3 # Sanity check the database
4 # Copyright (C) 2000, 2001, 2002, 2003 James Troup <james@nocrew.org>
5 # $Id: tea,v 1.23 2003-09-07 13:52:17 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, tempfile, 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
66 ################################################################################
68 def process_dir (unused, dirname, filenames):
69 global waste, db_files, excluded;
71 if dirname.find('/disks-') != -1 or dirname.find('upgrade-') != -1:
73 # hack; can't handle .changes files
74 if dirname.find('proposed-updates') != -1:
76 for name in filenames:
77 filename = os.path.abspath(dirname+'/'+name);
78 filename = filename.replace('potato-proposed-updates', 'proposed-updates');
79 if os.path.isfile(filename) and not os.path.islink(filename) and not db_files.has_key(filename) and not excluded.has_key(filename):
80 waste += os.stat(filename)[stat.ST_SIZE];
83 ################################################################################
88 print "Building list of database files...";
89 q = projectB.query("SELECT l.path, f.filename FROM files f, location l WHERE f.location = l.id")
94 filename = os.path.abspath(i[0] + i[1]);
95 db_files[filename] = "";
96 if os.access(filename, os.R_OK) == 0:
97 utils.warn("'%s' doesn't exist." % (filename));
99 file = utils.open_file(Cnf["Dir::Override"]+'override.unreferenced');
100 for filename in file.readlines():
101 filename = filename[:-1];
102 excluded[filename] = "";
104 print "Checking against existent files...";
106 os.path.walk(Cnf["Dir::Root"]+'dists/', process_dir, None);
109 print "%s wasted..." % (utils.size_type(waste));
111 ################################################################################
116 for component in Cnf.SubTree("Component").List():
117 if component == "mixed":
119 component = component.lower();
120 list_filename = '%s%s_%s_source.list' % (Cnf["Dir::Lists"], suite, component);
121 list_file = utils.open_file(list_filename);
122 for line in list_file.readlines():
125 utils.parse_changes(file, dsc_whitespace_rules=1);
126 except utils.invalid_dsc_format_exc, line:
127 utils.warn("syntax error in .dsc file '%s', line %s." % (file, line));
131 utils.warn("Found %s invalid .dsc files." % (count));
133 ################################################################################
135 def check_override():
136 for suite in [ "stable", "unstable" ]:
138 print "-------------"
140 suite_id = db_access.get_suite_id(suite);
141 q = projectB.query("""
142 SELECT DISTINCT b.package FROM binaries b, bin_associations ba
143 WHERE b.id = ba.bin AND ba.suite = %s AND NOT EXISTS
144 (SELECT 1 FROM override o WHERE o.suite = %s AND o.package = b.package)"""
145 % (suite_id, suite_id));
147 q = projectB.query("""
148 SELECT DISTINCT s.source FROM source s, src_associations sa
149 WHERE s.id = sa.source AND sa.suite = %s AND NOT EXISTS
150 (SELECT 1 FROM override o WHERE o.suite = %s and o.package = s.source)"""
151 % (suite_id, suite_id));
154 ################################################################################
156 # Ensure that the source files for any given package is all in one
157 # directory so that 'apt-get source' works...
159 def check_source_in_one_dir():
160 # Not the most enterprising method, but hey...
162 q = projectB.query("SELECT id FROM source;");
163 for i in q.getresult():
165 q2 = projectB.query("""
166 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"""
171 for j in q2.getresult():
172 filename = j[0] + j[1];
173 path = os.path.dirname(filename);
176 first_filename = filename;
177 elif first_path != path:
178 symlink = path + '/' + os.path.basename(first_filename);
179 if not os.path.exists(symlink):
181 print "WOAH, we got a live one here... %s [%s] {%s}" % (filename, source_id, symlink);
184 print "Found %d source packages where the source is not all in one directory." % (broken_count);
186 ################################################################################
189 print "Getting file information from database...";
190 q = projectB.query("SELECT l.path, f.filename, f.md5sum, f.size FROM files f, location l WHERE f.location = l.id")
193 print "Checking file md5sums & sizes...";
195 filename = os.path.abspath(i[0] + i[1]);
199 file = utils.open_file(filename);
201 utils.warn("can't open '%s'." % (filename));
203 md5sum = apt_pkg.md5sum(file);
204 size = os.stat(filename)[stat.ST_SIZE];
205 if md5sum != db_md5sum:
206 utils.warn("**WARNING** md5sum mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, md5sum, db_md5sum));
208 utils.warn("**WARNING** size mismatch for '%s' ('%s' [current] vs. '%s' [db])." % (filename, size, db_size));
212 ################################################################################
214 # Check all files for timestamps in the future; common from hardware
215 # (e.g. alpha) which have far-future dates as their default dates.
