3 # Various different sanity checks
4 # Copyright (C) 2000, 2001, 2002, 2003, 2004, 2006 James Troup <james@nocrew.org>
6 # This program is free software; you can redistribute it and/or modify
7 # it under the terms of the GNU General Public License as published by
8 # the Free Software Foundation; either version 2 of the License, or
9 # (at your option) any later version.
11 # This program is distributed in the hope that it will be useful,
12 # but WITHOUT ANY WARRANTY; without even the implied warranty of
13 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 # GNU General Public License for more details.
16 # You should have received a copy of the GNU General Public License
17 # along with this program; if not, write to the Free Software
18 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
20 ################################################################################
22 # And, lo, a great and menacing voice rose from the depths, and with
23 # great wrath and vehemence it's voice boomed across the
24 # land... ``hehehehehehe... that *tickles*''
27 ################################################################################
29 import commands, os, pg, stat, string, sys, time
30 import dak.lib.database, dak.lib.utils
31 import apt_pkg, apt_inst
33 ################################################################################
42 current_time = time.time()
44 ################################################################################
46 def usage(exit_code=0):
47 print """Usage: dak check-archive MODE
48 Run various sanity checks of the archive and/or database.
50 -h, --help show this help and exit.
52 The following MODEs are available:
54 md5sums - validate the md5sums stored in the database
55 files - check files in the database against what's in the archive
56 dsc-syntax - validate the syntax of .dsc files in the archive
57 missing-overrides - check for missing overrides
58 source-in-one-dir - ensure the source for each package is in one directory
59 timestamps - check for future timestamps in .deb's
60 tar-gz-in-dsc - ensure each .dsc lists a .tar.gz file
61 validate-indices - ensure files mentioned in Packages & Sources exist
62 files-not-symlinks - check files in the database aren't symlinks
63 validate-builddeps - validate build-dependencies of .dsc files in the archive
67 ################################################################################
69 def process_dir (unused, dirname, filenames):
70 global waste, db_files, excluded
72 if dirname.find('/disks-') != -1 or dirname.find('upgrade-') != -1:
74 # hack; can't handle .changes files
75 if dirname.find('proposed-updates') != -1:
77 for name in filenames:
78 filename = os.path.abspath(dirname+'/'+name)
79 filename = filename.replace('potato-proposed-updates', 'proposed-updates')
80 if os.path.isfile(filename) and not os.path.islink(filename) and not db_files.has_key(filename) and not excluded.has_key(filename):
81 waste += os.stat(filename)[stat.ST_SIZE]
84 ################################################################################
89 print "Building list of database files..."
90 q = projectB.query("SELECT l.path, f.filename FROM files f, location l WHERE f.location = l.id")
95 filename = os.path.abspath(i[0] + i[1])
96 db_files[filename] = ""
97 if os.access(filename, os.R_OK) == 0:
98 dak.lib.utils.warn("'%s' doesn't exist." % (filename))
100 filename = Cnf["Dir::Override"]+'override.unreferenced'
101 if os.path.exists(filename):
102 file = dak.lib.utils.open_file(filename)
103 for filename in file.readlines():
104 filename = filename[:-1]
105 excluded[filename] = ""
107 print "Checking against existent files..."
