4 Script to automate some parts of checking NEW packages
6 Most functions are written in a functional programming style. They
7 return a string avoiding the side effect of directly printing the string
8 to stdout. Those functions can be used in multithreaded parts of dak.
10 @contact: Debian FTP Master <ftpmaster@debian.org>
11 @copyright: 2000, 2001, 2002, 2003, 2006 James Troup <james@nocrew.org>
12 @copyright: 2009 Joerg Jaspert <joerg@debian.org>
13 @license: GNU General Public License version 2 or later
16 # This program is free software; you can redistribute it and/or modify
17 # it under the terms of the GNU General Public License as published by
18 # the Free Software Foundation; either version 2 of the License, or
19 # (at your option) any later version.
21 # This program is distributed in the hope that it will be useful,
22 # but WITHOUT ANY WARRANTY; without even the implied warranty of
23 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
24 # GNU General Public License for more details.
26 # You should have received a copy of the GNU General Public License
27 # along with this program; if not, write to the Free Software
28 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
30 ################################################################################
32 # <Omnic> elmo wrote docs?!!?!?!?!?!?!
33 # <aj> as if he wasn't scary enough before!!
34 # * aj imagines a little red furry toy sitting hunched over a computer
35 # tapping furiously and giggling to himself
36 # <aj> eventually he stops, and his heads slowly spins around and you
37 # see this really evil grin and then he sees you, and picks up a
38 # knife from beside the keyboard and throws it at you, and as you
39 # breathe your last breath, he starts giggling again
40 # <aj> but i should be telling this to my psychiatrist, not you guys,
43 ################################################################################
45 # suppress some deprecation warnings in squeeze related to md5 module
47 warnings.filterwarnings('ignore', \
48 "the md5 module is deprecated; use hashlib instead", \
62 from daklib import utils
63 from daklib.dbconn import DBConn, get_component_by_package_suite
64 from daklib.regexes import html_escaping, re_html_escaping, re_version, re_spacestrip, \
65 re_contrib, re_nonfree, re_localhost, re_newlinespace, \
66 re_package, re_doc_directory
68 ################################################################################
71 Cnf = utils.get_conf()
73 printed = threading.local()
74 printed.copyrights = {}
75 package_relations = {} #: Store relations of packages for later output
77 # default is to not output html.
80 ################################################################################
82 def usage (exit_code=0):
83 print """Usage: dak examine-package [PACKAGE]...
86 -h, --help show this help and exit
87 -H, --html-output output html page with inspection result
88 -f, --file-name filename for the html page
90 PACKAGE can be a .changes, .dsc, .deb or .udeb filename."""
94 ################################################################################
95 # probably xml.sax.saxutils would work as well
97 def escape_if_needed(s):
99 return re_html_escaping.sub(lambda x: html_escaping.get(x.group(0)), s)
103 def headline(s, level=2, bodyelement=None):
107 <tr><th colspan="2" class="title" onclick="toggle('%(bodyelement)s', 'table-row-group', 'table-row-group')">%(title)s <span class="toggle-msg">(click to toggle)</span></th></tr>
108 </thead>\n"""%{"bodyelement":bodyelement,"title":utils.html_escape(s)}
110 return "<h%d>%s</h%d>\n" % (level, utils.html_escape(s), level)
112 return "---- %s ----\n" % (s)
114 # Colour definitions, 'end' isn't really for use
118 'contrib': "\033[33m",
119 'nonfree': "\033[31m",
120 'provides': "\033[35m",
124 'maintainer': "\033[32m",
125 'distro': "\033[1m\033[41m"}
128 'main': ('<span style="color: aqua">',"</span>"),
129 'contrib': ('<span style="color: yellow">',"</span>"),
130 'nonfree': ('<span style="color: red">',"</span>"),
131 'provides': ('<span style="color: magenta">',"</span>"),
132 'arch': ('<span style="color: green">',"</span>"),
133 'bold': ('<span style="font-weight: bold">',"</span>"),
134 'maintainer': ('<span style="color: green">',"</span>"),
135 'distro': ('<span style="font-weight: bold; background-color: red">',"</span>")}
137 def colour_output(s, colour):
139 return ("%s%s%s" % (html_colours[colour][0], utils.html_escape(s), html_colours[colour][1]))
141 return ("%s%s%s" % (ansi_colours[colour], s, ansi_colours['end']))
143 def escaped_text(s, strip=False):
147 return "<pre>%s</pre>" % (s)
151 def formatted_text(s, strip=False):
155 return "<pre>%s</pre>" % (utils.html_escape(s))
161 return """<tr><td>"""+s+"""</td></tr>"""
165 def format_field(k,v):
