# Copyright 1998-2007 Gentoo Foundation # Distributed under the terms of the GNU General Public License v2 # $Id$ __all__ = ["PreservedLibsRegistry", "LinkageMap", "vardbapi", "vartree", "dblink"] + \ ["write_contents", "tar_contents"] from portage.checksum import perform_md5 from portage.const import CACHE_PATH, CONFIG_MEMORY_FILE, \ PORTAGE_PACKAGE_ATOM, PRIVATE_PATH, VDB_PATH from portage.data import portage_gid, portage_uid, secpass from portage.dbapi import dbapi from portage.dep import use_reduce, paren_reduce, isvalidatom, \ isjustname, dep_getkey, match_from_list from portage.exception import CommandNotFound, \ InvalidData, InvalidPackageName, \ FileNotFound, PermissionDenied, UnsupportedAPIException from portage.locks import lockdir, unlockdir from portage.output import bold, red, green from portage.update import fixdbentries from portage.util import apply_secpass_permissions, ConfigProtect, ensure_dirs, \ writemsg, writemsg_level, write_atomic, atomic_ofstream, writedict, \ grabfile, grabdict, normalize_path, new_protect_filename, getlibpaths from portage.versions import pkgsplit, catpkgsplit, catsplit, best, pkgcmp from portage import listdir, dep_expand, digraph, flatten, key_expand, \ doebuild_environment, doebuild, env_update, prepare_build_dirs, \ abssymlink, movefile, _movefile, bsd_chflags, cpv_getkey from portage.elog import elog_process from portage.elog.filtering import filter_mergephases, filter_unmergephases from portage.cache.mappings import slot_dict_class import os, re, shutil, stat, errno, copy, subprocess import logging import shlex from itertools import izip try: import cPickle as pickle except ImportError: import pickle class PreservedLibsRegistry(object): """ This class handles the tracking of preserved library objects """ def __init__(self, filename, autocommit=True): """ @param filename: absolute path for saving the preserved libs records @type filename: String @param autocommit: determines if the file is written after every update @type autocommit: Boolean """ self._filename = filename self._autocommit = autocommit self.load() self.pruneNonExisting() def load(self): """ Reload the registry data from file """ try: self._data = pickle.load(open(self._filename, "r")) except (EOFError, IOError), e: if isinstance(e, EOFError) or e.errno == errno.ENOENT: self._data = {} elif e.errno == PermissionDenied.errno: raise PermissionDenied(self._filename) else: raise e self._data_orig = self._data.copy() def store(self): """ Store the registry data to file. No need to call this if autocommit was enabled. """ if os.environ.get("SANDBOX_ON") == "1" or \ self._data == self._data_orig: return try: f = atomic_ofstream(self._filename) pickle.dump(self._data, f) f.close() except EnvironmentError, e: if e.errno != PermissionDenied.errno: writemsg("!!! %s %s\n" % (e, self._filename), noiselevel=-1) else: self._data_orig = self._data.copy() def register(self, cpv, slot, counter, paths): """ Register new objects in the registry. If there is a record with the same packagename (internally derived from cpv) and slot it is overwritten with the new data. @param cpv: package instance that owns the objects @type cpv: CPV (as String) @param slot: the value of SLOT of the given package instance @type slot: String @param counter: vdb counter value for the package instace @type counter: Integer @param paths: absolute paths of objects that got preserved during an update @type paths: List """ cp = "/".join(catpkgsplit(cpv)[:2]) cps = cp+":"+slot if len(paths) == 0 and cps in self._data \ and self._data[cps][0] == cpv and int(self._data[cps][1]) == int(counter): del self._data[cps] elif len(paths) > 0: self._data[cps] = (cpv, counter, paths) if self._autocommit: self.store() def unregister(self, cpv, slot, counter): """ Remove a previous registration of preserved objects for the given package. @param cpv: package instance whose records should be removed @type cpv: CPV (as String) @param slot: the value of SLOT of the given package instance @type slot: String """ self.register(cpv, slot, counter, []) def pruneNonExisting(self): """ Remove all records for objects that no longer exist on the filesystem. """ for cps in self._data.keys(): cpv, counter, paths = self._data[cps] paths = [f for f in paths if os.path.exists(f)] if len(paths) > 0: self._data[cps] = (cpv, counter, paths) else: del self._data[cps] if self._autocommit: self.store() def hasEntries(self): """ Check if this registry contains any records. """ return len(self._data) > 0 def getPreservedLibs(self): """ Return a mapping of packages->preserved objects. @returns mapping of package instances to preserved objects @rtype Dict cpv->list-of-paths """ rValue = {} for cps in self._data: rValue[self._data[cps][0]] = self._data[cps][2] return rValue class LinkageMap(object): """Models dynamic linker dependencies.""" _needed_aux_key = "NEEDED.ELF.2" _soname_map_class = slot_dict_class( ("consumers", "providers"), prefix="") def __init__(self, vardbapi): self._dbapi = vardbapi self._root = self._dbapi.root self._libs = {} self._obj_properties = {} self._obj_key_cache = {} self._defpath = set() self._path_key_cache = {} def _clear_cache(self): self._libs.clear() self._obj_properties.clear() self._obj_key_cache.clear() self._defpath.clear() self._path_key_cache.clear() def _path_key(self, path): key = self._path_key_cache.get(path) if key is None: key = self._ObjectKey(path, self._root) self._path_key_cache[path] = key return key def _obj_key(self, path): key = self._obj_key_cache.get(path) if key is None: key = self._ObjectKey(path, self._root) self._obj_key_cache[path] = key return key class _ObjectKey(object): """Helper class used as _obj_properties keys for objects.""" __slots__ = ("__weakref__", "_key") def __init__(self, obj, root): """ This takes a path to an object. @param object: path to a file @type object: string (example: '/usr/bin/bar') """ self._key = self._generate_object_key(obj, root) def __hash__(self): return hash(self._key) def __eq__(self, other): return self._key == other._key def _generate_object_key(self, obj, root): """ Generate object key for a given object. @param object: path to a file @type object: string (example: '/usr/bin/bar') @rtype: 2-tuple of types (long, int) if object exists. string if object does not exist. @return: 1. 2-tuple of object's inode and device from a stat call, if object exists. 2. realpath of object if object does not exist. """ abs_path = os.path.join(root, obj.lstrip(os.sep)) try: object_stat = os.stat(abs_path) except OSError: # Use the realpath as the key if the file does not exists on the # filesystem. return os.path.realpath(abs_path) # Return a tuple of the device and inode. return (object_stat.st_dev, object_stat.st_ino) def file_exists(self): """ Determine if the file for this key exists on the filesystem. @rtype: Boolean @return: 1. True if the file exists. 2. False if the file does not exist or is a broken symlink. """ return isinstance(self._key, tuple) class _LibGraphNode(_ObjectKey): __slots__ = ("alt_paths",) def __init__(self, obj, root): LinkageMap._ObjectKey.__init__(self, obj, root) self.alt_paths = set() def __str__(self): return str(sorted(self.alt_paths)) def rebuild(self, exclude_pkgs=None, include_file=None): """ Raises CommandNotFound if there are preserved libs and the scanelf binary is not available. """ root = self._root root_len = len(root) - 1 self._clear_cache() self._defpath.update(getlibpaths(self._root)) libs = self._libs obj_key_cache = self._obj_key_cache obj_properties = self._obj_properties lines = [] # Data from include_file is processed first so that it # overrides any data from previously installed files. if include_file is not None: lines += grabfile(include_file) aux_keys = [self._needed_aux_key] for cpv in self._dbapi.cpv_all(): if exclude_pkgs is not None and cpv in exclude_pkgs: continue lines += self._dbapi.aux_get(cpv, aux_keys)[0].split('\n') # Cache NEEDED.