109 os.path.walk(Cnf["Dir::Root"]+'pool/', process_dir, None)
112 print "%s wasted..." % (dak.lib.utils.size_type(waste))
114 ################################################################################
119 for component in Cnf.SubTree("Component").List():
120 if component == "mixed":
122 component = component.lower()
123 list_filename = '%s%s_%s_source.list' % (Cnf["Dir::Lists"], suite, component)
124 list_file = dak.lib.utils.open_file(list_filename)
125 for line in list_file.readlines():
128 dak.lib.utils.parse_changes(file, signing_rules=1)
129 except dak.lib.utils.invalid_dsc_format_exc, line:
130 dak.lib.utils.warn("syntax error in .dsc file '%s', line %s." % (file, line))
134 dak.lib.utils.warn("Found %s invalid .dsc files." % (count))
136 ################################################################################
138 def check_override():
139 for suite in [ "stable", "unstable" ]:
143 suite_id = dak.lib.database.get_suite_id(suite)
144 q = projectB.query("""
145 SELECT DISTINCT b.package FROM binaries b, bin_associations ba
146 WHERE b.id = ba.bin AND ba.suite = %s AND NOT EXISTS
147 (SELECT 1 FROM override o WHERE o.suite = %s AND o.package = b.package)"""
148 % (suite_id, suite_id))
150 q = projectB.query("""
151 SELECT DISTINCT s.source FROM source s, src_associations sa
152 WHERE s.id = sa.source AND sa.suite = %s AND NOT EXISTS
153 (SELECT 1 FROM override o WHERE o.suite = %s and o.package = s.source)"""
154 % (suite_id, suite_id))
157 ################################################################################
159 # Ensure that the source files for any given package is all in one
160 # directory so that 'apt-get source' works...
162 def check_source_in_one_dir():
163 # Not the most enterprising method, but hey...
165 q = projectB.query("SELECT id FROM source;")
166 for i in q.getresult():
168 q2 = projectB.query("""
169 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"""
174 for j in q2.getresult():
175 filename = j[0] + j[1]
176 path = os.path.dirname(filename)
179 first_filename = filename
180 elif first_path != path:
181 symlink = path + '/' + os.path.basename(first_filename)
182 if not os.path.exists(symlink):
184 print "WOAH, we got a live one here... %s [%s] {%s}" % (filename, source_id, symlink)
187 print "Found %d source packages where the source is not all in one directory." % (broken_count)
189 ################################################################################
192 print "Getting file information from database..."
193 q = projectB.query("SELECT l.path, f.filename, f.md5sum, f.size FROM files f, location l WHERE f.location = l.id")
196 print "Checking file md5sums & sizes..."
198 filename = os.path.abspath(i[0] + i[1])
202 file = dak.lib.utils.open_file(filename)
204 dak.lib.utils.warn("can't open '%s'." % (filename))
206 md5sum = apt_pkg.md5sum(file)
207 size = os.stat(filename)[stat.ST_SIZE]
208 if md5sum != db_md5sum:
209 dak.lib.utils.warn("**WARNING** md5sum mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, md5sum, db_md5sum))
211 dak.lib.utils.warn("**WARNING** size mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, size, db_size))
215 ################################################################################
217 # Check all files for timestamps in the future; common from hardware
218 # (e.g. alpha) which have far-future dates as their default dates.
220 def Ent(Kind,Name,Link,Mode,UID,GID,Size,MTime,Major,Minor):
223 if MTime > current_time:
224 future_files[current_file] = MTime
225 print "%s: %s '%s','%s',%u,%u,%u,%u,%u,%u,%u" % (current_file, Kind,Name,Link,Mode,UID,GID,Size, MTime, Major, Minor)
227 def check_timestamps():
230 q = projectB.query("SELECT l.path, f.filename FROM files f, location l WHERE f.location = l.id AND f.filename ~ '.deb$'")
235 filename = os.path.abspath(i[0] + i[1])
236 if os.access(filename, os.R_OK):
237 file = dak.lib.utils.open_file(filename)
238 current_file = filename
239 sys.stderr.write("Processing %s.\n" % (filename))
240 apt_inst.debExtract(file,Ent,"control.tar.gz")
242 apt_inst.debExtract(file,Ent,"data.tar.gz")
244 print "Checked %d files (out of %d)." % (count, len(db_files.keys()))
246 ################################################################################
248 def check_missing_tar_gz_in_dsc():
251 print "Building list of database files..."
252 q = projectB.query("SELECT l.path, f.filename FROM files f, location l WHERE f.location = l.id AND f.filename ~ '.dsc$'")
255 print "Checking %d files..." % len(ql)
257 print "No files to check."
