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, sys, time
30 import apt_pkg, apt_inst
31 from daklib import database
32 from daklib import utils
33 from daklib.regexes import re_issource
35 ################################################################################
44 current_time = time.time()
46 ################################################################################
48 def usage(exit_code=0):
49 print """Usage: dak check-archive MODE
50 Run various sanity checks of the archive and/or database.
52 -h, --help show this help and exit.
54 The following MODEs are available:
56 checksums - validate the checksums stored in the database
57 files - check files in the database against what's in the archive
58 dsc-syntax - validate the syntax of .dsc files in the archive
59 missing-overrides - check for missing overrides
60 source-in-one-dir - ensure the source for each package is in one directory
61 timestamps - check for future timestamps in .deb's
62 tar-gz-in-dsc - ensure each .dsc lists a .tar.gz file
63 validate-indices - ensure files mentioned in Packages & Sources exist
64 files-not-symlinks - check files in the database aren't symlinks
65 validate-builddeps - validate build-dependencies of .dsc files in the archive
69 ################################################################################
71 def process_dir (unused, dirname, filenames):
72 global waste, db_files, excluded
74 if dirname.find('/disks-') != -1 or dirname.find('upgrade-') != -1:
76 # hack; can't handle .changes files
77 if dirname.find('proposed-updates') != -1:
79 for name in filenames:
80 filename = os.path.abspath(dirname+'/'+name)
81 filename = filename.replace('potato-proposed-updates', 'proposed-updates')
82 if os.path.isfile(filename) and not os.path.islink(filename) and not db_files.has_key(filename) and not excluded.has_key(filename):
83 waste += os.stat(filename)[stat.ST_SIZE]
84 print "%s" % (filename)
86 ################################################################################
91 print "Building list of database files..."
92 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")
95 print "Missing files:"
98 filename = os.path.abspath(i[0] + i[1])
99 db_files[filename] = ""
100 if os.access(filename, os.R_OK) == 0:
102 print "(last used: %s) %s" % (i[2], filename)
104 print "%s" % (filename)
107 filename = Cnf["Dir::Override"]+'override.unreferenced'
108 if os.path.exists(filename):
109 f = utils.open_file(filename)
110 for filename in f.readlines():
111 filename = filename[:-1]
112 excluded[filename] = ""
114 print "Existent files not in db:"
116 os.path.walk(Cnf["Dir::Root"]+'pool/', process_dir, None)
119 print "%s wasted..." % (utils.size_type(waste))
121 ################################################################################
126 for component in Cnf.SubTree("Component").List():
127 if component == "mixed":
129 component = component.lower()
130 list_filename = '%s%s_%s_source.list' % (Cnf["Dir::Lists"], suite, component)
131 list_file = utils.open_file(list_filename)
132 for line in list_file.readlines():
135 utils.parse_changes(f, signing_rules=1)
136 except InvalidDscError, line:
137 utils.warn("syntax error in .dsc file '%s', line %s." % (f, line))
141 utils.warn("Found %s invalid .dsc files." % (count))
143 ################################################################################
145 def check_override():
146 for suite in [ "stable", "unstable" ]:
150 suite_id = database.get_suite_id(suite)
151 q = projectB.query("""
152 SELECT DISTINCT b.package FROM binaries b, bin_associations ba
153 WHERE b.id = ba.bin AND ba.suite = %s AND NOT EXISTS
154 (SELECT 1 FROM override o WHERE o.suite = %s AND o.package = b.package)"""
155 % (suite_id, suite_id))
157 q = projectB.query("""
158 SELECT DISTINCT s.source FROM source s, src_associations sa
159 WHERE s.id = sa.source AND sa.suite = %s AND NOT EXISTS
160 (SELECT 1 FROM override o WHERE o.suite = %s and o.package = s.source)"""
161 % (suite_id, suite_id))
164 ################################################################################
166 # Ensure that the source files for any given package is all in one
167 # directory so that 'apt-get source' works...
169 def check_source_in_one_dir():
170 # Not the most enterprising method, but hey...
172 q = projectB.query("SELECT id FROM source;")
173 for i in q.getresult():
175 q2 = projectB.query("""
176 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"""
181 for j in q2.getresult():
182 filename = j[0] + j[1]
183 path = os.path.dirname(filename)
186 first_filename = filename
187 elif first_path != path:
188 symlink = path + '/' + os.path.basename(first_filename)
189 if not os.path.exists(symlink):
191 print "WOAH, we got a live one here... %s [%s] {%s}" % (filename, source_id, symlink)
194 print "Found %d source packages where the source is not all in one directory." % (broken_count)
196 ################################################################################
198 def check_checksums():
199 print "Getting file information from database..."
