3 # Various different sanity checks
4 # Copyright (C) 2000, 2001, 2002, 2003, 2004 James Troup <james@nocrew.org>
5 # $Id: tea,v 1.27 2004-04-01 17:13:11 troup Exp $
7 # This program is free software; you can redistribute it and/or modify
8 # it under the terms of the GNU General Public License as published by
9 # the Free Software Foundation; either version 2 of the License, or
10 # (at your option) any later version.
12 # This program is distributed in the hope that it will be useful,
13 # but WITHOUT ANY WARRANTY; without even the implied warranty of
14 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 # GNU General Public License for more details.
17 # You should have received a copy of the GNU General Public License
18 # along with this program; if not, write to the Free Software
19 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
21 ################################################################################
23 # And, lo, a great and menacing voice rose from the depths, and with
24 # great wrath and vehemence it's voice boomed across the
25 # land... ``hehehehehehe... that *tickles*''
28 ################################################################################
30 import commands, os, pg, stat, string, sys, time;
31 import db_access, utils;
32 import apt_pkg, apt_inst;
34 ################################################################################
43 current_time = time.time();
45 ################################################################################
47 def usage(exit_code=0):
48 print """Usage: tea MODE
49 Run various sanity checks of the archive and/or database.
51 -h, --help show this help and exit.
53 The following MODEs are available:
55 md5sums - validate the md5sums stored in the database
56 files - check files in the database against what's in the archive
57 dsc-syntax - validate the syntax of .dsc files in the archive
58 missing-overrides - check for missing overrides
59 source-in-one-dir - ensure the source for each package is in one directory
60 timestamps - check for future timestamps in .deb's
61 tar-gz-in-dsc - ensure each .dsc lists a .tar.gz file
62 validate-indices - ensure files mentioned in Packages & Sources exist
63 files-not-symlinks - check files in the database aren't symlinks
64 validate-builddeps - validate build-dependencies of .dsc files in the archive
68 ################################################################################
70 def process_dir (unused, dirname, filenames):
71 global waste, db_files, excluded;
73 if dirname.find('/disks-') != -1 or dirname.find('upgrade-') != -1:
75 # hack; can't handle .changes files
76 if dirname.find('proposed-updates') != -1:
78 for name in filenames:
79 filename = os.path.abspath(dirname+'/'+name);
80 filename = filename.replace('potato-proposed-updates', 'proposed-updates');
81 if os.path.isfile(filename) and not os.path.islink(filename) and not db_files.has_key(filename) and not excluded.has_key(filename):
82 waste += os.stat(filename)[stat.ST_SIZE];
85 ################################################################################
90 print "Building list of database files...";
91 q = projectB.query("SELECT l.path, f.filename FROM files f, location l WHERE f.location = l.id")
96 filename = os.path.abspath(i[0] + i[1]);
97 db_files[filename] = "";
98 if os.access(filename, os.R_OK) == 0:
99 utils.warn("'%s' doesn't exist." % (filename));
101 file = utils.open_file(Cnf["Dir::Override"]+'override.unreferenced');
102 for filename in file.readlines():
103 filename = filename[:-1];
104 excluded[filename] = "";
106 print "Checking against existent files...";
108 os.path.walk(Cnf["Dir::Root"]+'dists/', process_dir, None);
111 print "%s wasted..." % (utils.size_type(waste));
113 ################################################################################
118 for component in Cnf.SubTree("Component").List():
119 if component == "mixed":
121 component = component.lower();
122 list_filename = '%s%s_%s_source.list' % (Cnf["Dir::Lists"], suite, component);
123 list_file = utils.open_file(list_filename);
124 for line in list_file.readlines():
127 utils.parse_changes(file, dsc_whitespace_rules=1);
128 except utils.invalid_dsc_format_exc, line:
129 utils.warn("syntax error in .dsc file '%s', line %s." % (file, line));
133 utils.warn("Found %s invalid .dsc files." % (count));
135 ################################################################################
137 def check_override():
138 for suite in [ "stable", "unstable" ]:
140 print "-------------"
142 suite_id = db_access.get_suite_id(suite);
143 q = projectB.query("""
144 SELECT DISTINCT b.package FROM binaries b, bin_associations ba
145 WHERE b.id = ba.bin AND ba.suite = %s AND NOT EXISTS
146 (SELECT 1 FROM override o WHERE o.suite = %s AND o.package = b.package)"""
147 % (suite_id, suite_id));
149 q = projectB.query("""
150 SELECT DISTINCT s.source FROM source s, src_associations sa
151 WHERE s.id = sa.source AND sa.suite = %s AND NOT EXISTS
152 (SELECT 1 FROM override o WHERE o.suite = %s and o.package = s.source)"""
153 % (suite_id, suite_id));
156 ################################################################################
158 # Ensure that the source files for any given package is all in one
159 # directory so that 'apt-get source' works...
161 def check_source_in_one_dir():
162 # Not the most enterprising method, but hey...
