3 """ Various different sanity checks
5 @contact: Debian FTP Master <ftpmaster@debian.org>
6 @copyright: (C) 2000, 2001, 2002, 2003, 2004, 2006 James Troup <james@nocrew.org>
7 @license: GNU General Public License version 2 or later
10 # This program is free software; you can redistribute it and/or modify
11 # it under the terms of the GNU General Public License as published by
12 # the Free Software Foundation; either version 2 of the License, or
13 # (at your option) any later version.
15 # This program is distributed in the hope that it will be useful,
16 # but WITHOUT ANY WARRANTY; without even the implied warranty of
17 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 # GNU General Public License for more details.
20 # You should have received a copy of the GNU General Public License
21 # along with this program; if not, write to the Free Software
22 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
24 ################################################################################
26 # And, lo, a great and menacing voice rose from the depths, and with
27 # great wrath and vehemence it's voice boomed across the
28 # land... ``hehehehehehe... that *tickles*''
31 ################################################################################
41 from daklib.dbconn import *
42 from daklib import utils
43 from daklib.config import Config
45 ################################################################################
47 db_files = {} #: Cache of filenames as known by the database
48 waste = 0.0 #: How many bytes are "wasted" by files not referenced in database
49 excluded = {} #: List of files which are excluded from files check
52 current_time = time.time() #: now()
54 ################################################################################
56 def usage(exit_code=0):
57 print """Usage: dak check-archive MODE
58 Run various sanity checks of the archive and/or database.
60 -h, --help show this help and exit.
62 The following MODEs are available:
64 checksums - validate the checksums stored in the database
65 files - check files in the database against what's in the archive
66 dsc-syntax - validate the syntax of .dsc files in the archive
67 missing-overrides - check for missing overrides
68 source-in-one-dir - ensure the source for each package is in one directory
69 timestamps - check for future timestamps in .deb's
70 files-in-dsc - ensure each .dsc references appropriate Files
71 validate-indices - ensure files mentioned in Packages & Sources exist
72 files-not-symlinks - check files in the database aren't symlinks
73 validate-builddeps - validate build-dependencies of .dsc files in the archive
77 ################################################################################
79 def process_dir (unused, dirname, filenames):
81 Process a directory and output every files name which is not listed already
82 in the C{filenames} or global C{excluded} dictionaries.
85 @param dirname: the directory to look at
88 @param filenames: Known filenames to ignore
90 global waste, db_files, excluded
92 if dirname.find('/disks-') != -1 or dirname.find('upgrade-') != -1:
94 # hack; can't handle .changes files
95 if dirname.find('proposed-updates') != -1:
97 for name in filenames:
98 filename = os.path.abspath(os.path.join(dirname,name))
99 if os.path.isfile(filename) and not os.path.islink(filename) and not db_files.has_key(filename) and not excluded.has_key(filename):
100 waste += os.stat(filename)[stat.ST_SIZE]
101 print "%s" % (filename)
103 ################################################################################
107 Prepare the dictionary of existing filenames, then walk through the archive
108 pool/ directory to compare it.
114 print "Building list of database files..."
115 q = DBConn().session().query(PoolFile).join(Location).order_by('path', 'location')
117 print "Missing files:"
121 filename = os.path.abspath(os.path.join(f.location.path, f.filename))
122 db_files[filename] = ""
123 if os.access(filename, os.R_OK) == 0:
125 print "(last used: %s) %s" % (f.last_used, filename)
127 print "%s" % (filename)
130 filename = os.path.join(cnf["Dir::Override"], 'override.unreferenced')
131 if os.path.exists(filename):
132 f = utils.open_file(filename)
133 for filename in f.readlines():
134 filename = filename[:-1]
135 excluded[filename] = ""
137 print "Existent files not in db:"
139 os.path.walk(os.path.join(cnf["Dir::Root"], 'pool/'), process_dir, None)
142 print "%s wasted..." % (utils.size_type(waste))
144 ################################################################################
148 Parse every .dsc file in the archive and check for it's validity.
153 for dsc_file in DBConn().session().query(DSCFile):
154 f = dsc_file.poolfile.fullpath
156 utils.parse_changes(f, signing_rules=1, dsc_file=1)
157 except InvalidDscError, line:
158 utils.warn("syntax error in .dsc file %s" % f)
160 except ChangesUnicodeError:
161 utils.warn("found invalid dsc file (%s), not properly utf-8 encoded" % f)
165 utils.warn("Found %s invalid .dsc files." % (count))
167 ################################################################################
169 def check_override():
171 Check for missing overrides in stable and unstable.
173 session = DBConn().session()
175 for suite_name in [ "stable", "unstable" ]:
177 print "-" * len(suite_name)
179 suite = get_suite(suite_name)
180 q = session.execute("""
181 SELECT DISTINCT b.package FROM binaries b, bin_associations ba
182 WHERE b.id = ba.bin AND ba.suite = :suiteid AND NOT EXISTS
183 (SELECT 1 FROM override o WHERE o.suite = :suiteid AND o.package = b.package)"""
184 % {'suiteid': suite.suite_id})
186 for j in q.fetchall():
189 q = session.execute("""
190 SELECT DISTINCT s.source FROM source s, src_associations sa
191 WHERE s.id = sa.source AND sa.suite = :suiteid AND NOT EXISTS
192 (SELECT 1 FROM override o WHERE o.suite = :suiteid and o.package = s.source)"""
193 % {'suiteid': suite.suite_id})
194 for j in q.fetchall():
197 ################################################################################
200 def check_source_in_one_dir():
202 Ensure that the source files for any given package is all in one
203 directory so that 'apt-get source' works...
