ################################################################################ def check_files(): + """ + Prepare the dictionary of existing filenames, then walk through the archive + pool/ directory to compare it. + """ global db_files + cnf = Config() + print "Building list of database files..." - 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() + q = DBConn().session().query(PoolFile).join(Location).order_by('path', 'location') print "Missing files:" db_files.clear() - for i in ql: - filename = os.path.abspath(i[0] + i[1]) + + for f in q.all(): + filename = os.path.abspath(os.path.join(f.location.path, f.filename)) db_files[filename] = "" if os.access(filename, os.R_OK) == 0: - if i[2]: - print "(last used: %s) %s" % (i[2], filename) + if f.last_used: + print "(last used: %s) %s" % (f.last_used, filename) else: print "%s" % (filename) - filename = Cnf["Dir::Override"]+'override.unreferenced' + filename = os.path.join(cnf["Dir::Override"], 'override.unreferenced') if os.path.exists(filename): f = utils.open_file(filename) for filename in f.readlines(): @@ -112,7 +137,7 @@ def check_files(): print "Existent files not in db:" - os.path.walk(Cnf["Dir::Root"]+'pool/', process_dir, None) + os.path.walk(os.path.join(cnf["Dir::Root"], 'pool/'), process_dir, None) print print "%s wasted..." % (utils.size_type(waste)) @@ -120,21 +145,28 @@ def check_files(): ################################################################################ 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(): - f = line[:-1] - try: - utils.parse_changes(f, signing_rules=1) - except InvalidDscError, line: - utils.warn("syntax error in .dsc file '%s', line %s." % (f, line)) - count += 1 + + for src in DBConn().session().query(DBSource).order_by(DBSource.source, DBSource.version): + f = src.poolfile.fullpath + try: + utils.parse_changes(f, signing_rules=1, dsc_file=1) + except InvalidDscError: + utils.warn("syntax error in .dsc file %s" % f) + count += 1 + except ChangesUnicodeError: + utils.warn("found invalid dsc file (%s), not properly utf-8 encoded" % f) + count += 1 + except CantOpenError: + utils.warn("missing dsc file (%s)" % f) + count += 1 + except Exception as e: + utils.warn("miscellaneous error parsing dsc file (%s): %s" % (f, str(e))) + count += 1 if count: utils.warn("Found %s invalid .dsc files." % (count)) @@ -142,86 +174,113 @@ def check_dscs(): ################################################################################ def check_override(): - for suite in [ "stable", "unstable" ]: - print suite - print "-"*len(suite) + """ + Check for missing overrides in stable and unstable. + """ + session = DBConn().session() + + for suite_name in [ "stable", "unstable" ]: + print suite_name + print "-" * len(suite_name) print - suite_id = database.get_suite_id(suite) - q = projectB.query(""" + suite = get_suite(suite_name) + q = session.execute(""" SELECT DISTINCT b.package FROM binaries b, bin_associations ba - WHERE b.id = ba.bin AND ba.suite = %s AND NOT EXISTS - (SELECT 1 FROM override o WHERE o.suite = %s AND o.package = b.package)""" - % (suite_id, suite_id)) - print q - q = projectB.query(""" + WHERE b.id = ba.bin AND ba.suite = :suiteid AND NOT EXISTS + (SELECT 1 FROM override o WHERE o.suite = :suiteid AND o.package = b.package)""" + % {'suiteid': suite.suite_id}) + + for j in q.fetchall(): + print j[0] + + q = session.execute(""" SELECT DISTINCT s.source FROM source s, src_associations sa - WHERE s.id = sa.source AND sa.suite = %s AND NOT EXISTS - (SELECT 1 FROM override o WHERE o.suite = %s and o.package = s.source)""" - % (suite_id, suite_id)) - print q + WHERE s.id = sa.source AND sa.suite = :suiteid AND NOT EXISTS + (SELECT 1 FROM override o WHERE o.suite = :suiteid and o.package = s.source)""" + % {'suiteid': suite.suite_id}) + for j in q.fetchall(): + print j[0] ################################################################################ -# 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;") - for i in q.getresult(): - source_id = i[0] - q2 = projectB.query(""" -SELECT l.path, f.filename FROM files f, dsc_files df, location l WHERE df.source = %s AND f.id = df.file AND l.id = f.location""" - % (source_id)) + + session = DBConn().session() + + q = session.query(DBSource) + for s in q.all(): first_path = "" first_filename = "" - broken = 0 - for j in q2.getresult(): - filename = j[0] + j[1] + broken = False + + qf = session.query(PoolFile).join(Location).join(DSCFile).filter_by(source_id=s.source_id) + for f in qf.all(): + # 0: path + # 1: filename + filename = os.path.join(f.location.path, f.filename) path = os.path.dirname(filename) + if first_path == "": first_path = path first_filename = filename elif first_path != path: symlink = path + '/' + os.path.basename(first_filename) if not os.path.exists(symlink): - broken = 1 - print "WOAH, we got a live one here... %s [%s] {%s}" % (filename, source_id, symlink) + broken = True + print "WOAH, we got a live one here... %s [%s] {%s}" % (filename, s.source_id, symlink) if broken: broken_count += 1 + print "Found %d source packages where the source is not all in one directory." % (broken_count) ################################################################################ - -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") - ql = q.getresult() - - print "Checking file md5sums & sizes..." - for i in ql: - filename = os.path.abspath(i[0] + i[1]) - db_md5sum = i[2] - db_size = int(i[3]) + q = DBConn().session().query(PoolFile) + + print "Checking file checksums & sizes..." + for f in q: + filename = os.path.abspath(os.path.join(f.location.path, f.filename)) + try: - f = utils.open_file(filename) + fi = utils.open_file(filename) except: utils.warn("can't open '%s'." % (filename)) continue - 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)) + if size != f.filesize: + utils.warn("**WARNING** size mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, size, f.filesize)) + + md5sum = apt_pkg.md5sum(fi) + if md5sum != f.md5sum: + utils.warn("**WARNING** md5sum mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, md5sum, f.md5sum)) + + fi.seek(0) + sha1sum = apt_pkg.sha1sum(fi) + if sha1sum != f.sha1sum: + utils.warn("**WARNING** sha1sum mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, sha1sum, f.sha1sum)) + + fi.seek(0) + sha256sum = apt_pkg.sha256sum(fi) + if sha256sum != f.sha256sum: + utils.warn("**WARNING** sha256sum mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, sha256sum, f.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 @@ -231,14 +290,20 @@ 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$'") - ql = q.getresult() + q = DBConn().session().query(PoolFile).filter(PoolFile.filename.like('.deb$')) + db_files.clear() count = 0 - for i in ql: - filename = os.path.abspath(i[0] + i[1]) + + for pf in q.all(): + filename = os.path.abspath(os.path.join(pf.location.path, pf.filename)) if os.access(filename, os.R_OK): f = utils.open_file(filename) current_file = filename @@ -247,38 +312,40 @@ def check_timestamps(): 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(): +def check_files_in_dsc(): + """ + Ensure each .dsc lists appropriate files in its Files field (according + to the format announced in its Format field). + """ count = 0 print "Building list of database files..." - q = projectB.query("SELECT l.path, f.filename FROM files f, location l WHERE f.location = l.id AND f.filename ~ '.dsc$'") - ql = q.getresult() - if ql: + q = DBConn().session().query(PoolFile).filter(PoolFile.filename.like('.dsc$')) + + if q.count() > 0: print "Checking %d files..." % len(ql) else: print "No files to check." - for i in ql: - filename = os.path.abspath(i[0] + i[1]) + + for pf in q.all(): + filename = os.path.abspath(os.path.join(pf.location.path + pf.filename)) + try: # NB: don't enforce .dsc syntax - dsc = utils.parse_changes(filename) + dsc = utils.parse_changes(filename, dsc_file=1) except: utils.fubar("error parsing .dsc file '%s'." % (filename)) - dsc_files = utils.build_file_list(dsc, is_a_dsc=1) - has_tar = 0 - for f in dsc_files.keys(): - m = utils.re_issource.match(f) - if not m: - 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." % (f)) + + reasons = utils.check_dsc_files(filename, dsc) + for r in reasons: + utils.warn(r) + + if len(reasons) > 0: count += 1 if count: @@ -288,10 +355,13 @@ def check_missing_tar_gz_in_dsc(): ################################################################################ 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)) @@ -326,11 +396,14 @@ 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 = 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)) @@ -347,10 +420,13 @@ 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)) - for arch in [ i.lower() for i in architectures ]: + for component in get_component_names(): + architectures = get_suite_architectures(suite) + for arch in [ i.arch_string.lower() for i in architectures ]: if arch == "source": validate_sources(suite, component) elif arch == "all": @@ -361,14 +437,15 @@ 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() + q = DBConn().session().query(PoolFile).filter(PoolFile.filename.like('.dsc$')) - for i in q_files: - filename = os.path.normpath(i[0] + i[1]) + for pf in q.all(): + filename = os.path.abspath(os.path.join(pf.location.path, pf.filename)) if os.access(filename, os.R_OK) == 0: utils.warn("%s: doesn't exist." % (filename)) else: @@ -382,7 +459,7 @@ def chk_bd_process_dir (unused, dirname, filenames): if not name.endswith(".dsc"): continue filename = os.path.abspath(dirname+'/'+name) - dsc = utils.parse_changes(filename) + dsc = utils.parse_changes(filename, dsc_file=1) for field_name in [ "build-depends", "build-depends-indep" ]: field = dsc.get(field_name) if field: @@ -395,22 +472,25 @@ def chk_bd_process_dir (unused, dirname, filenames): ################################################################################ def check_build_depends(): - os.path.walk(Cnf["Dir::Root"], chk_bd_process_dir, None) + """ Validate build-dependencies of .dsc files in the archive """ + cnf = Config() + os.path.walk(cnf["Dir::Root"], chk_bd_process_dir, None) ################################################################################ def main (): - global Cnf, projectB, db_files, waste, excluded + global db_files, waste, excluded + + cnf = Config() - 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) + args = apt_pkg.ParseCommandLine(cnf.Cnf, Arguments, sys.argv) - Options = Cnf.SubTree("Check-Archive::Options") + Options = cnf.SubTree("Check-Archive::Options") if Options["Help"]: usage() @@ -422,11 +502,11 @@ def main (): usage(1) mode = args[0].lower() - projectB = pg.connect(Cnf["DB::Name"], Cnf["DB::Host"], int(Cnf["DB::Port"])) - database.init(Cnf, projectB) + # Initialize DB + DBConn() - if mode == "md5sums": - check_md5sums() + if mode == "checksums": + check_checksums() elif mode == "files": check_files() elif mode == "dsc-syntax": @@ -437,8 +517,8 @@ def main (): check_source_in_one_dir() elif mode == "timestamps": check_timestamps() - elif mode == "tar-gz-in-dsc": - check_missing_tar_gz_in_dsc() + elif mode == "files-in-dsc": + check_files_in_dsc() elif mode == "validate-indices": check_indices_files_exist() elif mode == "files-not-symlinks":