200 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")
203 print "Checking file checksums & sizes..."
205 filename = os.path.abspath(i[0] + i[1])
211 f = utils.open_file(filename)
213 utils.warn("can't open '%s'." % (filename))
215 md5sum = apt_pkg.md5sum(f)
216 size = os.stat(filename)[stat.ST_SIZE]
217 if md5sum != db_md5sum:
218 utils.warn("**WARNING** md5sum mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, md5sum, db_md5sum))
220 utils.warn("**WARNING** size mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, size, db_size))
221 # Until the main database is filled, we need to not spit 500,000 warnings
222 # every time we scan the archive. Yet another hack (TM) which can go away
223 # once this is all working
224 if db_sha1sum is not None and db_sha1sum != '':
226 sha1sum = apt_pkg.sha1sum(f)
227 if sha1sum != db_sha1sum:
228 utils.warn("**WARNING** sha1sum mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, sha1sum, db_sha1sum))
230 if db_sha256sum is not None and db_sha256sum != '':
232 sha256sum = apt_pkg.sha256sum(f)
233 if sha256sum != db_sha256sum:
234 utils.warn("**WARNING** sha256sum mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, sha256sum, db_sha256sum))
238 ################################################################################
240 # Check all files for timestamps in the future; common from hardware
241 # (e.g. alpha) which have far-future dates as their default dates.
243 def Ent(Kind,Name,Link,Mode,UID,GID,Size,MTime,Major,Minor):
246 if MTime > current_time:
247 future_files[current_file] = MTime
248 print "%s: %s '%s','%s',%u,%u,%u,%u,%u,%u,%u" % (current_file, Kind,Name,Link,Mode,UID,GID,Size, MTime, Major, Minor)
250 def check_timestamps():
253 q = projectB.query("SELECT l.path, f.filename FROM files f, location l WHERE f.location = l.id AND f.filename ~ '.deb$'")
258 filename = os.path.abspath(i[0] + i[1])
259 if os.access(filename, os.R_OK):
260 f = utils.open_file(filename)
261 current_file = filename
262 sys.stderr.write("Processing %s.\n" % (filename))
263 apt_inst.debExtract(f, Ent, "control.tar.gz")
265 apt_inst.debExtract(f, Ent, "data.tar.gz")
267 print "Checked %d files (out of %d)." % (count, len(db_files.keys()))
269 ################################################################################
271 def check_missing_tar_gz_in_dsc():
274 print "Building list of database files..."
275 q = projectB.query("SELECT l.path, f.filename FROM files f, location l WHERE f.location = l.id AND f.filename ~ '.dsc$'")
278 print "Checking %d files..." % len(ql)
280 print "No files to check."
282 filename = os.path.abspath(i[0] + i[1])
284 # NB: don't enforce .dsc syntax
285 dsc = utils.parse_changes(filename)
287 utils.fubar("error parsing .dsc file '%s'." % (filename))
288 dsc_files = utils.build_file_list(dsc, is_a_dsc=1)
290 for f in dsc_files.keys():
291 m = re_issource.match(f)
293 utils.fubar("%s not recognised as source." % (f))
295 if ftype == "orig.tar.gz" or ftype == "tar.gz":
298 utils.warn("%s has no .tar.gz in the .dsc file." % (f))
302 utils.warn("Found %s invalid .dsc files." % (count))
305 ################################################################################
307 def validate_sources(suite, component):
308 filename = "%s/dists/%s/%s/source/Sources.gz" % (Cnf["Dir::Root"], suite, component)
309 print "Processing %s..." % (filename)
310 # apt_pkg.ParseTagFile needs a real file handle and can't handle a GzipFile instance...
311 (fd, temp_filename) = utils.temp_filename()
312 (result, output) = commands.getstatusoutput("gunzip -c %s > %s" % (filename, temp_filename))
314 sys.stderr.write("Gunzip invocation failed!\n%s\n" % (output))
316 sources = utils.open_file(temp_filename)
317 Sources = apt_pkg.ParseTagFile(sources)
318 while Sources.Step():
319 source = Sources.Section.Find('Package')
320 directory = Sources.Section.Find('Directory')
321 files = Sources.Section.Find('Files')
322 for i in files.split('\n'):
323 (md5, size, name) = i.split()
324 filename = "%s/%s/%s" % (Cnf["Dir::Root"], directory, name)
325 if not os.path.exists(filename):
326 if directory.find("potato") == -1:
327 print "W: %s missing." % (filename)
329 pool_location = utils.poolify (source, component)
330 pool_filename = "%s/%s/%s" % (Cnf["Dir::Pool"], pool_location, name)
331 if not os.path.exists(pool_filename):
332 print "E: %s missing (%s)." % (filename, pool_filename)
335 pool_filename = os.path.normpath(pool_filename)
336 filename = os.path.normpath(filename)
337 src = utils.clean_symlink(pool_filename, filename, Cnf["Dir::Root"])
338 print "Symlinking: %s -> %s" % (filename, src)
339 #os.symlink(src, filename)
341 os.unlink(temp_filename)
343 ########################################
345 def validate_packages(suite, component, architecture):
346 filename = "%s/dists/%s/%s/binary-%s/Packages.gz" \
347 % (Cnf["Dir::Root"], suite, component, architecture)
348 print "Processing %s..." % (filename)
349 # apt_pkg.ParseTagFile needs a real file handle and can't handle a GzipFile instance...