206 # Not the most enterprising method, but hey...
209 session = DBConn().session()
211 q = session.query(DBSource)
217 qf = session.query(PoolFile).join(Location).join(DSCFile).filter_by(source_id=s.source_id)
221 filename = os.path.join(f.location.path, f.filename)
222 path = os.path.dirname(filename)
226 first_filename = filename
227 elif first_path != path:
228 symlink = path + '/' + os.path.basename(first_filename)
229 if not os.path.exists(symlink):
231 print "WOAH, we got a live one here... %s [%s] {%s}" % (filename, s.source_id, symlink)
235 print "Found %d source packages where the source is not all in one directory." % (broken_count)
237 ################################################################################
238 def check_checksums():
242 print "Getting file information from database..."
243 q = DBConn().session().query(PoolFile)
245 print "Checking file checksums & sizes..."
247 filename = os.path.abspath(os.path.join(f.location.path, f.filename))
250 fi = utils.open_file(filename)
252 utils.warn("can't open '%s'." % (filename))
255 size = os.stat(filename)[stat.ST_SIZE]
256 if size != f.filesize:
257 utils.warn("**WARNING** size mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, size, f.filesize))
259 md5sum = apt_pkg.md5sum(fi)
260 if md5sum != f.md5sum:
261 utils.warn("**WARNING** md5sum mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, md5sum, f.md5sum))
264 sha1sum = apt_pkg.sha1sum(fi)
265 if sha1sum != f.sha1sum:
266 utils.warn("**WARNING** sha1sum mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, sha1sum, f.sha1sum))
269 sha256sum = apt_pkg.sha256sum(fi)
270 if sha256sum != f.sha256sum:
271 utils.warn("**WARNING** sha256sum mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, sha256sum, f.sha256sum))
275 ################################################################################
278 def Ent(Kind,Name,Link,Mode,UID,GID,Size,MTime,Major,Minor):
281 if MTime > current_time:
282 future_files[current_file] = MTime
283 print "%s: %s '%s','%s',%u,%u,%u,%u,%u,%u,%u" % (current_file, Kind,Name,Link,Mode,UID,GID,Size, MTime, Major, Minor)
285 def check_timestamps():
287 Check all files for timestamps in the future; common from hardware
288 (e.g. alpha) which have far-future dates as their default dates.
293 q = DBConn().session().query(PoolFile).filter(PoolFile.filename.like('.deb$'))
299 filename = os.path.abspath(os.path.join(pf.location.path, pf.filename))
300 if os.access(filename, os.R_OK):
301 f = utils.open_file(filename)
302 current_file = filename
303 sys.stderr.write("Processing %s.\n" % (filename))
304 apt_inst.debExtract(f, Ent, "control.tar.gz")
306 apt_inst.debExtract(f, Ent, "data.tar.gz")
309 print "Checked %d files (out of %d)." % (count, len(db_files.keys()))
311 ################################################################################
313 def check_files_in_dsc():
315 Ensure each .dsc lists appropriate files in its Files field (according
316 to the format announced in its Format field).
320 print "Building list of database files..."
321 q = DBConn().session().query(PoolFile).filter(PoolFile.filename.like('.dsc$'))
324 print "Checking %d files..." % len(ql)
326 print "No files to check."
329 filename = os.path.abspath(os.path.join(pf.location.path + pf.filename))
332 # NB: don't enforce .dsc syntax
333 dsc = utils.parse_changes(filename, dsc_file=1)
335 utils.fubar("error parsing .dsc file '%s'." % (filename))
337 reasons = utils.check_dsc_files(filename, dsc)
345 utils.warn("Found %s invalid .dsc files." % (count))
348 ################################################################################
350 def validate_sources(suite, component):
352 Ensure files mentioned in Sources exist
354 filename = "%s/dists/%s/%s/source/Sources.gz" % (Cnf["Dir::Root"], suite, component)
355 print "Processing %s..." % (filename)
356 # apt_pkg.ParseTagFile needs a real file handle and can't handle a GzipFile instance...