259 filename = os.path.abspath(i[0] + i[1])
261 # NB: don't enforce .dsc syntax
262 dsc = dak.lib.utils.parse_changes(filename)
264 dak.lib.utils.fubar("error parsing .dsc file '%s'." % (filename))
265 dsc_files = dak.lib.utils.build_file_list(dsc, is_a_dsc=1)
267 for file in dsc_files.keys():
268 m = dak.lib.utils.re_issource.match(file)
270 dak.lib.utils.fubar("%s not recognised as source." % (file))
272 if type == "orig.tar.gz" or type == "tar.gz":
275 dak.lib.utils.warn("%s has no .tar.gz in the .dsc file." % (file))
279 dak.lib.utils.warn("Found %s invalid .dsc files." % (count))
282 ################################################################################
284 def validate_sources(suite, component):
285 filename = "%s/dists/%s/%s/source/Sources.gz" % (Cnf["Dir::Root"], suite, component)
286 print "Processing %s..." % (filename)
287 # apt_pkg.ParseTagFile needs a real file handle and can't handle a GzipFile instance...
288 temp_filename = dak.lib.utils.temp_filename()
289 (result, output) = commands.getstatusoutput("gunzip -c %s > %s" % (filename, temp_filename))
291 sys.stderr.write("Gunzip invocation failed!\n%s\n" % (output))
293 sources = dak.lib.utils.open_file(temp_filename)
294 Sources = apt_pkg.ParseTagFile(sources)
295 while Sources.Step():
296 source = Sources.Section.Find('Package')
297 directory = Sources.Section.Find('Directory')
298 files = Sources.Section.Find('Files')
299 for i in files.split('\n'):
300 (md5, size, name) = i.split()
301 filename = "%s/%s/%s" % (Cnf["Dir::Root"], directory, name)
302 if not os.path.exists(filename):
303 if directory.find("potato") == -1:
304 print "W: %s missing." % (filename)
306 pool_location = dak.lib.utils.poolify (source, component)
307 pool_filename = "%s/%s/%s" % (Cnf["Dir::Pool"], pool_location, name)
308 if not os.path.exists(pool_filename):
309 print "E: %s missing (%s)." % (filename, pool_filename)
312 pool_filename = os.path.normpath(pool_filename)
313 filename = os.path.normpath(filename)
314 src = dak.lib.utils.clean_symlink(pool_filename, filename, Cnf["Dir::Root"])
315 print "Symlinking: %s -> %s" % (filename, src)
316 #os.symlink(src, filename)
318 os.unlink(temp_filename)
320 ########################################
322 def validate_packages(suite, component, architecture):
323 filename = "%s/dists/%s/%s/binary-%s/Packages.gz" \
324 % (Cnf["Dir::Root"], suite, component, architecture)
325 print "Processing %s..." % (filename)
326 # apt_pkg.ParseTagFile needs a real file handle and can't handle a GzipFile instance...
327 temp_filename = dak.lib.utils.temp_filename()
328 (result, output) = commands.getstatusoutput("gunzip -c %s > %s" % (filename, temp_filename))
330 sys.stderr.write("Gunzip invocation failed!\n%s\n" % (output))
332 packages = dak.lib.utils.open_file(temp_filename)
333 Packages = apt_pkg.ParseTagFile(packages)
334 while Packages.Step():
335 filename = "%s/%s" % (Cnf["Dir::Root"], Packages.Section.Find('Filename'))
336 if not os.path.exists(filename):
337 print "W: %s missing." % (filename)
339 os.unlink(temp_filename)
341 ########################################
343 def check_indices_files_exist():
344 for suite in [ "stable", "testing", "unstable" ]:
345 for component in Cnf.ValueList("Suite::%s::Components" % (suite)):
346 architectures = Cnf.ValueList("Suite::%s::Architectures" % (suite))
347 for arch in map(string.lower, architectures):
349 validate_sources(suite, component)
353 validate_packages(suite, component, arch)
355 ################################################################################
357 def check_files_not_symlinks():
358 print "Building list of database files... ",
360 q = projectB.query("SELECT l.path, f.filename, f.id FROM files f, location l WHERE f.location = l.id")
361 print "done. (%d seconds)" % (int(time.time()-before))
362 q_files = q.getresult()
365 # q = projectB.query("SELECT l.path, c.name, l.id FROM location l, component c WHERE l.component = c.id")
366 # for i in q.getresult():
367 # path = os.path.normpath(i[0] + i[1])
368 # locations[path] = (i[0], i[2])
370 # q = projectB.query("BEGIN WORK")
372 filename = os.path.normpath(i[0] + i[1])
374 if os.access(filename, os.R_OK) == 0:
375 dak.lib.utils.warn("%s: doesn't exist." % (filename))
377 if os.path.islink(filename):
378 dak.lib.utils.warn("%s: is a symlink." % (filename))
379 # You probably don't want to use the rest of this...