167 return """<tr><td class="key">%s:</td><td class="val">%s</td></tr>"""%(k,v)
169 return "%s: %s"%(k,v)
171 def foldable_output(title, elementnameprefix, content, norow=False):
172 d = {'elementnameprefix':elementnameprefix}
175 result += """<div id="%(elementnameprefix)s-wrap"><a name="%(elementnameprefix)s" />
176 <table class="infobox rfc822">\n"""%d
177 result += headline(title, bodyelement="%(elementnameprefix)s-body"%d)
179 result += """ <tbody id="%(elementnameprefix)s-body" class="infobody">\n"""%d
181 result += content + "\n"
183 result += output_row(content) + "\n"
185 result += """</tbody></table></div>"""
188 ################################################################################
190 def get_depends_parts(depend) :
191 v_match = re_version.match(depend)
193 d_parts = { 'name' : v_match.group(1), 'version' : v_match.group(2) }
195 d_parts = { 'name' : depend , 'version' : '' }
198 def get_or_list(depend) :
199 or_list = depend.split("|")
202 def get_comma_list(depend) :
203 dep_list = depend.split(",")
206 def split_depends (d_str) :
207 # creates a list of lists of dictionaries of depends (package,version relation)
209 d_str = re_spacestrip.sub('',d_str)
211 # first split depends string up amongs comma delimiter
212 dep_list = get_comma_list(d_str)
214 while d < len(dep_list):
215 # put depends into their own list
216 depends_tree.append([dep_list[d]])
219 while d < len(depends_tree):
221 # split up Or'd depends into a multi-item list
222 depends_tree[d] = get_or_list(depends_tree[d][0])
223 while k < len(depends_tree[d]):
224 # split depends into {package, version relation}
225 depends_tree[d][k] = get_depends_parts(depends_tree[d][k])
230 def read_control (filename):
237 deb_file = utils.open_file(filename)
239 extracts = apt_inst.debExtractControl(deb_file)
240 control = apt_pkg.ParseSection(extracts)
242 print formatted_text("can't parse control info")
248 control_keys = control.keys()
250 if control.has_key("Depends"):
251 depends_str = control.Find("Depends")
252 # create list of dependancy lists
253 depends = split_depends(depends_str)
255 if control.has_key("Recommends"):
256 recommends_str = control.Find("Recommends")
257 recommends = split_depends(recommends_str)
259 if control.has_key("Section"):
260 section_str = control.Find("Section")
262 c_match = re_contrib.search(section_str)
263 nf_match = re_nonfree.search(section_str)
266 section = colour_output(section_str, 'contrib')
269 section = colour_output(section_str, 'nonfree')
272 section = colour_output(section_str, 'main')
273 if control.has_key("Architecture"):
274 arch_str = control.Find("Architecture")
275 arch = colour_output(arch_str, 'arch')
277 if control.has_key("Maintainer"):
278 maintainer = control.Find("Maintainer")
279 localhost = re_localhost.search(maintainer)
282 maintainer = colour_output(maintainer, 'maintainer')
284 maintainer = escape_if_needed(maintainer)
286 return (control, control_keys, section, depends, recommends, arch, maintainer)
288 def read_changes_or_dsc (suite, filename, session = None):
291 dsc_file = utils.open_file(filename)
293 dsc = utils.parse_changes(filename, dsc_file=1)
295 return formatted_text("can't parse .dsc control info")
298 filecontents = strip_pgp_signature(filename)
300 for l in filecontents.split('\n'):
301 m = re.match(r'([-a-zA-Z0-9]*):', l)
303 keysinorder.append(m.group(1))
306 if k in ("build-depends","build-depends-indep"):
307 dsc[k] = create_depends_string(suite, split_depends(dsc[k]), session)
308 elif k == "architecture":
309 if (dsc["architecture"] != "any"):
310 dsc['architecture'] = colour_output(dsc["architecture"], 'arch')
311 elif k == "distribution":
312 if dsc["distribution"] not in ('unstable', 'experimental'):
313 dsc['distribution'] = colour_output(dsc["distribution"], 'distro')
314 elif k in ("files","changes","description"):
316 dsc[k] = formatted_text(dsc[k], strip=True)
318 dsc[k] = ('\n'+'\n'.join(map(lambda x: ' '+x, dsc[k].split('\n')))).rstrip()
320 dsc[k] = escape_if_needed(dsc[k])
322 keysinorder = filter(lambda x: not x.lower().startswith('checksums-'), keysinorder)
324 filecontents = '\n'.join(map(lambda x: format_field(x,dsc[x.lower()]), keysinorder))+'\n'
327 def get_provides(suite):
329 session = DBConn().session()
330 query = '''SELECT DISTINCT value
331 FROM binaries_metadata m
332 JOIN bin_associations b
337 WHERE key = 'Provides' )
341 WHERE suite_name = '%(suite)s'
342 OR codename = '%(suite)s')''' % \
344 for p in session.execute(query):
346 for i in e.split(','):
347 provides.add(i.strip())
351 def create_depends_string (suite, depends_tree, session = None):
353 if suite == 'experimental':
354 suite_list = ['experimental','unstable']
360 for l in depends_tree:
361 if (comma_count >= 2):
367 # doesn't do version lookup yet.