217 def Ent(Kind,Name,Link,Mode,UID,GID,Size,MTime,Major,Minor):
220 if MTime > current_time:
221 future_files[current_file] = MTime;
222 print "%s: %s '%s','%s',%u,%u,%u,%u,%u,%u,%u" % (current_file, Kind,Name,Link,Mode,UID,GID,Size, MTime, Major, Minor);
224 def check_timestamps():
227 q = projectB.query("SELECT l.path, f.filename FROM files f, location l WHERE f.location = l.id AND f.filename ~ '.deb$'")
232 filename = os.path.abspath(i[0] + i[1]);
233 if os.access(filename, os.R_OK):
234 file = utils.open_file(filename);
235 current_file = filename;
236 sys.stderr.write("Processing %s.\n" % (filename));
237 apt_inst.debExtract(file,Ent,"control.tar.gz");
239 apt_inst.debExtract(file,Ent,"data.tar.gz");
241 print "Checked %d files (out of %d)." % (count, len(db_files.keys()));
243 ################################################################################
245 def check_missing_tar_gz_in_dsc():
248 print "Building list of database files...";
249 q = projectB.query("SELECT l.path, f.filename FROM files f, location l WHERE f.location = l.id AND f.filename ~ '.dsc$'");
252 print "Checking %d files..." % len(ql);
254 print "No files to check."
256 filename = os.path.abspath(i[0] + i[1]);
258 # NB: don't enforce .dsc syntax
259 dsc = utils.parse_changes(filename);
261 utils.fubar("error parsing .dsc file '%s'." % (filename));
262 dsc_files = utils.build_file_list(dsc, is_a_dsc=1);
264 for file in dsc_files.keys():
265 m = utils.re_issource.match(file);
267 utils.fubar("%s not recognised as source." % (file));
269 if type == "orig.tar.gz" or type == "tar.gz":
272 utils.warn("%s has no .tar.gz in the .dsc file." % (file));
276 utils.warn("Found %s invalid .dsc files." % (count));
279 ################################################################################
281 def validate_sources(suite, component):
282 filename = "%s/dists/%s/%s/source/Sources.gz" % (Cnf["Dir::Root"], suite, component);
283 print "Processing %s..." % (filename);
284 # apt_pkg.ParseTagFile needs a real file handle and can't handle a GzipFile instance...
285 temp_filename = tempfile.mktemp();
286 fd = os.open(temp_filename, os.O_RDWR|os.O_CREAT|os.O_EXCL, 0700);
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'):
300 (md5, size, name) = s;
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 = 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 = utils.clean_symlink(pool_filename, filename, Cnf["Dir::Root"]);
315 print "Symlinking: %s -> %s" % (filename, src);
316 #os.symlink(src, filename);
317 os.unlink(temp_filename);
319 ########################################
321 def validate_packages(suite, component, architecture):
322 filename = "%s/dists/%s/%s/binary-%s/Packages" \
323 % (Cnf["Dir::Root"], suite, component, architecture);
324 print "Processing %s..." % (filename);
325 packages = utils.open_file(filename);
326 Packages = apt_pkg.ParseTagFile(packages);
327 while Packages.Step():
328 filename = "%s/%s" % (Cnf["Dir::Root"], Packages.Section.Find('Filename'));
329 if not os.path.exists(filename):
330 print "W: %s missing." % (filename);
333 ########################################
335 def check_indices_files_exist():
336 for suite in [ "stable", "testing", "unstable" ]:
337 for component in Cnf.ValueList("Suite::%s::Components" % (suite)):
338 architectures = Cnf.ValueList("Suite::%s::Architectures" % (suite));
339 for arch in map(string.lower, architectures):
341 validate_sources(suite, component);
345 validate_packages(suite, component, arch);
347 ################################################################################
350 global Cnf, projectB, db_files, waste, excluded;
352 Cnf = utils.get_conf();
353 Arguments = [('h',"help","Tea::Options::Help")];
355 if not Cnf.has_key("Tea::Options::%s" % (i)):
356 Cnf["Tea::Options::%s" % (i)] = "";
358 args = apt_pkg.ParseCommandLine(Cnf, Arguments, sys.argv);
360 Options = Cnf.SubTree("Tea::Options")
365 utils.warn("tea requires at least one argument");
368 utils.warn("tea accepts only one argument");
370 mode = args[0].lower();
372 projectB = pg.connect(Cnf["DB::Name"], Cnf["DB::Host"], int(Cnf["DB::Port"]));
373 db_access.init(Cnf, projectB);
375 if mode == "md5sums":
377 elif mode == "files":
379 elif mode == "dsc-syntax":
381 elif mode == "missing-overrides":
383 elif mode == "source-in-one-dir":
384 check_source_in_one_dir();
385 elif mode == "timestamps":
387 elif mode == "tar-gz-in-dsc":
388 check_missing_tar_gz_in_dsc();
389 elif mode == "validate-indices":
390 check_indices_files_exist();
392 utils.warn("unknown mode '%s'" % (mode));
395 ################################################################################
397 if __name__ == '__main__':