357 (fd, temp_filename) = utils.temp_filename()
358 (result, output) = commands.getstatusoutput("gunzip -c %s > %s" % (filename, temp_filename))
360 sys.stderr.write("Gunzip invocation failed!\n%s\n" % (output))
362 sources = utils.open_file(temp_filename)
363 Sources = apt_pkg.ParseTagFile(sources)
364 while Sources.Step():
365 source = Sources.Section.Find('Package')
366 directory = Sources.Section.Find('Directory')
367 files = Sources.Section.Find('Files')
368 for i in files.split('\n'):
369 (md5, size, name) = i.split()
370 filename = "%s/%s/%s" % (Cnf["Dir::Root"], directory, name)
371 if not os.path.exists(filename):
372 if directory.find("potato") == -1:
373 print "W: %s missing." % (filename)
375 pool_location = utils.poolify (source, component)
376 pool_filename = "%s/%s/%s" % (Cnf["Dir::Pool"], pool_location, name)
377 if not os.path.exists(pool_filename):
378 print "E: %s missing (%s)." % (filename, pool_filename)
381 pool_filename = os.path.normpath(pool_filename)
382 filename = os.path.normpath(filename)
383 src = utils.clean_symlink(pool_filename, filename, Cnf["Dir::Root"])
384 print "Symlinking: %s -> %s" % (filename, src)
385 #os.symlink(src, filename)
387 os.unlink(temp_filename)
389 ########################################
391 def validate_packages(suite, component, architecture):
393 Ensure files mentioned in Packages exist
395 filename = "%s/dists/%s/%s/binary-%s/Packages.gz" \
396 % (Cnf["Dir::Root"], suite, component, architecture)
397 print "Processing %s..." % (filename)
398 # apt_pkg.ParseTagFile needs a real file handle and can't handle a GzipFile instance...
399 (fd, temp_filename) = utils.temp_filename()
400 (result, output) = commands.getstatusoutput("gunzip -c %s > %s" % (filename, temp_filename))
402 sys.stderr.write("Gunzip invocation failed!\n%s\n" % (output))
404 packages = utils.open_file(temp_filename)
405 Packages = apt_pkg.ParseTagFile(packages)
406 while Packages.Step():
407 filename = "%s/%s" % (Cnf["Dir::Root"], Packages.Section.Find('Filename'))
408 if not os.path.exists(filename):
409 print "W: %s missing." % (filename)
411 os.unlink(temp_filename)
413 ########################################
415 def check_indices_files_exist():
417 Ensure files mentioned in Packages & Sources exist
419 for suite in [ "stable", "testing", "unstable" ]:
420 for component in Cnf.ValueList("Suite::%s::Components" % (suite)):
421 architectures = get_suite_architectures(suite)
422 for arch in [ i.arch_string.lower() for i in architectures ]:
424 validate_sources(suite, component)
428 validate_packages(suite, component, arch)
430 ################################################################################
432 def check_files_not_symlinks():
434 Check files in the database aren't symlinks
436 print "Building list of database files... ",
438 q = DBConn().session().query(PoolFile).filter(PoolFile.filename.like('.dsc$'))
441 filename = os.path.abspath(os.path.join(pf.location.path, pf.filename))
442 if os.access(filename, os.R_OK) == 0:
443 utils.warn("%s: doesn't exist." % (filename))
445 if os.path.islink(filename):
446 utils.warn("%s: is a symlink." % (filename))
448 ################################################################################
450 def chk_bd_process_dir (unused, dirname, filenames):
451 for name in filenames:
452 if not name.endswith(".dsc"):
454 filename = os.path.abspath(dirname+'/'+name)
455 dsc = utils.parse_changes(filename, dsc_file=1)
456 for field_name in [ "build-depends", "build-depends-indep" ]:
457 field = dsc.get(field_name)
460 apt_pkg.ParseSrcDepends(field)
462 print "E: [%s] %s: %s" % (filename, field_name, field)
465 ################################################################################
467 def check_build_depends():
468 """ Validate build-dependencies of .dsc files in the archive """
470 os.path.walk(cnf["Dir::Root"], chk_bd_process_dir, None)
472 ################################################################################
475 global db_files, waste, excluded
479 Arguments = [('h',"help","Check-Archive::Options::Help")]
481 if not cnf.has_key("Check-Archive::Options::%s" % (i)):
482 cnf["Check-Archive::Options::%s" % (i)] = ""
484 args = apt_pkg.ParseCommandLine(cnf.Cnf, Arguments, sys.argv)
486 Options = cnf.SubTree("Check-Archive::Options")
491 utils.warn("dak check-archive requires at least one argument")
494 utils.warn("dak check-archive accepts only one argument")
496 mode = args[0].lower()
501 if mode == "checksums":
503 elif mode == "files":
505 elif mode == "dsc-syntax":
507 elif mode == "missing-overrides":
509 elif mode == "source-in-one-dir":
510 check_source_in_one_dir()
511 elif mode == "timestamps":
513 elif mode == "files-in-dsc":
515 elif mode == "validate-indices":
516 check_indices_files_exist()
517 elif mode == "files-not-symlinks":
518 check_files_not_symlinks()
519 elif mode == "validate-builddeps":
520 check_build_depends()
522 utils.warn("unknown mode '%s'" % (mode))
525 ################################################################################
527 if __name__ == '__main__':