350 (fd, temp_filename) = utils.temp_filename()
351 (result, output) = commands.getstatusoutput("gunzip -c %s > %s" % (filename, temp_filename))
353 sys.stderr.write("Gunzip invocation failed!\n%s\n" % (output))
355 packages = utils.open_file(temp_filename)
356 Packages = apt_pkg.ParseTagFile(packages)
357 while Packages.Step():
358 filename = "%s/%s" % (Cnf["Dir::Root"], Packages.Section.Find('Filename'))
359 if not os.path.exists(filename):
360 print "W: %s missing." % (filename)
362 os.unlink(temp_filename)
364 ########################################
366 def check_indices_files_exist():
367 for suite in [ "stable", "testing", "unstable" ]:
368 for component in Cnf.ValueList("Suite::%s::Components" % (suite)):
369 architectures = Cnf.ValueList("Suite::%s::Architectures" % (suite))
370 for arch in [ i.lower() for i in architectures ]:
372 validate_sources(suite, component)
376 validate_packages(suite, component, arch)
378 ################################################################################
380 def check_files_not_symlinks():
381 print "Building list of database files... ",
383 q = projectB.query("SELECT l.path, f.filename, f.id FROM files f, location l WHERE f.location = l.id")
384 print "done. (%d seconds)" % (int(time.time()-before))
385 q_files = q.getresult()
388 filename = os.path.normpath(i[0] + i[1])
389 if os.access(filename, os.R_OK) == 0:
390 utils.warn("%s: doesn't exist." % (filename))
392 if os.path.islink(filename):
393 utils.warn("%s: is a symlink." % (filename))
395 ################################################################################
397 def chk_bd_process_dir (unused, dirname, filenames):
398 for name in filenames:
399 if not name.endswith(".dsc"):
401 filename = os.path.abspath(dirname+'/'+name)
402 dsc = utils.parse_changes(filename)
403 for field_name in [ "build-depends", "build-depends-indep" ]:
404 field = dsc.get(field_name)
407 apt_pkg.ParseSrcDepends(field)
409 print "E: [%s] %s: %s" % (filename, field_name, field)
412 ################################################################################
414 def check_build_depends():
415 os.path.walk(Cnf["Dir::Root"], chk_bd_process_dir, None)
417 ################################################################################
420 global Cnf, projectB, db_files, waste, excluded
422 Cnf = utils.get_conf()
423 Arguments = [('h',"help","Check-Archive::Options::Help")]
425 if not Cnf.has_key("Check-Archive::Options::%s" % (i)):
426 Cnf["Check-Archive::Options::%s" % (i)] = ""
428 args = apt_pkg.ParseCommandLine(Cnf, Arguments, sys.argv)
430 Options = Cnf.SubTree("Check-Archive::Options")
435 utils.warn("dak check-archive requires at least one argument")
438 utils.warn("dak check-archive accepts only one argument")
440 mode = args[0].lower()
442 projectB = pg.connect(Cnf["DB::Name"], Cnf["DB::Host"], int(Cnf["DB::Port"]))
443 database.init(Cnf, projectB)
445 if mode == "checksums":
447 elif mode == "files":
449 elif mode == "dsc-syntax":
451 elif mode == "missing-overrides":
453 elif mode == "source-in-one-dir":
454 check_source_in_one_dir()
455 elif mode == "timestamps":
457 elif mode == "tar-gz-in-dsc":
458 check_missing_tar_gz_in_dsc()
459 elif mode == "validate-indices":
460 check_indices_files_exist()
461 elif mode == "files-not-symlinks":
462 check_files_not_symlinks()
463 elif mode == "validate-builddeps":
464 check_build_depends()
466 utils.warn("unknown mode '%s'" % (mode))
469 ################################################################################
471 if __name__ == '__main__':