164 q = projectB.query("SELECT id FROM source;");
165 for i in q.getresult():
167 q2 = projectB.query("""
168 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"""
173 for j in q2.getresult():
174 filename = j[0] + j[1];
175 path = os.path.dirname(filename);
178 first_filename = filename;
179 elif first_path != path:
180 symlink = path + '/' + os.path.basename(first_filename);
181 if not os.path.exists(symlink):
183 print "WOAH, we got a live one here... %s [%s] {%s}" % (filename, source_id, symlink);
186 print "Found %d source packages where the source is not all in one directory." % (broken_count);
188 ################################################################################
191 print "Getting file information from database...";
192 q = projectB.query("SELECT l.path, f.filename, f.md5sum, f.size FROM files f, location l WHERE f.location = l.id")
195 print "Checking file md5sums & sizes...";
197 filename = os.path.abspath(i[0] + i[1]);
201 file = utils.open_file(filename);
203 utils.warn("can't open '%s'." % (filename));
205 md5sum = apt_pkg.md5sum(file);
206 size = os.stat(filename)[stat.ST_SIZE];
207 if md5sum != db_md5sum:
208 utils.warn("**WARNING** md5sum mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, md5sum, db_md5sum));
210 utils.warn("**WARNING** size mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, size, db_size));
214 ################################################################################
216 # Check all files for timestamps in the future; common from hardware
217 # (e.g. alpha) which have far-future dates as their default dates.
219 def Ent(Kind,Name,Link,Mode,UID,GID,Size,MTime,Major,Minor):
222 if MTime > current_time:
223 future_files[current_file] = MTime;
224 print "%s: %s '%s','%s',%u,%u,%u,%u,%u,%u,%u" % (current_file, Kind,Name,Link,Mode,UID,GID,Size, MTime, Major, Minor);
226 def check_timestamps():
229 q = projectB.query("SELECT l.path, f.filename FROM files f, location l WHERE f.location = l.id AND f.filename ~ '.deb$'")
234 filename = os.path.abspath(i[0] + i[1]);
235 if os.access(filename, os.R_OK):
236 file = utils.open_file(filename);
237 current_file = filename;
238 sys.stderr.write("Processing %s.\n" % (filename));
239 apt_inst.debExtract(file,Ent,"control.tar.gz");
241 apt_inst.debExtract(file,Ent,"data.tar.gz");
243 print "Checked %d files (out of %d)." % (count, len(db_files.keys()));
245 ################################################################################
247 def check_missing_tar_gz_in_dsc():
250 print "Building list of database files...";
251 q = projectB.query("SELECT l.path, f.filename FROM files f, location l WHERE f.location = l.id AND f.filename ~ '.dsc$'");
254 print "Checking %d files..." % len(ql);
256 print "No files to check."
258 filename = os.path.abspath(i[0] + i[1]);
260 # NB: don't enforce .dsc syntax
261 dsc = utils.parse_changes(filename);
263 utils.fubar("error parsing .dsc file '%s'." % (filename));
264 dsc_files = utils.build_file_list(dsc, is_a_dsc=1);
266 for file in dsc_files.keys():
267 m = utils.re_issource.match(file);
269 utils.fubar("%s not recognised as source." % (file));
271 if type == "orig.tar.gz" or type == "tar.gz":
274 utils.warn("%s has no .tar.gz in the .dsc file." % (file));
278 utils.warn("Found %s invalid .dsc files." % (count));
281 ################################################################################
283 def validate_sources(suite, component):
284 filename = "%s/dists/%s/%s/source/Sources.gz" % (Cnf["Dir::Root"], suite, component);
285 print "Processing %s..." % (filename);
286 # apt_pkg.ParseTagFile needs a real file handle and can't handle a GzipFile instance...