* files avoid doing excessive IO for every rebuild. self._dbapi.flush_cache() # have to call scanelf for preserved libs here as they aren't # registered in NEEDED.ELF.2 files if self._dbapi.plib_registry and self._dbapi.plib_registry.getPreservedLibs(): args = ["/usr/bin/scanelf", "-qF", "%a;%F;%S;%r;%n"] for items in self._dbapi.plib_registry.getPreservedLibs().values(): args.extend(os.path.join(root, x.lstrip("." + os.sep)) \ for x in items) try: proc = subprocess.Popen(args, stdout=subprocess.PIPE) except EnvironmentError, e: if e.errno != errno.ENOENT: raise raise CommandNotFound(args[0]) else: for l in proc.stdout: l = l[3:].rstrip("\n") if not l: continue fields = l.split(";") if len(fields) < 5: writemsg_level("\nWrong number of fields " + \ "returned from scanelf: %s\n\n" % (l,), level=logging.ERROR, noiselevel=-1) continue fields[1] = fields[1][root_len:] lines.append(";".join(fields)) proc.wait() for l in lines: l = l.rstrip("\n") if not l: continue fields = l.split(";") if len(fields) < 5: writemsg_level("\nWrong number of fields " + \ "in %s: %s\n\n" % (self._needed_aux_key, l), level=logging.ERROR, noiselevel=-1) continue arch = fields[0] obj = fields[1] soname = fields[2] path = set([normalize_path(x) \ for x in filter(None, fields[3].replace( "${ORIGIN}", os.path.dirname(obj)).replace( "$ORIGIN", os.path.dirname(obj)).split(":"))]) needed = filter(None, fields[4].split(",")) obj_key = self._obj_key(obj) indexed = True myprops = obj_properties.get(obj_key) if myprops is None: indexed = False myprops = (arch, needed, path, soname, set()) obj_properties[obj_key] = myprops # All object paths are added into the obj_properties tuple. myprops[4].add(obj) # Don't index the same file more that once since only one # set of data can be correct and therefore mixing data # may corrupt the index (include_file overrides previously # installed). if indexed: continue arch_map = libs.get(arch) if arch_map is None: arch_map = {} libs[arch] = arch_map if soname: soname_map = arch_map.get(soname) if soname_map is None: soname_map = self._soname_map_class( providers=set(), consumers=set()) arch_map[soname] = soname_map soname_map.providers.add(obj_key) for needed_soname in needed: soname_map = arch_map.get(needed_soname) if soname_map is None: soname_map = self._soname_map_class( providers=set(), consumers=set()) arch_map[needed_soname] = soname_map soname_map.consumers.add(obj_key) def listBrokenBinaries(self, debug=False): """ Find binaries and their needed sonames, which have no providers. @param debug: Boolean to enable debug output @type debug: Boolean @rtype: dict (example: {'/usr/bin/foo': set(['libbar.so'])}) @return: The return value is an object -> set-of-sonames mapping, where object is a broken binary and the set consists of sonames needed by object that have no corresponding libraries to fulfill the dependency. """ class _LibraryCache(object): """ Caches properties associated with paths. The purpose of this class is to prevent multiple instances of _ObjectKey for the same paths. """ def __init__(cache_self): cache_self.cache = {} def get(cache_self, obj): """ Caches and returns properties associated with an object. @param obj: absolute path (can be symlink) @type obj: string (example: '/usr/lib/libfoo.so') @rtype: 4-tuple with types (string or None, string or None, 2-tuple, Boolean) @return: 4-tuple with the following components: 1. arch as a string or None if it does not exist, 2. soname as a string or None if it does not exist, 3. obj_key as 2-tuple, 4. Boolean representing whether the object exists. (example: ('libfoo.so.1', (123L, 456L), True)) """ if obj in cache_self.cache: return cache_self.cache[obj] else: obj_key = self._obj_key(obj) # Check that the library exists on the filesystem. if obj_key.file_exists(): # Get the arch and soname from LinkageMap._obj_properties if # it exists. Otherwise, None. arch, _, _, soname, _ = \ self._obj_properties.get(obj_key, (None,)*5) return cache_self.cache.setdefault(obj, \ (arch, soname, obj_key, True)) else: return cache_self.cache.setdefault(obj, \ (None, None, obj_key, False)) rValue = {} cache = _LibraryCache() providers = self.listProviders() # Iterate over all obj_keys and their providers. for obj_key, sonames in providers.items(): arch, _, path, _, objs = self._obj_properties[obj_key] path = path.union(self._defpath) # Iterate over each needed soname and the set of library paths that # fulfill the soname to determine if the dependency is broken. for soname, libraries in sonames.items(): # validLibraries is used to store libraries, which satisfy soname, # so if no valid libraries are found, the soname is not satisfied # for obj_key. If unsatisfied, objects associated with obj_key # must be emerged. validLibraries = set() # It could be the case that the library to satisfy the soname is # not in the obj's runpath, but a symlink to the library is (eg # libnvidia-tls.so.1 in nvidia-drivers). Also, since LinkageMap # does not catalog symlinks, broken or missing symlinks may go # unnoticed. As a result of these cases, check that a file with # the same name as the soname exists in obj's runpath. # XXX If we catalog symlinks in LinkageMap, this could be improved. for directory in path: cachedArch, cachedSoname, cachedKey, cachedExists = \ cache.get(os.path.join(directory, soname)) # Check that this library provides the needed soname. Doing # this, however, will cause consumers of libraries missing # sonames to be unnecessarily emerged. (eg libmix.so) if cachedSoname == soname and cachedArch == arch: validLibraries.add(cachedKey) if debug and cachedKey not in \ set(map(self._obj_key_cache.get, libraries)): # XXX This is most often due to soname symlinks not in # a library's directory. We could catalog symlinks in # LinkageMap to avoid checking for this edge case here. print "Found provider outside of findProviders:", \ os.path.join(directory, soname), "->", \ self._obj_properties[cachedKey][4], libraries # A valid library has been found, so there is no need to # continue. break if debug and cachedArch == arch and \ cachedKey in self._obj_properties: print "Broken symlink or missing/bad soname:", \ os.path.join(directory, soname), '->', \ self._obj_properties[cachedKey], "with soname", \ cachedSoname, "but expecting", soname # This conditional checks if there are no libraries to satisfy the # soname (empty set). if not validLibraries: for obj in objs: rValue.setdefault(obj, set()).add(soname) # If no valid libraries have been found by this point, then # there are no files named with the soname within obj's runpath, # but if there are libraries (from the providers mapping), it is # likely that soname symlinks or the actual libraries are # missing or broken. Thus those libraries are added to rValue # in order to emerge corrupt library packages. for lib in libraries: rValue.setdefault(lib, set()).add(soname) if debug: if not os.path.isfile(lib): print "Missing library:", lib else: print "Possibly missing symlink:", \ os.path.join(os.path.dirname(lib), soname) return rValue def listProviders(self): """ Find the providers for all object keys in LinkageMap. @rtype: dict (example: {(123L, 456L): {'libbar.so': set(['/lib/libbar.so.1.5'])}}) @return: The return value is an object key -> providers mapping, where providers is a mapping of soname -> set-of-library-paths returned from the findProviders method. """ rValue = {} if not self._libs: self.rebuild() # Iterate over all object keys within LinkageMap. for obj_key in self._obj_properties: rValue.setdefault(obj_key, self.findProviders(obj_key)) return rValue def isMasterLink(self, obj): """ Determine whether an object is a master link. @param obj: absolute path to an object @type obj: string (example: '/usr/bin/foo') @rtype: Boolean @return: 1. True if obj is a master link 2. False if obj is not a master link """ basename = os.path.basename(obj) obj_key = self._obj_key(obj) if obj_key not in self._obj_properties: raise KeyError("%s (%s) not in object list" % (obj_key, obj)) soname = self._obj_properties[obj_key][3] return (len(basename) < len(soname)) def listLibraryObjects(self): """ Return a list of library objects. Known limitation: library objects lacking an soname are not included. @rtype: list of strings @return: list of paths to all providers """ rValue = [] if not self._libs: self.rebuild() for arch_map in self._libs.itervalues(): for soname_map in arch_map.itervalues(): for obj_key in soname_map.providers: rValue.extend(self._obj_properties[obj_key][4]) return rValue def getSoname(self, obj): """ Return the soname associated with an object. @param obj: absolute path to an object @type obj: string (example: '/usr/bin/bar') @rtype: string @return: soname as a string """ if not self._libs: self.rebuild() if isinstance(obj, self._ObjectKey): obj_key = obj if obj_key not in self._obj_properties: raise KeyError("%s not in object list" % obj_key) return self._obj_properties[obj_key][3] if obj not in self._obj_key_cache: raise KeyError("%s not in object list" % obj) return self._obj_properties[self._obj_key_cache[obj]][3] def findProviders(self, obj): """ Find providers for an object or object key. This method may be called with a key from _obj_properties. In