380 # print "%s: is a symlink." % (filename)
381 # dest = os.readlink(filename)
382 # if not os.path.isabs(dest):
383 # dest = os.path.normpath(os.path.join(os.path.dirname(filename), dest))
384 # print "--> %s" % (dest)
385 # # Determine suitable location ID
386 # # [in what must be the suckiest way possible?]
388 # for path in locations.keys():
389 # if dest.find(path) == 0:
390 # (location, location_id) = locations[path]
392 # if not location_id:
393 # dak.lib.utils.fubar("Can't find location for %s (%s)." % (dest, filename))
394 # new_filename = dest.replace(location, "")
395 # q = projectB.query("UPDATE files SET filename = '%s', location = %s WHERE id = %s" % (new_filename, location_id, file_id))
396 # q = projectB.query("COMMIT WORK")
398 ################################################################################
400 def chk_bd_process_dir (unused, dirname, filenames):
401 for name in filenames:
402 if not name.endswith(".dsc"):
404 filename = os.path.abspath(dirname+'/'+name)
405 dsc = dak.lib.utils.parse_changes(filename)
406 for field_name in [ "build-depends", "build-depends-indep" ]:
407 field = dsc.get(field_name)
410 apt_pkg.ParseSrcDepends(field)
412 print "E: [%s] %s: %s" % (filename, field_name, field)
415 ################################################################################
417 def check_build_depends():
418 os.path.walk(Cnf["Dir::Root"], chk_bd_process_dir, None)
420 ################################################################################
423 global Cnf, projectB, db_files, waste, excluded
425 Cnf = dak.lib.utils.get_conf()
426 Arguments = [('h',"help","Check-Archive::Options::Help")]
428 if not Cnf.has_key("Check-Archive::Options::%s" % (i)):
429 Cnf["Check-Archive::Options::%s" % (i)] = ""
431 args = apt_pkg.ParseCommandLine(Cnf, Arguments, sys.argv)
433 Options = Cnf.SubTree("Check-Archive::Options")
438 dak.lib.utils.warn("dak check-archive requires at least one argument")
441 dak.lib.utils.warn("dak check-archive accepts only one argument")
443 mode = args[0].lower()
445 projectB = pg.connect(Cnf["DB::Name"], Cnf["DB::Host"], int(Cnf["DB::Port"]))
446 dak.lib.database.init(Cnf, projectB)
448 if mode == "md5sums":
450 elif mode == "files":
452 elif mode == "dsc-syntax":
454 elif mode == "missing-overrides":
456 elif mode == "source-in-one-dir":
457 check_source_in_one_dir()
458 elif mode == "timestamps":
460 elif mode == "tar-gz-in-dsc":
461 check_missing_tar_gz_in_dsc()
462 elif mode == "validate-indices":
463 check_indices_files_exist()
464 elif mode == "files-not-symlinks":
465 check_files_not_symlinks()
466 elif mode == "validate-builddeps":
467 check_build_depends()
469 dak.lib.utils.warn("unknown mode '%s'" % (mode))
472 ################################################################################
474 if __name__ == '__main__':