287 temp_filename = utils.temp_filename();
288 (result, output) = commands.getstatusoutput("gunzip -c %s > %s" % (filename, temp_filename));
290 sys.stderr.write("Gunzip invocation failed!\n%s\n" % (output));
292 sources = utils.open_file(temp_filename);
293 Sources = apt_pkg.ParseTagFile(sources);
294 while Sources.Step():
295 source = Sources.Section.Find('Package');
296 directory = Sources.Section.Find('Directory');
297 files = Sources.Section.Find('Files');
298 for i in files.split('\n'):
299 (md5, size, name) = i.split();
300 filename = "%s/%s/%s" % (Cnf["Dir::Root"], directory, name);
301 if not os.path.exists(filename):
302 if directory.find("potato") == -1:
303 print "W: %s missing." % (filename);
305 pool_location = utils.poolify (source, component);
306 pool_filename = "%s/%s/%s" % (Cnf["Dir::Pool"], pool_location, name);
307 if not os.path.exists(pool_filename):
308 print "E: %s missing (%s)." % (filename, pool_filename);
311 pool_filename = os.path.normpath(pool_filename);
312 filename = os.path.normpath(filename);
313 src = utils.clean_symlink(pool_filename, filename, Cnf["Dir::Root"]);
314 print "Symlinking: %s -> %s" % (filename, src);
315 #os.symlink(src, filename);
316 os.unlink(temp_filename);
318 ########################################
320 def validate_packages(suite, component, architecture):
321 filename = "%s/dists/%s/%s/binary-%s/Packages" \
322 % (Cnf["Dir::Root"], suite, component, architecture);
323 print "Processing %s..." % (filename);
324 packages = utils.open_file(filename);
325 Packages = apt_pkg.ParseTagFile(packages);
326 while Packages.Step():
327 filename = "%s/%s" % (Cnf["Dir::Root"], Packages.Section.Find('Filename'));
328 if not os.path.exists(filename):
329 print "W: %s missing." % (filename);
332 ########################################
334 def check_indices_files_exist():
335 for suite in [ "stable", "testing", "unstable" ]:
336 for component in Cnf.ValueList("Suite::%s::Components" % (suite)):
337 architectures = Cnf.ValueList("Suite::%s::Architectures" % (suite));
338 for arch in map(string.lower, architectures):
340 validate_sources(suite, component);
344 validate_packages(suite, component, arch);
346 ################################################################################
348 def check_files_not_symlinks():
349 print "Building list of database files... ",;
350 before = time.time();
351 q = projectB.query("SELECT l.path, f.filename, f.id FROM files f, location l WHERE f.location = l.id")
352 print "done. (%d seconds)" % (int(time.time()-before));
353 q_files = q.getresult();
356 # q = projectB.query("SELECT l.path, c.name, l.id FROM location l, component c WHERE l.component = c.id");
357 # for i in q.getresult():
358 # path = os.path.normpath(i[0] + i[1]);
359 # locations[path] = (i[0], i[2]);
361 # q = projectB.query("BEGIN WORK");
363 filename = os.path.normpath(i[0] + i[1]);
365 if os.access(filename, os.R_OK) == 0:
366 utils.warn("%s: doesn't exist." % (filename));
368 if os.path.islink(filename):
369 utils.warn("%s: is a symlink." % (filename));
370 # You probably don't want to use the rest of this...
371 # print "%s: is a symlink." % (filename);
372 # dest = os.readlink(filename);
373 # if not os.path.isabs(dest):
374 # dest = os.path.normpath(os.path.join(os.path.dirname(filename), dest));
375 # print "--> %s" % (dest);
376 # # Determine suitable location ID
377 # # [in what must be the suckiest way possible?]
378 # location_id = None;
379 # for path in locations.keys():
380 # if dest.find(path) == 0:
381 # (location, location_id) = locations[path];
383 # if not location_id:
384 # utils.fubar("Can't find location for %s (%s)." % (dest, filename));
385 # new_filename = dest.replace(location, "");
386 # q = projectB.query("UPDATE files SET filename = '%s', location = %s WHERE id = %s" % (new_filename, location_id, file_id));
387 # q = projectB.query("COMMIT WORK");
389 ################################################################################
391 def chk_bd_process_dir (unused, dirname, filenames):
392 for name in filenames:
393 if not name.endswith(".dsc"):
395 filename = os.path.abspath(dirname+'/'+name);
396 dsc = utils.parse_changes(filename);
397 for field_name in [ "build-depends", "build-depends-indep" ]:
398 field = dsc.get(field_name);
401 apt_pkg.ParseSrcDepends(field);
403 print "E: [%s] %s: %s" % (filename, field_name, field);
406 ################################################################################
408 def check_build_depends():
409 os.path.walk(Cnf["Dir::Root"], chk_bd_process_dir, None);
411 ################################################################################
414 global Cnf, projectB, db_files, waste, excluded;
416 Cnf = utils.get_conf();
417 Arguments = [('h',"help","Tea::Options::Help")];
419 if not Cnf.has_key("Tea::Options::%s" % (i)):
420 Cnf["Tea::Options::%s" % (i)] = "";
422 args = apt_pkg.ParseCommandLine(Cnf, Arguments, sys.argv);
424 Options = Cnf.SubTree("Tea::Options")
429 utils.warn("tea requires at least one argument");
432 utils.warn("tea accepts only one argument");
434 mode = args[0].lower();
436 projectB = pg.connect(Cnf["DB::Name"], Cnf["DB::Host"], int(Cnf["DB::Port"]));
437 db_access.init(Cnf, projectB);
439 if mode == "md5sums":
441 elif mode == "files":
443 elif mode == "dsc-syntax":
445 elif mode == "missing-overrides":
447 elif mode == "source-in-one-dir":
448 check_source_in_one_dir();
449 elif mode == "timestamps":
451 elif mode == "tar-gz-in-dsc":
452 check_missing_tar_gz_in_dsc();
453 elif mode == "validate-indices":
454 check_indices_files_exist();
455 elif mode == "files-not-symlinks":
456 check_files_not_symlinks();
457 elif mode == "validate-builddeps":
458 check_build_depends();
460 utils.warn("unknown mode '%s'" % (mode));
463 ################################################################################
465 if __name__ == '__main__':