some cases, not all valid libraries are returned. This may occur when an soname symlink referencing a library is in an object's runpath while the actual library is not. We should consider cataloging symlinks within LinkageMap as this would avoid those cases and would be a better model of library dependencies (since the dynamic linker actually searches for files named with the soname in the runpaths). @param obj: absolute path to an object or a key from _obj_properties @type obj: string (example: '/usr/bin/bar') or _ObjectKey @rtype: dict (example: {'libbar.so': set(['/lib/libbar.so.1.5'])}) @return: The return value is a soname -> set-of-library-paths, where set-of-library-paths satisfy soname. """ rValue = {} if not self._libs: self.rebuild() # Determine the obj_key from the arguments. if isinstance(obj, self._ObjectKey): obj_key = obj if obj_key not in self._obj_properties: raise KeyError("%s not in object list" % obj_key) else: obj_key = self._obj_key(obj) if obj_key not in self._obj_properties: raise KeyError("%s (%s) not in object list" % (obj_key, obj)) arch, needed, path, _, _ = self._obj_properties[obj_key] path_keys = set(self._path_key(x) for x in path.union(self._defpath)) for soname in needed: rValue[soname] = set() if arch not in self._libs or soname not in self._libs[arch]: continue # For each potential provider of the soname, add it to rValue if it # resides in the obj's runpath. for provider_key in self._libs[arch][soname].providers: providers = self._obj_properties[provider_key][4] for provider in providers: if self._path_key(os.path.dirname(provider)) in path_keys: rValue[soname].add(provider) return rValue def findConsumers(self, obj): """ Find consumers of an object or object key. This method may be called with a key from _obj_properties. If this method is going to be called with an object key, to avoid not catching shadowed libraries, do not pass new _ObjectKey instances to this method. Instead pass the obj as a string. In some cases, not all consumers are returned. This may occur when an soname symlink referencing a library is in an object's runpath while the actual library is not. For example, this problem is noticeable for binutils since it's libraries are added to the path via symlinks that are gemerated in the /usr/$CHOST/lib/ directory by binutils-config. Failure to recognize consumers of these symlinks makes preserve-libs fail to preserve binutils libs that are needed by these unrecognized consumers. @param obj: absolute path to an object or a key from _obj_properties @type obj: string (example: '/usr/bin/bar') or _ObjectKey @rtype: set of strings (example: set(['/bin/foo', '/usr/bin/bar'])) @return: The return value is a soname -> set-of-library-paths, where set-of-library-paths satisfy soname. """ rValue = set() if not self._libs: self.rebuild() # Determine the obj_key and the set of objects matching the arguments. if isinstance(obj, self._ObjectKey): obj_key = obj if obj_key not in self._obj_properties: raise KeyError("%s not in object list" % obj_key) objs = self._obj_properties[obj_key][4] else: objs = set([obj]) obj_key = self._obj_key(obj) if obj_key not in self._obj_properties: raise KeyError("%s (%s) not in object list" % (obj_key, obj)) # If there is another version of this lib with the # same soname and the master link points to that # other version, this lib will be shadowed and won't # have any consumers. if not isinstance(obj, self._ObjectKey): soname = self._obj_properties[obj_key][3] master_link = os.path.join(self._root, os.path.dirname(obj).lstrip(os.path.sep), soname) try: master_st = os.stat(master_link) obj_st = os.stat(obj) except OSError: pass else: if (obj_st.st_dev, obj_st.st_ino) != \ (master_st.st_dev, master_st.st_ino): return set() # Determine the directory(ies) from the set of objects. objs_dir_keys = set(self._path_key(os.path.dirname(x)) for x in objs) defpath_keys = set(self._path_key(x) for x in self._defpath) arch, _, _, soname, _ = self._obj_properties[obj_key] if arch in self._libs and soname in self._libs[arch]: # For each potential consumer, add it to rValue if an object from the # arguments resides in the consumer's runpath. for consumer_key in self._libs[arch][soname].consumers: _, _, path, _, consumer_objs = \ self._obj_properties[consumer_key] path_keys = defpath_keys.union(self._path_key(x) for x in path) if objs_dir_keys.intersection(path_keys): rValue.update(consumer_objs) return rValue class vardbapi(dbapi): _excluded_dirs = ["CVS", "lost+found"] _excluded_dirs = [re.escape(x) for x in _excluded_dirs] _excluded_dirs = re.compile(r'^(\..*|-MERGING-.*|' + \ "|".join(_excluded_dirs) + r')$') _aux_cache_version = "1" _owners_cache_version = "1" # Number of uncached packages to trigger cache update, since # it's wasteful to update it for every vdb change. _aux_cache_threshold = 5 _aux_cache_keys_re = re.compile(r'^NEEDED\..*$') _aux_multi_line_re = re.compile(r'^(CONTENTS|NEEDED\..*)$') def __init__(self, root, categories=None, settings=None, vartree=None): """ The categories parameter is unused since the dbapi class now has a categories property that is generated from the available packages. """ self.root = root[:] #cache for category directory mtimes self.mtdircache = {} #cache for dependency checks self.matchcache = {} #cache for cp_list results self.cpcache = {} self.blockers = None if settings is None: from portage import settings self.settings = settings if vartree is None: from portage import db vartree = db[root]["vartree"] self.vartree = vartree self._aux_cache_keys = set( ["CHOST", "COUNTER", "DEPEND", "DESCRIPTION", "EAPI", "HOMEPAGE", "IUSE", "KEYWORDS", "LICENSE", "PDEPEND", "PROPERTIES", "PROVIDE", "RDEPEND", "repository", "RESTRICT" , "SLOT", "USE"]) self._aux_cache_obj = None self._aux_cache_filename = os.path.join(self.root, CACHE_PATH.lstrip(os.path.sep), "vdb_metadata.pickle") self._counter_path = os.path.join(root, CACHE_PATH.lstrip(os.path.sep), "counter") try: self.plib_registry = PreservedLibsRegistry( os.path.join(self.root, PRIVATE_PATH, "preserved_libs_registry")) except PermissionDenied: # apparently this user isn't allowed to access PRIVATE_PATH self.plib_registry = None self.linkmap = LinkageMap(self) self._owners = self._owners_db(self) def getpath(self, mykey, filename=None): rValue = os.path.join(self.root, VDB_PATH, mykey) if filename != None: rValue = os.path.join(rValue, filename) return rValue def cpv_exists(self, mykey): "Tells us whether an actual ebuild exists on disk (no masking)" return os.path.exists(self.getpath(mykey)) def cpv_counter(self, mycpv): "This method will grab the COUNTER. Returns a counter value." try: return long(self.aux_get(mycpv, ["COUNTER"])[0]) except (KeyError, ValueError): pass writemsg_level(("portage: COUNTER for %s was corrupted; " + \ "resetting to value of 0\n") % (mycpv,), level=logging.ERROR, noiselevel=-1) return 0 def cpv_inject(self, mycpv): "injects a real package into our on-disk database; assumes mycpv is valid and doesn't already exist" os.makedirs(self.getpath(mycpv)) counter = self.counter_tick(self.root, mycpv=mycpv) # write local package counter so that emerge clean does the right thing write_atomic(self.getpath(mycpv, filename="COUNTER"), str(counter)) def isInjected(self, mycpv): if self.cpv_exists(mycpv): if os.path.exists(self.getpath(mycpv, filename="INJECTED")): return True if not os.path.exists(self.getpath(mycpv, filename="CONTENTS")): return True return False def move_ent(self, mylist): origcp = mylist[1] newcp = mylist[2] # sanity check for cp in [origcp, newcp]: if not (isvalidatom(cp) and isjustname(cp)): raise InvalidPackageName(cp) origmatches = self.match(origcp, use_cache=0) moves = 0 if not origmatches: return moves for mycpv in origmatches: mycpsplit = catpkgsplit(mycpv) mynewcpv = newcp + "-" + mycpsplit[2] mynewcat = newcp.split("/")[0] if mycpsplit[3] != "r0": mynewcpv += "-" + mycpsplit[3] mycpsplit_new = catpkgsplit(mynewcpv) origpath = self.getpath(mycpv) if not os.path.exists(origpath): continue moves += 1 if not os.path.exists(self.getpath(mynewcat)): #create the directory os.makedirs(self.getpath(mynewcat)) newpath = self.getpath(mynewcpv) if os.path.exists(newpath): #dest already exists; keep this puppy where it is. continue _movefile(origpath, newpath, mysettings=self.settings) # We need to rename the ebuild now. old_pf = catsplit(mycpv)[1] new_pf = catsplit(mynewcpv)[1] if new_pf != old_pf: try: os.rename(os.path.join(newpath, old_pf + ".ebuild"), os.path.join(newpath, new_pf + ".ebuild")) except EnvironmentError, e: if e.errno != errno.ENOENT: raise del e write_atomic(os.path.join(newpath, "PF"), new_pf+"\n") write_atomic(os.path.join(newpath, "CATEGORY"), mynewcat+"\n") fixdbentries([mylist], newpath) return moves def cp_list(self, mycp, use_cache=1): mysplit=catsplit(mycp) if mysplit[0] == '*': mysplit[0] = mysplit[0][1:] try: mystat = os.stat(self.getpath(mysplit[0]))[stat.ST_MTIME] except OSError: mystat = 0 if use_cache and mycp in self.cpcache: cpc = self.cpcache[mycp] if cpc[0] == mystat: return cpc[1][:] cat_dir = self.getpath(mysplit[0]) try: dir_list = os.listdir(cat_dir) except EnvironmentError, e: if e.errno == PermissionDenied.errno: raise PermissionDenied(cat_dir) del e dir_list = [] returnme = [] for x in dir_list: if self._excluded_dirs.match(x) is not None: continue ps = pkgsplit(x) if not ps: self.invalidentry(os.path.join(self.getpath(mysplit[0]), x)) continue if len(mysplit) > 1: if ps[0] == mysplit[1]: returnme.append(mysplit[0]+"/"+x) self._cpv_sort_ascending(returnme) if use_cache: self.cpcache[mycp] = [mystat, returnme[:]] elif mycp in self.cpcache: del self.cpcache[mycp] return returnme def cpv_all(self, use_cache=1): """ Set use_cache=0 to bypass the portage.cachedir() cache in cases when the accuracy of mtime staleness checks should not be trusted (generally this is only necessary in critical sections that involve merge or unmerge of packages). """ returnme = [] basepath = os.path.join(self.root, VDB_PATH) + os.path.sep if use_cache: from portage import listdir else: def listdir(p, **kwargs): try: return [x for x in os.listdir(p) \ if os.path.isdir(os.path.join(p, x))] except EnvironmentError, e: if e.errno == PermissionDenied.errno: raise PermissionDenied(p) del e return [] for x in listdir(basepath, EmptyOnError=1, ignorecvs=1, dirsonly=1): if self._excluded_dirs.match(x) is not None: continue if not self._category_re.match(x): continue for y in listdir(basepath + x, EmptyOnError=1, dirsonly=1): if self._excluded_dirs.match(y) is not None: continue subpath = x + "/" + y # -MERGING- should never be a cpv, nor should files. try: if catpkgsplit(subpath) is None: self.invalidentry(self.getpath(subpath)) continue except InvalidData: self.invalidentry(self.getpath(subpath)) continue returnme.append(subpath) return returnme def cp_all(self, use_cache=1): mylist = self.cpv_all(use_cache=use_cache) d={} for y in mylist: if y[0] == '*': y = y[1:] try: mysplit = catpkgsplit(y) except InvalidData: self.invalidentry(self.getpath(y)) continue if not mysplit: self.invalidentry(self.getpath(y)) continue d[mysplit[0]+"/"+mysplit[1]] = None return d.keys() def checkblockers(self, origdep): pass def _clear_cache(self): self.mtdircache.clear() self.matchcache.clear() self.cpcache.clear() self._aux_cache_obj = None def _add(self, pkg_dblink): self._clear_pkg_cache(pkg_dblink) def _remove(self, pkg_dblink): self._clear_pkg_cache(pkg_dblink) def _clear_pkg_cache(self, pkg_dblink): # Due to 1 second mtime granularity in = self._aux_cache_threshold and \ secpass >= 2: self._owners.populate() # index any unindexed contents valid_nodes = set(self.cpv_all()) for cpv in self._aux_cache["packages"].keys(): if cpv not in valid_nodes: del self._aux_cache["packages"][cpv] del self._aux_cache["modified"] try: f = atomic_ofstream(self._aux_cache_filename) pickle.dump(self._aux_cache, f, -1) f.close() apply_secpass_permissions( self._aux_cache_filename, gid=portage_gid, mode=0644) except (IOError, OSError), e: pass self._aux_cache["modified"] = set() @property def _aux_cache(self): if self._aux_cache_obj is None: self._aux_cache_init() return self._aux_cache_obj def _aux_cache_init(self): aux_cache = None try: f = open(self._aux_cache_filename) mypickle = pickle.Unpickler(f) mypickle.find_global = None aux_cache = mypickle.load() f.close() del f except (IOError, OSError, EOFError, pickle.UnpicklingError), e: if isinstance(e, pickle.UnpicklingError): writemsg("!!! Error loading '%s': %s\n" % \ (self._aux_cache_filename, str(e)), noiselevel=-1) del e if not aux_cache or \ not isinstance(aux_cache, dict) or \ aux_cache.get("version") != self._aux_cache_version or \ not aux_cache.get("packages"): aux_cache = {"version": self._aux_cache_version} aux_cache["packages"] = {} owners = aux_cache.get("owners") if owners is not None: if not isinstance(owners, dict): owners = None elif "version" not in owners: owners = None elif owners["version"] != self._owners_cache_version: owners = None elif "base_names" not in owners: owners = None elif not isinstance(owners["base_names"], dict): owners = None if owners is None: owners = { "base_names" : {}, "version" : self._owners_cache_version } aux_cache["owners"] = owners aux_cache["modified"] = set() self._aux_cache_obj = aux_cache def aux_get(self, mycpv, wants): """This automatically caches selected keys that are frequently needed by emerge for dependency calculations. The cached metadata is considered valid if the mtime of the package directory has not changed since the data was cached. The cache is stored in a pickled dict object with the following format: {version:"1", "packages":{cpv1:(mtime,{k1,v1, k2,v2, ...}), cpv2...}} If an error occurs while loading the cache pickle or the version is unrecognized, the cache will simple be recreated from scratch (it is completely disposable). """ cache_these_wants = self._aux_cache_keys.intersection(wants) for x in wants: if self._aux_cache_keys_re.match(x) is not None: cache_these_wants.add(x) if not cache_these_wants: return self._aux_get(mycpv, wants) cache_these = set(self._aux_cache_keys) cache_these.update(cache_these_wants) mydir = self.getpath(mycpv) mydir_stat = None try: mydir_stat = os.stat(mydir) except OSError, e: if e.errno != errno.ENOENT: raise raise KeyError(mycpv) mydir_mtime = long(mydir_stat.st_mtime) pkg_data = self._aux_cache["packages"].get(mycpv) pull_me = cache_these.union(wants) mydata = {"_mtime_" : mydir_mtime} cache_valid = False cache_incomplete = False cache_mtime = None metadata = None if pkg_data is not None: if not isinstance(pkg_data, tuple) or len(pkg_data) != 2: pkg_data = None else: cache_mtime, metadata = pkg_data if not isinstance(cache_mtime, (long, int)) or \ not isinstance(metadata, dict): pkg_data = None if pkg_data: cache_mtime, metadata = pkg_data cache_valid = cache_mtime == mydir_mtime if cache_valid: mydata.update(metadata) pull_me.difference_update(mydata) if pull_me: # pull any needed data and cache it aux_keys = list(pull_me) for k, v in izip(aux_keys, self._aux_get(mycpv, aux_keys, st=mydir_stat)): mydata[k] = v if not cache_valid or cache_these.difference(metadata): cache_data = {} if cache_valid and metadata: cache_data.update(metadata) for aux_key in cache_these: cache_data[aux_key] = mydata[aux_key] self._aux_cache["packages"][mycpv] = (mydir_mtime, cache_data) self._aux_cache["modified"].add(mycpv) return [mydata[x] for x in wants] def _aux_get(self, mycpv, wants, st=None): mydir = self.getpath(mycpv) if st is None: try: st = os.stat(mydir) except OSError, e: if e.errno == errno.ENOENT: raise KeyError(mycpv) elif e.errno == PermissionDenied.errno: raise PermissionDenied(mydir) else: raise if not stat.S_ISDIR(st.st_mode): raise KeyError(mycpv) results = [] for x in wants: if x == "_mtime_": results.append(st.st_mtime) continue try: myf = open(os.path.join(mydir, x), "r") try: myd = myf.read() finally: myf.close() # Preserve \n for metadata that is known to # contain multiple lines. if self._aux_multi_line_re.match(x) is None: myd = " ".join(myd.split()) except IOError: myd = "" if x == "EAPI" and not myd: results.append("0") else: results.append(myd) return results def aux_update(self, cpv, values): cat, pkg = catsplit(cpv) mylink = dblink(cat, pkg, self.root, self.settings, treetype="vartree", vartree=self.vartree) if not mylink.exists(): raise KeyError(cpv) for k, v in values.iteritems(): if v: mylink.setfile(k, v) else: try: os.unlink(os.path.join(self.getpath(cpv), k)) except EnvironmentError: pass def counter_tick(self, myroot, mycpv=None): return self.counter_tick_core(myroot, incrementing=1, mycpv=mycpv) def get_counter_tick_core(self, myroot, mycpv=None): """ Use this method to retrieve the counter instead of having to trust the value of a global counter file that can lead to invalid COUNTER generation. When cache is valid, the package COUNTER files are not read and we rely on the timestamp of the package directory to validate cache. The stat calls should only take a short time, so performance is sufficient without having to rely on a potentially corrupt global counter file. The global counter file located at $CACHE_PATH/counter serves to record the counter of the last installed package and it also corresponds to the total number of installation actions that have occurred in the history of this package database. """ cp_list = self.cp_list max_counter = 0 for cp in self.cp_all(): for cpv in cp_list(cp): try: counter = int(self.aux_get(cpv, ["COUNTER"])[0]) except (KeyError, OverflowError, ValueError): continue if counter > max_counter: max_counter = counter new_vdb = False counter = -1 try: cfile = open(self._counter_path, "r") except EnvironmentError, e: new_vdb = not bool(self.cpv_all()) if not new_vdb: writemsg("!!! Unable to read COUNTER file: '%s'\n" % \ self._counter_path, noiselevel=-1) writemsg("!!! %s\n" % str(e), noiselevel=-1) del e else: try: try: counter = long(cfile.readline().strip()) finally: cfile.close() except (OverflowError, ValueError), e: writemsg("!!! COUNTER file is corrupt: '%s'\n" % \ self._counter_path, noiselevel=-1) writemsg("!!! %s\n" % str(e), noiselevel=-1) del e # We must ensure that we return a counter # value that is at least as large as the # highest one from the installed packages, # since having a corrupt value that is too low # can trigger incorrect AUTOCLEAN behavior due # to newly installed packages having lower # COUNTERs than the previous version in the # same slot. if counter > max_counter: max_counter = counter if counter < 0 and not new_vdb: writemsg("!!! Initializing COUNTER to " + \ "value of %d\n" % max_counter, noiselevel=-1) return max_counter + 1 def counter_tick_core(self, myroot, incrementing=1, mycpv=None): "This method will grab the next COUNTER value and record it back to the global file. Returns new counter value." counter = self.get_counter_tick_core(myroot, mycpv=mycpv) - 1 if incrementing: #increment counter counter += 1 # update new global counter file write_atomic(self._counter_path, str(counter)) return counter def _dblink(self, cpv): category, pf = catsplit(cpv) return dblink(category, pf, self.root, self.settings, vartree=self.vartree, treetype="vartree") def removeFromContents(self, pkg, paths, relative_paths=True): """ @param pkg: cpv for an installed package @type pkg: string @param paths: paths of files to remove from contents @type paths: iterable """ if not hasattr(pkg, "getcontents"): pkg = self._dblink(pkg) root = self.root root_len = len(root) - 1 new_contents = pkg.getcontents().copy() removed = 0 for filename in paths: filename = normalize_path(filename) if relative_paths: relative_filename = filename else: relative_filename = filename[root_len:] contents_key = pkg._match_contents(relative_filename, root) if contents_key: del new_contents[contents_key] removed += 1 if removed: f = atomic_ofstream(os.path.join(pkg.dbdir, "CONTENTS")) write_contents(new_contents, root, f) f.close() pkg._clear_contents_cache() class _owners_cache(object): """ This class maintains an hash table that serves to index package contents by mapping the basename of file to a list of possible packages that own it. This is used to optimize owner lookups by narrowing the search down to a smaller number of packages. """ try: from hashlib import md5 as _new_hash except ImportError: from md5 import new as _new_hash _hash_bits = 16 _hex_chars = _hash_bits / 4 def __init__(self, vardb): self._vardb = vardb def add(self, cpv): root_len = len(self._vardb.root) contents = self._vardb._dblink(cpv).getcontents() pkg_hash = self._hash_pkg(cpv) if not contents: # Empty path is a code used to represent empty contents. self._add_path("", pkg_hash) for x in contents: self._add_path(x[root_len:], pkg_hash) self._vardb._aux_cache["modified"].add(cpv) def _add_path(self, path, pkg_hash): """ Empty path is a code that represents empty contents. """ if path: name = os.path.basename(path.rstrip(os.path.sep)) if not name: return else: name = path name_hash = self._hash_str(name) base_names = self._vardb._aux_cache["owners"]["base_names"] pkgs = base_names.get(name_hash) if pkgs is None: pkgs = {} base_names[name_hash] = pkgs pkgs[pkg_hash] = None def _hash_str(self, s): h = self._new_hash() h.update(s) h = h.hexdigest() h = h[-self._hex_chars:] h = int(h, 16) return h def _hash_pkg(self, cpv): counter, mtime = self._vardb.aux_get( cpv, ["COUNTER", "_mtime_"]) try: counter = int(counter) except ValueError: counter = 0 return (cpv, counter, mtime) class _owners_db(object): def __init__(self, vardb): self._vardb = vardb def populate(self): self._populate() def _populate(self): owners_cache = vardbapi._owners_cache(self._vardb) cached_hashes = set() base_names = self._vardb._aux_cache["owners"]["base_names"] # Take inventory of all cached package hashes. for name, hash_values in base_names.items(): if not isinstance(hash_values, dict): del base_names[name] continue cached_hashes.update(hash_values) # Create sets of valid package hashes and uncached packages. uncached_pkgs = set() hash_pkg = owners_cache._hash_pkg valid_pkg_hashes = set() for cpv in self._vardb.cpv_all(): hash_value = hash_pkg(cpv) valid_pkg_hashes.add(hash_value) if hash_value not in cached_hashes: uncached_pkgs.add(cpv) # Cache any missing packages. for cpv in uncached_pkgs: owners_cache.add(cpv) # Delete any stale cache. stale_hashes = cached_hashes.difference(valid_pkg_hashes) if stale_hashes: for base_name_hash, bucket in base_names.items(): for hash_value in stale_hashes.intersection(bucket): del bucket[hash_value] if not bucket: del base_names[base_name_hash] return owners_cache def get_owners(self, path_iter): """ @return the owners as a dblink -> set(files) mapping. """ owners = {} for owner, f in self.iter_owners(path_iter): owned_files = owners.get(owner) if owned_files is None: owned_files = set() owners[owner] = owned_files owned_files.add(f) return owners def getFileOwnerMap(self, path_iter): owners = self.get_owners(path_iter) file_owners = {} for pkg_dblink, files in owners.iteritems(): for f in files: owner_set = file_owners.get(f) if owner_set is None: owner_set = set() file_owners[f] = owner_set owner_set.add(pkg_dblink) return file_owners def iter_owners(self, path_iter): """ Iterate over tuples of (dblink, path). In order to avoid consuming too many resources for too much time, resources are only allocated for the duration of a given iter_owners() call. Therefore, to maximize reuse of resources when searching for multiple files, it's best to search for them all in a single call. """ owners_cache = self._populate() vardb = self._vardb root = vardb.root hash_pkg = owners_cache._hash_pkg hash_str = owners_cache._hash_str base_names = self._vardb._aux_cache["owners"]["base_names"] dblink_cache = {} def dblink(cpv): x = dblink_cache.get(cpv) if x is None: x = self._vardb._dblink(cpv) dblink_cache[cpv] = x return x for path in path_iter: name = os.path.basename(path.rstrip(os.path.sep)) if not name: continue name_hash = hash_str(name) pkgs = base_names.get(name_hash) if pkgs is not None: for hash_value in pkgs: if not isinstance(hash_value, tuple) or \ len(hash_value) != 3: continue cpv, counter, mtime = hash_value if not isinstance(cpv, basestring): continue try: current_hash = hash_pkg(cpv) except KeyError: continue if current_hash != hash_value: continue if dblink(cpv).isowner(path, root): yield dblink(cpv), path class vartree(object): "this tree will scan a var/db/pkg database located at root (passed to init)" def __init__(self, root="/", virtual=None, clone=None, categories=None, settings=None): if clone: writemsg("vartree.