X-Git-Url: https://git.decadent.org.uk/gitweb/?a=blobdiff_plain;f=dak%2Fcheck_archive.py;h=93cc832c1fb0efd41748e894bf607894a20c9c5f;hb=042841e1fe5cfe41755dc53b39e36b6ed19dc794;hp=227cda97b86a3fee80a0eaa3555a7c5c4e2b597d;hpb=d0a12f9f8e9fe208a277e020d244a55eb6a874c4;p=dak.git diff --git a/dak/check_archive.py b/dak/check_archive.py old mode 100644 new mode 100755 index 227cda97..93cc832c --- a/dak/check_archive.py +++ b/dak/check_archive.py @@ -28,8 +28,8 @@ import commands, os, pg, stat, sys, time import apt_pkg, apt_inst -import daklib.database -import daklib.utils +from daklib import database +from daklib import utils ################################################################################ @@ -52,7 +52,7 @@ 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 @@ -105,7 +105,7 @@ def check_files(): filename = Cnf["Dir::Override"]+'override.unreferenced' if os.path.exists(filename): - f = daklib.utils.open_file(filename) + f = utils.open_file(filename) for filename in f.readlines(): filename = filename[:-1] excluded[filename] = "" @@ -115,7 +115,7 @@ def check_files(): 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)) ################################################################################ @@ -127,17 +127,17 @@ def check_dscs(): 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(): f = line[:-1] try: - daklib.utils.parse_changes(f, signing_rules=1) - except daklib.utils.invalid_dsc_format_exc, line: - daklib.utils.warn("syntax error in .dsc file '%s', line %s." % (f, 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)) ################################################################################ @@ -146,7 +146,7 @@ def check_override(): 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 @@ -194,27 +194,41 @@ SELECT l.path, f.filename FROM files f, dsc_files df, location l WHERE df.source ################################################################################ -def check_md5sums(): +def check_checksums(): 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]) db_md5sum = i[2] - db_size = int(i[3]) + db_sha1sum = i[3] + db_sha256sum = i[4] + db_size = int(i[5]) try: - f = 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(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)) + # Until the main database is filled, we need to not spit 500,000 warnings + # every time we scan the archive. Yet another hack (TM) which can go away + # once this is all working + if db_sha1sum is not None and db_sha1sum != '': + 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)) + + if db_sha256sum is not None and db_sha256sum != '': + 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." @@ -240,7 +254,7 @@ def check_timestamps(): for i in ql: filename = os.path.abspath(i[0] + i[1]) if os.access(filename, os.R_OK): - f = daklib.utils.open_file(filename) + f = utils.open_file(filename) current_file = filename sys.stderr.write("Processing %s.\n" % (filename)) apt_inst.debExtract(f, Ent, "control.tar.gz") @@ -265,24 +279,24 @@ 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 f in dsc_files.keys(): - m = daklib.utils.re_issource.match(f) + m = utils.re_issource.match(f) if not m: - daklib.utils.fubar("%s not recognised as source." % (f)) + 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." % (f)) + 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)) ################################################################################ @@ -291,12 +305,12 @@ def validate_sources(suite, component): 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() + 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') @@ -309,7 +323,7 @@ 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) @@ -317,7 +331,7 @@ 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() @@ -330,12 +344,12 @@ def validate_packages(suite, component, architecture): % (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() + 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')) @@ -367,39 +381,13 @@ def check_files_not_symlinks(): 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] 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)) ################################################################################ @@ -408,7 +396,7 @@ 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: @@ -428,7 +416,7 @@ 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)): @@ -441,18 +429,18 @@ def main (): 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": @@ -472,7 +460,7 @@ 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) ################################################################################