369 component = get_component_by_package_suite(d['name'], suite_list, \
371 if component is not None:
373 if d['version'] != '' :
374 adepends += " (%s)" % (d['version'])
376 if component == "contrib":
377 result += colour_output(adepends, "contrib")
378 elif component == "non-free":
379 result += colour_output(adepends, "nonfree")
381 result += colour_output(adepends, "main")
384 if d['version'] != '' :
385 adepends += " (%s)" % (d['version'])
387 provides = get_provides(suite)
388 if d['name'] in provides:
389 result += colour_output(adepends, "provides")
391 result += colour_output(adepends, "bold")
396 def output_package_relations ():
398 Output the package relations, if there is more than one package checked in this run.
401 if len(package_relations) < 2:
402 # Only list something if we have more than one binary to compare
403 package_relations.clear()
407 for package in package_relations:
408 for relation in package_relations[package]:
409 to_print += "%-15s: (%s) %s\n" % (package, relation, package_relations[package][relation])
411 package_relations.clear()
412 return foldable_output("Package relations", "relations", to_print)
414 def output_deb_info(suite, filename, packagename, session = None):
415 (control, control_keys, section, depends, recommends, arch, maintainer) = read_control(filename)
418 return formatted_text("no control info")
420 if not package_relations.has_key(packagename):
421 package_relations[packagename] = {}
422 for key in control_keys :
424 field_value = create_depends_string(suite, depends, session)
425 package_relations[packagename][key] = field_value
426 elif key == 'Recommends':
427 field_value = create_depends_string(suite, recommends, session)
428 package_relations[packagename][key] = field_value
429 elif key == 'Section':
430 field_value = section
431 elif key == 'Architecture':
433 elif key == 'Maintainer':
434 field_value = maintainer
435 elif key == 'Description':
437 field_value = formatted_text(control.Find(key), strip=True)
439 desc = control.Find(key)
440 desc = re_newlinespace.sub('\n ', desc)
441 field_value = escape_if_needed(desc)
443 field_value = escape_if_needed(control.Find(key))
444 to_print += " "+format_field(key,field_value)+'\n'
447 def do_command (command, filename, escaped=0):
448 o = os.popen("%s %s" % (command, filename))
450 return escaped_text(o.read())
452 return formatted_text(o.read())
454 def do_lintian (filename):
456 return do_command("lintian --show-overrides --color html", filename, 1)
458 return do_command("lintian --show-overrides --color always", filename, 1)
460 def get_copyright (deb_filename):
463 package = re_package.sub(r'\1', deb_filename)
464 o = os.popen("dpkg-deb -c %s | egrep 'usr(/share)?/doc/[^/]*/copyright' | awk '{print $6}' | head -n 1" % (deb_filename))
465 cright = o.read()[:-1]
468 return formatted_text("WARNING: No copyright found, please check package manually.")
470 doc_directory = re_doc_directory.sub(r'\1', cright)
471 if package != doc_directory:
472 return formatted_text("WARNING: wrong doc directory (expected %s, got %s)." % (package, doc_directory))
474 o = os.popen("dpkg-deb --fsys-tarfile %s | tar xvOf - %s 2>/dev/null" % (deb_filename, cright))
476 copyrightmd5 = md5.md5(cright).hexdigest()
479 if printed.copyrights.has_key(copyrightmd5) and printed.copyrights[copyrightmd5] != "%s (%s)" % (package, deb_filename):
480 res += formatted_text( "NOTE: Copyright is the same as %s.\n\n" % \
481 (printed.copyrights[copyrightmd5]))
483 printed.copyrights[copyrightmd5] = "%s (%s)" % (package, deb_filename)
484 return res+formatted_text(cright)
486 def get_readme_source (dsc_filename):
487 tempdir = utils.temp_dirname()
490 cmd = "dpkg-source --no-check --no-copy -x %s %s" % (dsc_filename, tempdir)
491 (result, output) = commands.getstatusoutput(cmd)
493 res = "How is education supposed to make me feel smarter? Besides, every time I learn something new, it pushes some\n old stuff out of my brain. Remember when I took that home winemaking course, and I forgot how to drive?\n"
494 res += "Error, couldn't extract source, WTF?\n"
495 res += "'dpkg-source -x' failed. return code: %s.\n\n" % (result)
499 path = os.path.join(tempdir, 'debian/README.source')
501 if os.path.exists(path):
502 res += do_command("cat", path)
504 res += "No README.source in this package\n\n"