__init__(): deprecated " + \ "use of clone parameter\n", noiselevel=-1) self.root = clone.root[:] self.dbapi = copy.deepcopy(clone.dbapi) self.populated = 1 from portage import config self.settings = config(clone=clone.settings) else: self.root = root[:] if settings is None: from portage import settings self.settings = settings # for key_expand calls if categories is None: categories = settings.categories self.dbapi = vardbapi(self.root, categories=categories, settings=settings, vartree=self) self.populated = 1 def getpath(self, mykey, filename=None): return self.dbapi.getpath(mykey, filename=filename) def zap(self, mycpv): return def inject(self, mycpv): return def get_provide(self, mycpv): myprovides = [] mylines = None try: mylines, myuse = self.dbapi.aux_get(mycpv, ["PROVIDE", "USE"]) if mylines: myuse = myuse.split() mylines = flatten(use_reduce(paren_reduce(mylines), uselist=myuse)) for myprovide in mylines: mys = catpkgsplit(myprovide) if not mys: mys = myprovide.split("/") myprovides += [mys[0] + "/" + mys[1]] return myprovides except SystemExit, e: raise except Exception, e: mydir = os.path.join(self.root, VDB_PATH, mycpv) writemsg("\nParse Error reading PROVIDE and USE in '%s'\n" % mydir, noiselevel=-1) if mylines: writemsg("Possibly Invalid: '%s'\n" % str(mylines), noiselevel=-1) writemsg("Exception: %s\n\n" % str(e), noiselevel=-1) return [] def get_all_provides(self): myprovides = {} for node in self.getallcpv(): for mykey in self.get_provide(node): if mykey in myprovides: myprovides[mykey] += [node] else: myprovides[mykey] = [node] return myprovides def dep_bestmatch(self, mydep, use_cache=1): "compatibility method -- all matches, not just visible ones" #mymatch=best(match(dep_expand(mydep,self.dbapi),self.dbapi)) mymatch = best(self.dbapi.match( dep_expand(mydep, mydb=self.dbapi, settings=self.settings), use_cache=use_cache)) if mymatch is None: return "" else: return mymatch def dep_match(self, mydep, use_cache=1): "compatibility method -- we want to see all matches, not just visible ones" #mymatch = match(mydep,self.dbapi) mymatch = self.dbapi.match(mydep, use_cache=use_cache) if mymatch is None: return [] else: return mymatch def exists_specific(self, cpv): return self.dbapi.cpv_exists(cpv) def getallcpv(self): """temporary function, probably to be renamed --- Gets a list of all category/package-versions installed on the system.""" return self.dbapi.cpv_all() def getallnodes(self): """new behavior: these are all *unmasked* nodes. There may or may not be available masked package for nodes in this nodes list.""" return self.dbapi.cp_all() def exists_specific_cat(self, cpv, use_cache=1): cpv = key_expand(cpv, mydb=self.dbapi, use_cache=use_cache, settings=self.settings) a = catpkgsplit(cpv) if not a: return 0 mylist = listdir(self.getpath(a[0]), EmptyOnError=1) for x in mylist: b = pkgsplit(x) if not b: self.dbapi.invalidentry(self.getpath(a[0], filename=x)) continue if a[1] == b[0]: return 1 return 0 def getebuildpath(self, fullpackage): cat, package = catsplit(fullpackage) return self.getpath(fullpackage, filename=package+".ebuild") def getnode(self, mykey, use_cache=1): mykey = key_expand(mykey, mydb=self.dbapi, use_cache=use_cache, settings=self.settings) if not mykey: return [] mysplit = catsplit(mykey) mydirlist = listdir(self.getpath(mysplit[0]),EmptyOnError=1) returnme = [] for x in mydirlist: mypsplit = pkgsplit(x) if not mypsplit: self.dbapi.invalidentry(self.getpath(mysplit[0], filename=x)) continue if mypsplit[0] == mysplit[1]: appendme = [mysplit[0]+"/"+x, [mysplit[0], mypsplit[0], mypsplit[1], mypsplit[2]]] returnme.append(appendme) return returnme def getslot(self, mycatpkg): "Get a slot for a catpkg; assume it exists." try: return self.dbapi.aux_get(mycatpkg, ["SLOT"])[0] except KeyError: return "" def hasnode(self, mykey, use_cache): """Does the particular node (cat/pkg key) exist?""" mykey = key_expand(mykey, mydb=self.dbapi, use_cache=use_cache, settings=self.settings) mysplit = catsplit(mykey) mydirlist = listdir(self.getpath(mysplit[0]), EmptyOnError=1) for x in mydirlist: mypsplit = pkgsplit(x) if not mypsplit: self.dbapi.invalidentry(self.getpath(mysplit[0], filename=x)) continue if mypsplit[0] == mysplit[1]: return 1 return 0 def populate(self): self.populated=1 class dblink(object): """ This class provides an interface to the installed package database At present this is implemented as a text backend in /var/db/pkg. """ import re _normalize_needed = re.compile(r'.*//.*|^[^/]|.+/$|(^|.*/)\.\.?(/.*|$)') _contents_split_counts = { "dev": 2, "dir": 2, "fif": 2, "obj": 4, "sym": 5 } # When looping over files for merge/unmerge, temporarily yield to the # scheduler each time this many files are processed. _file_merge_yield_interval = 20 def __init__(self, cat, pkg, myroot, mysettings, treetype=None, vartree=None, blockers=None, scheduler=None): """ Creates a DBlink object for a given CPV. The given CPV may not be present in the database already. @param cat: Category @type cat: String @param pkg: Package (PV) @type pkg: String @param myroot: Typically ${ROOT} @type myroot: String (Path) @param mysettings: Typically portage.config @type mysettings: An instance of portage.config @param treetype: one of ['porttree','bintree','vartree'] @type treetype: String @param vartree: an instance of vartree corresponding to myroot. @type vartree: vartree """ self.cat = cat self.pkg = pkg self.mycpv = self.cat + "/" + self.pkg self.mysplit = list(catpkgsplit(self.mycpv)[1:]) self.mysplit[0] = "%s/%s" % (self.cat, self.mysplit[0]) self.treetype = treetype if vartree is None: from portage import db vartree = db[myroot]["vartree"] self.vartree = vartree self._blockers = blockers self._scheduler = scheduler self.dbroot = normalize_path(os.path.join(myroot, VDB_PATH)) self.dbcatdir = self.dbroot+"/"+cat self.dbpkgdir = self.dbcatdir+"/"+pkg self.dbtmpdir = self.dbcatdir+"/-MERGING-"+pkg self.dbdir = self.dbpkgdir self._lock_vdb = None self.settings = mysettings self._verbose = self.settings.get("PORTAGE_VERBOSE") == "1" self.myroot=myroot protect_obj = ConfigProtect(myroot, shlex.split(mysettings.get("CONFIG_PROTECT", "")), shlex.split(mysettings.get("CONFIG_PROTECT_MASK", ""))) self.updateprotect = protect_obj.updateprotect self.isprotected = protect_obj.isprotected self._installed_instance = None self.contentscache = None self._contents_inodes = None self._contents_basenames = None self._linkmap_broken = False def lockdb(self): if self._lock_vdb: raise AssertionError("Lock already held.") # At least the parent needs to exist for the lock file. ensure_dirs(self.dbroot) self._lock_vdb = lockdir(self.dbroot) def unlockdb(self): if self._lock_vdb: unlockdir(self._lock_vdb) self._lock_vdb = None def getpath(self): "return path to location of db information (for >>> informational display)" return self.dbdir def exists(self): "does the db entry exist? boolean." return os.path.exists(self.dbdir) def delete(self): """ Remove this entry from the database """ if not os.path.exists(self.dbdir): return # Check validity of self.dbdir before attempting to remove it. if not self.dbdir.startswith(self.dbroot): writemsg("portage.dblink.delete(): invalid dbdir: %s\n" % \ self.dbdir, noiselevel=-1) return shutil.rmtree(self.dbdir) self.vartree.dbapi._remove(self) def clearcontents(self): """ For a given db entry (self), erase the CONTENTS values. """ if os.path.exists(self.dbdir+"/CONTENTS"): os.unlink(self.dbdir+"/CONTENTS") def _clear_contents_cache(self): self.contentscache = None self._contents_inodes = None self._contents_basenames = None def getcontents(self): """ Get the installed files of a given package (aka what that package installed) """ contents_file = os.path.join(self.dbdir, "CONTENTS") if self.contentscache is not None: return self.contentscache pkgfiles = {} try: myc = open(contents_file,"r") except EnvironmentError, e: if e.errno != errno.ENOENT: raise del e self.contentscache = pkgfiles return pkgfiles mylines = myc.readlines() myc.close() null_byte = "\0" normalize_needed = self._normalize_needed contents_split_counts = self._contents_split_counts myroot = self.myroot if myroot == os.path.sep: myroot = None pos = 0 errors = [] for pos, line in enumerate(mylines): if null_byte in line: # Null bytes are a common indication of corruption. errors.append((pos + 1, "Null byte found in CONTENTS entry")) continue line = line.rstrip("\n") # Split on " " so that even file paths that # end with spaces can be handled. mydat = line.split(" ") entry_type = mydat[0] # empty string if line is empty correct_split_count = contents_split_counts.get(entry_type) if correct_split_count and len(mydat) > correct_split_count: # Apparently file paths contain spaces, so reassemble # the split have the correct_split_count. newsplit = [entry_type] spaces_total = len(mydat) - correct_split_count if entry_type == "sym": try: splitter = mydat.index("->", 2, len(mydat) - 2) except ValueError: errors.append((pos + 1, "Unrecognized CONTENTS entry")) continue spaces_in_path = splitter - 2 spaces_in_target = spaces_total - spaces_in_path newsplit.append(" ".join(mydat[1:splitter])) newsplit.append("->") target_end = splitter + spaces_in_target + 2 newsplit.append(" ".join(mydat[splitter + 1:target_end])) newsplit.extend(mydat[target_end:]) else: path_end = spaces_total + 2 newsplit.append(" ".join(mydat[1:path_end])) newsplit.extend(mydat[path_end:]) mydat = newsplit # we do this so we can remove from non-root filesystems # (use the ROOT var to allow maintenance on other partitions) try: if normalize_needed.match(mydat[1]): mydat[1] = normalize_path(mydat[1]) if not mydat[1].startswith(os.path.sep): mydat[1] = os.path.sep + mydat[1] if myroot: mydat[1] = os.path.join(myroot, mydat[1].lstrip(os.path.sep)) if mydat[0] == "obj": #format: type, mtime, md5sum pkgfiles[mydat[1]] = [mydat[0], mydat[3], mydat[2]] elif mydat[0] == "dir": #format: type pkgfiles[mydat[1]] = [mydat[0]] elif mydat[0] == "sym": #format: type, mtime, dest pkgfiles[mydat[1]] = [mydat[0], mydat[4], mydat[3]] elif mydat[0] == "dev": #format: type pkgfiles[mydat[1]] = [mydat[0]] elif mydat[0]=="fif": #format: type pkgfiles[mydat[1]] = [mydat[0]] else: errors.append((pos + 1, "Unrecognized CONTENTS entry")) except (KeyError, IndexError): errors.append((pos + 1, "Unrecognized CONTENTS entry")) if errors: writemsg("!!! Parse error in '%s'\n" % contents_file, noiselevel=-1) for pos, e in errors: writemsg("!!! line %d: %s\n" % (pos, e), noiselevel=-1) self.contentscache = pkgfiles return pkgfiles def unmerge(self, pkgfiles=None, trimworld=1, cleanup=1, ldpath_mtimes=None, others_in_slot=None): """ Calls prerm Unmerges a given package (CPV) calls postrm calls cleanrm calls env_update @param pkgfiles: files to unmerge (generally self.getcontents() ) @type pkgfiles: Dictionary @param trimworld: Remove CPV from world file if True, not if False @type trimworld: Boolean @param cleanup: cleanup to pass to doebuild (see doebuild) @type cleanup: Boolean @param ldpath_mtimes: mtimes to pass to env_update (see env_update) @type ldpath_mtimes: Dictionary @param others_in_slot: all dblink instances in this slot, excluding self @type others_in_slot: list @rtype: Integer @returns: 1. os.EX_OK if everything went well. 2. return code of the failed phase (for prerm, postrm, cleanrm) Notes: The caller must ensure that lockdb() and unlockdb() are called before and after this method. """ showMessage = self._display_merge if self.vartree.dbapi._categories is not None: self.vartree.dbapi._categories = None # When others_in_slot is supplied, the security check has already been # done for this slot, so it shouldn't be repeated until the next # replacement or unmerge operation. if others_in_slot is None: slot = self.vartree.dbapi.aux_get(self.mycpv, ["SLOT"])[0] slot_matches = self.vartree.dbapi.match( "%s:%s" % (dep_getkey(self.mycpv), slot)) others_in_slot = [] for cur_cpv in slot_matches: if cur_cpv == self.mycpv: continue others_in_slot.append(dblink(self.cat, catsplit(cur_cpv)[1], self.vartree.root, self.settings, vartree=self.vartree, treetype="vartree")) retval = self._security_check([self] + others_in_slot) if retval: return retval contents = self.getcontents() # Now, don't assume that the name of the ebuild is the same as the # name of the dir; the package may have been moved. myebuildpath = None ebuild_phase = "prerm" log_path = None mystuff = os.listdir(self.dbdir) for x in mystuff: if x.endswith(".ebuild"): myebuildpath = os.path.join(self.dbdir, self.pkg + ".ebuild") if x[:-7] != self.pkg: # Clean up after vardbapi.move_ent() breakage in # portage versions before 2.1.2 os.rename(os.path.join(self.dbdir, x), myebuildpath) write_atomic(os.path.join(self.dbdir, "PF"), self.pkg+"\n") break self.settings.setcpv(self.mycpv, mydb=self.vartree.dbapi) if myebuildpath: try: doebuild_environment(myebuildpath, "prerm", self.myroot, self.settings, 0, 0, self.vartree.dbapi) except UnsupportedAPIException, e: # Sometimes this happens due to corruption of the EAPI file. writemsg("!!! FAILED prerm: %s\n" % \ os.path.join(self.dbdir, "EAPI"), noiselevel=-1) writemsg("%s\n" % str(e), noiselevel=-1) myebuildpath = None else: catdir = os.path.dirname(self.settings["PORTAGE_BUILDDIR"]) ensure_dirs(os.path.dirname(catdir), uid=portage_uid, gid=portage_gid, mode=070, mask=0) builddir_lock = None catdir_lock = None scheduler = self._scheduler retval = -1 try: if myebuildpath: catdir_lock = lockdir(catdir) ensure_dirs(catdir, uid=portage_uid, gid=portage_gid, mode=070, mask=0) builddir_lock = lockdir( self.settings["PORTAGE_BUILDDIR"]) try: unlockdir(catdir_lock) finally: catdir_lock = None prepare_build_dirs(self.myroot, self.settings, 1) log_path = self.settings.get("PORTAGE_LOG_FILE") if scheduler is None: retval = doebuild(myebuildpath, ebuild_phase, self.myroot, self.settings, cleanup=cleanup, use_cache=0, mydbapi=self.vartree.dbapi, tree=self.treetype, vartree=self.vartree) else: retval = scheduler.dblinkEbuildPhase( self, self.vartree.dbapi, myebuildpath, ebuild_phase) # XXX: Decide how to handle failures here. if retval != os.EX_OK: writemsg("!!! FAILED prerm: %s\n" % retval, noiselevel=-1) self._unmerge_pkgfiles(pkgfiles, others_in_slot) self._clear_contents_cache() # Remove the registration of preserved libs for this pkg instance plib_registry = self.vartree.dbapi.plib_registry plib_registry.unregister(self.mycpv, self.settings["SLOT"], self.vartree.dbapi.cpv_counter(self.mycpv)) if myebuildpath: ebuild_phase = "postrm" if scheduler is None: retval = doebuild(myebuildpath, ebuild_phase, self.myroot, self.settings, use_cache=0, tree=self.treetype, mydbapi=self.vartree.dbapi, vartree=self.vartree) else: retval = scheduler.dblinkEbuildPhase( self, self.vartree.dbapi, myebuildpath, ebuild_phase) # XXX: Decide how to handle failures here. if retval != os.EX_OK: writemsg("!!! FAILED postrm: %s\n" % retval, noiselevel=-1) # Skip this if another package in the same slot has just been # merged on top of this package, since the other package has # already called LinkageMap.rebuild() and passed it's NEEDED file # in as an argument. if not others_in_slot: self._linkmap_rebuild(exclude_pkgs=(self.mycpv,)) # remove preserved libraries that don't have any consumers left cpv_lib_map = self._find_unused_preserved_libs() if cpv_lib_map: self._remove_preserved_libs(cpv_lib_map) for cpv, removed in cpv_lib_map.iteritems(): if not self.vartree.dbapi.cpv_exists(cpv): for dblnk in others_in_slot: if dblnk.mycpv == cpv: # This one just got merged so it doesn't # register with cpv_exists() yet. self.vartree.dbapi.removeFromContents( dblnk, removed) break continue self.vartree.dbapi.removeFromContents(cpv, removed) finally: if builddir_lock: try: if myebuildpath: if retval != os.EX_OK: msg_lines = [] msg = ("The '%s' " % ebuild_phase) + \ ("phase of the '%s' package " % self.mycpv) + \ ("has failed with exit value %s." % retval) from textwrap import wrap msg_lines.extend(wrap(msg, 72)) msg_lines.append("") ebuild_name = os.path.basename(myebuildpath) ebuild_dir = os.path.dirname(myebuildpath) msg = "The problem occurred while executing " + \ ("the ebuild file named '%s' " % ebuild_name) + \ ("located in the '%s' directory. " \ % ebuild_dir) + \ "If necessary, manually remove " + \ "the environment.bz2 file and/or the " + \ "ebuild file located in that directory." msg_lines.extend(wrap(msg, 72)) msg_lines.append("") msg = "Removal " + \ "of the environment.bz2 file is " + \ "preferred since it may allow the " + \ "removal phases to execute successfully. " + \ "The ebuild will be " + \ "sourced and the eclasses " + \ "from the current portage tree will be used " + \ "when necessary. Removal of " + \ "the ebuild file will cause the " + \ "pkg_prerm() and pkg_postrm() removal " + \ "phases to be skipped entirely." msg_lines.extend(wrap(msg, 72)) self._eerror(ebuild_phase, msg_lines) # process logs created during pre/postrm elog_process(self.mycpv, self.settings, phasefilter=filter_unmergephases) if retval == os.EX_OK: if scheduler is None: doebuild(myebuildpath, "cleanrm", self.myroot, self.settings, tree="vartree", mydbapi=self.vartree.dbapi, vartree=self.vartree) else: scheduler.dblinkEbuildPhase( self, self.vartree.dbapi, myebuildpath, "cleanrm") finally: unlockdir(builddir_lock) try: if myebuildpath and not catdir_lock: # Lock catdir for removal if empty. catdir_lock = lockdir(catdir) finally: if catdir_lock: try: os.rmdir(catdir) except OSError, e: if e.errno not in (errno.ENOENT, errno.ENOTEMPTY, errno.EEXIST): raise del e unlockdir(catdir_lock) if log_path is not None and os.path.exists(log_path): # Restore this since it gets lost somewhere above and it # needs to be set for _display_merge() to be able to log. # Note that the log isn't necessarily supposed to exist # since if PORT_LOGDIR is unset then it's a temp file # so it gets cleaned above. self.settings["PORTAGE_LOG_FILE"] = log_path else: self.settings.pop("PORTAGE_LOG_FILE", None) env_update(target_root=self.myroot, prev_mtimes=ldpath_mtimes, contents=contents, env=self.settings.environ(), writemsg_level=self._display_merge) return os.EX_OK def _display_merge(self, msg, level=0, noiselevel=0): if not self._verbose and noiselevel >= 0 and level < logging.WARN: return if self._scheduler is not None: self._scheduler.dblinkDisplayMerge(self, msg, level=level, noiselevel=noiselevel) return writemsg_level(msg, level=level, noiselevel=noiselevel) def _unmerge_pkgfiles(self, pkgfiles, others_in_slot): """ Unmerges the contents of a package from the liveFS Removes the VDB entry for self @param pkgfiles: typically self.getcontents() @type pkgfiles: Dictionary { filename: [ 'type', '?', 'md5sum' ] } @param others_in_slot: all dblink instances in this slot, excluding self @type others_in_slot: list @rtype: None """ showMessage = self._display_merge scheduler = self._scheduler if not pkgfiles: showMessage("No package files given... Grabbing a set.