507 shutil.rmtree(tempdir)
509 if errno.errorcode[e.errno] != 'EACCES':
510 res += "%s: couldn't remove tmp dir %s for source tree." % (dsc_filename, tempdir)
514 def check_dsc (suite, dsc_filename, session = None):
515 (dsc) = read_changes_or_dsc(suite, dsc_filename, session)
516 return foldable_output(dsc_filename, "dsc", dsc, norow=True) + \
518 foldable_output("lintian check for %s" % dsc_filename,
519 "source-lintian", do_lintian(dsc_filename)) + \
521 foldable_output("README.source for %s" % dsc_filename,
522 "source-readmesource", get_readme_source(dsc_filename))
524 def check_deb (suite, deb_filename, session = None):
525 filename = os.path.basename(deb_filename)
526 packagename = filename.split('_')[0]
528 if filename.endswith(".udeb"):
533 result = foldable_output("control file for %s" % (filename), "binary-%s-control"%packagename,
534 output_deb_info(suite, deb_filename, packagename, session), norow=True) + "\n"
537 result += foldable_output("skipping lintian check for udeb",
538 "binary-%s-lintian"%packagename, "") + "\n"
540 result += foldable_output("lintian check for %s" % (filename),
541 "binary-%s-lintian"%packagename, do_lintian(deb_filename)) + "\n"
543 result += foldable_output("contents of %s" % (filename), "binary-%s-contents"%packagename,
544 do_command("dpkg -c", deb_filename)) + "\n"
547 result += foldable_output("skipping copyright for udeb",
548 "binary-%s-copyright"%packagename, "") + "\n"
550 result += foldable_output("copyright of %s" % (filename),
551 "binary-%s-copyright"%packagename, get_copyright(deb_filename)) + "\n"
553 result += foldable_output("file listing of %s" % (filename),
554 "binary-%s-file-listing"%packagename, do_command("ls -l", deb_filename))
558 # Read a file, strip the signature and return the modified contents as
560 def strip_pgp_signature (filename):
561 inputfile = utils.open_file (filename)
565 for line in inputfile.readlines():
573 if line.startswith("-----BEGIN PGP SIGNED MESSAGE"):
576 if line.startswith("-----BEGIN PGP SIGNATURE"):
579 if line.startswith("-----END PGP SIGNATURE"):
586 def display_changes(suite, changes_filename):
588 changes = read_changes_or_dsc(suite, changes_filename)
589 printed.copyrights = {}
590 return foldable_output(changes_filename, "changes", changes, norow=True)
592 def check_changes (changes_filename):
594 changes = utils.parse_changes (changes_filename)
595 except ChangesUnicodeError:
596 utils.warn("Encoding problem with changes file %s" % (changes_filename))
597 print display_changes(changes['distribution'], changes_filename)
599 files = utils.build_file_list(changes)
600 for f in files.keys():
601 if f.endswith(".deb") or f.endswith(".udeb"):
602 print check_deb(changes['distribution'], f)
603 if f.endswith(".dsc"):
604 print check_dsc(changes['distribution'], f)
608 global Cnf, db_files, waste, excluded
610 # Cnf = utils.get_conf()
612 Arguments = [('h',"help","Examine-Package::Options::Help"),
613 ('H',"html-output","Examine-Package::Options::Html-Output"),
615 for i in [ "Help", "Html-Output", "partial-html" ]:
616 if not Cnf.has_key("Examine-Package::Options::%s" % (i)):
617 Cnf["Examine-Package::Options::%s" % (i)] = ""
619 args = apt_pkg.ParseCommandLine(Cnf,Arguments,sys.argv)
620 Options = Cnf.SubTree("Examine-Package::Options")
625 if Options["Html-Output"]:
629 stdout_fd = sys.stdout
633 if not Options["Html-Output"]:
634 # Pipe output for each argument through less
635 less_fd = os.popen("less -R -", 'w', 0)
636 # -R added to display raw control chars for colour
639 if f.endswith(".changes"):
641 elif f.endswith(".deb") or f.endswith(".udeb"):
642 # default to unstable when we don't have a .changes file
643 # perhaps this should be a command line option?
644 print check_deb('unstable', f)
645 elif f.endswith(".dsc"):
646 print check_dsc('unstable', f)
648 utils.fubar("Unrecognised file type: '%s'." % (f))
650 print output_package_relations()
651 if not Options["Html-Output"]:
652 # Reset stdout here so future less invocations aren't FUBAR
654 sys.stdout = stdout_fd
656 if errno.errorcode[e.errno] == 'EPIPE':
657 utils.warn("[examine-package] Caught EPIPE; skipping.")
661 except KeyboardInterrupt:
662 utils.warn("[examine-package] Caught C-c; skipping.")
665 #######################################################################################
667 if __name__ == '__main__':