\n") pkgfiles = self.getcontents() if others_in_slot is None: others_in_slot = [] slot = self.vartree.dbapi.aux_get(self.mycpv, ["SLOT"])[0] slot_matches = self.vartree.dbapi.match( "%s:%s" % (dep_getkey(self.mycpv), slot)) for cur_cpv in slot_matches: if cur_cpv == self.mycpv: continue others_in_slot.append(dblink(self.cat, catsplit(cur_cpv)[1], self.vartree.root, self.settings, vartree=self.vartree, treetype="vartree")) dest_root = normalize_path(self.vartree.root).rstrip(os.path.sep) + \ os.path.sep dest_root_len = len(dest_root) - 1 conf_mem_file = os.path.join(dest_root, CONFIG_MEMORY_FILE) cfgfiledict = grabdict(conf_mem_file) stale_confmem = [] unmerge_orphans = "unmerge-orphans" in self.settings.features if pkgfiles: self.updateprotect() mykeys = pkgfiles.keys() mykeys.sort() mykeys.reverse() #process symlinks second-to-last, directories last. mydirs = [] ignored_unlink_errnos = ( errno.EBUSY, errno.ENOENT, errno.ENOTDIR, errno.EISDIR) ignored_rmdir_errnos = ( errno.EEXIST, errno.ENOTEMPTY, errno.EBUSY, errno.ENOENT, errno.ENOTDIR, errno.EISDIR) modprotect = os.path.join(self.vartree.root, "lib/modules/") def unlink(file_name, lstatobj): if bsd_chflags: if lstatobj.st_flags != 0: bsd_chflags.lchflags(file_name, 0) parent_name = os.path.dirname(file_name) # Use normal stat/chflags for the parent since we want to # follow any symlinks to the real parent directory. pflags = os.stat(parent_name).st_flags if pflags != 0: bsd_chflags.chflags(parent_name, 0) try: if not stat.S_ISLNK(lstatobj.st_mode): # Remove permissions to ensure that any hardlinks to # suid/sgid files are rendered harmless. os.chmod(file_name, 0) os.unlink(file_name) finally: if bsd_chflags and pflags != 0: # Restore the parent flags we saved before unlinking bsd_chflags.chflags(parent_name, pflags) def show_unmerge(zing, desc, file_type, file_name): showMessage("%s %s %s %s\n" % \ (zing, desc.ljust(8), file_type, file_name)) for i, objkey in enumerate(mykeys): if scheduler is not None and \ 0 == i % self._file_merge_yield_interval: scheduler.scheduleYield() obj = normalize_path(objkey) file_data = pkgfiles[objkey] file_type = file_data[0] statobj = None try: statobj = os.stat(obj) except OSError: pass lstatobj = None try: lstatobj = os.lstat(obj) except (OSError, AttributeError): pass islink = lstatobj is not None and stat.S_ISLNK(lstatobj.st_mode) if lstatobj is None: show_unmerge("---", "!found", file_type, obj) continue if obj.startswith(dest_root): relative_path = obj[dest_root_len:] is_owned = False for dblnk in others_in_slot: if dblnk.isowner(relative_path, dest_root): is_owned = True break if is_owned: # A new instance of this package claims the file, so # don't unmerge it. show_unmerge("---", "replaced", file_type, obj) continue elif relative_path in cfgfiledict: stale_confmem.append(relative_path) # next line includes a tweak to protect modules from being unmerged, # but we don't protect modules from being overwritten if they are # upgraded. We effectively only want one half of the config protection # functionality for /lib/modules. For portage-ng both capabilities # should be able to be independently specified. if obj.startswith(modprotect): show_unmerge("---", "cfgpro", file_type, obj) continue # Don't unlink symlinks to directories here since that can # remove /lib and /usr/lib symlinks. if unmerge_orphans and \ lstatobj and not stat.S_ISDIR(lstatobj.st_mode) and \ not (islink and statobj and stat.S_ISDIR(statobj.st_mode)) and \ not self.isprotected(obj): try: unlink(obj, lstatobj) except EnvironmentError, e: if e.errno not in ignored_unlink_errnos: raise del e show_unmerge("<<<", "", file_type, obj) continue lmtime = str(lstatobj[stat.ST_MTIME]) if (pkgfiles[objkey][0] not in ("dir", "fif", "dev")) and (lmtime != pkgfiles[objkey][1]): show_unmerge("---", "!mtime", file_type, obj) continue if pkgfiles[objkey][0] == "dir": if statobj is None or not stat.S_ISDIR(statobj.st_mode): show_unmerge("---", "!dir", file_type, obj) continue mydirs.append(obj) elif pkgfiles[objkey][0] == "sym": if not islink: show_unmerge("---", "!sym", file_type, obj) continue # Go ahead and unlink symlinks to directories here when # they're actually recorded as symlinks in the contents. # Normally, symlinks such as /lib -> lib64 are not recorded # as symlinks in the contents of a package. If a package # installs something into ${D}/lib/, it is recorded in the # contents as a directory even if it happens to correspond # to a symlink when it's merged to the live filesystem. try: unlink(obj, lstatobj) show_unmerge("<<<", "", file_type, obj) except (OSError, IOError),e: if e.errno not in ignored_unlink_errnos: raise del e show_unmerge("!!!", "", file_type, obj) elif pkgfiles[objkey][0] == "obj": if statobj is None or not stat.S_ISREG(statobj.st_mode): show_unmerge("---", "!obj", file_type, obj) continue mymd5 = None try: mymd5 = perform_md5(obj, calc_prelink=1) except FileNotFound, e: # the file has disappeared between now and our stat call show_unmerge("---", "!obj", file_type, obj) continue # string.lower is needed because db entries used to be in upper-case. The # string.lower allows for backwards compatibility. if mymd5 != pkgfiles[objkey][2].lower(): show_unmerge("---", "!md5", file_type, obj) continue try: unlink(obj, lstatobj) except (OSError, IOError), e: if e.errno not in ignored_unlink_errnos: raise del e show_unmerge("<<<", "", file_type, obj) elif pkgfiles[objkey][0] == "fif": if not stat.S_ISFIFO(lstatobj[stat.ST_MODE]): show_unmerge("---", "!fif", file_type, obj) continue show_unmerge("---", "", file_type, obj) elif pkgfiles[objkey][0] == "dev": show_unmerge("---", "", file_type, obj) mydirs.sort() mydirs.reverse() for obj in mydirs: try: if bsd_chflags: lstatobj = os.lstat(obj) if lstatobj.st_flags != 0: bsd_chflags.lchflags(obj, 0) parent_name = os.path.dirname(obj) # Use normal stat/chflags for the parent since we want to # follow any symlinks to the real parent directory. pflags = os.stat(parent_name).st_flags if pflags != 0: bsd_chflags.chflags(parent_name, 0) try: os.rmdir(obj) finally: if bsd_chflags and pflags != 0: # Restore the parent flags we saved before unlinking bsd_chflags.chflags(parent_name, pflags) show_unmerge("<<<", "", "dir", obj) except EnvironmentError, e: if e.errno not in ignored_rmdir_errnos: raise if e.errno != errno.ENOENT: show_unmerge("---", "!empty", "dir", obj) del e # Remove stale entries from config memory. if stale_confmem: for filename in stale_confmem: del cfgfiledict[filename] writedict(cfgfiledict, conf_mem_file) #remove self from vartree database so that our own virtual gets zapped if we're the last node self.vartree.zap(self.mycpv) def isowner(self, filename, destroot): """ Check if a file belongs to this package. This may result in a stat call for the parent directory of every installed file, since the inode numbers are used to work around the problem of ambiguous paths caused by symlinked directories. The results of stat calls are cached to optimize multiple calls to this method. @param filename: @type filename: @param destroot: @type destroot: @rtype: Boolean @returns: 1. True if this package owns the file. 2. False if this package does not own the file. """ return bool(self._match_contents(filename, destroot)) def _match_contents(self, filename, destroot): """ The matching contents entry is returned, which is useful since the path may differ from the one given by the caller, due to symlinks. @rtype: String @return: the contents entry corresponding to the given path, or False if the file is not owned by this package. """ destfile = normalize_path( os.path.join(destroot, filename.lstrip(os.path.sep))) pkgfiles = self.getcontents() if pkgfiles and destfile in pkgfiles: return destfile if pkgfiles: basename = os.path.basename(destfile) if self._contents_basenames is None: self._contents_basenames = set( os.path.basename(x) for x in pkgfiles) if basename not in self._contents_basenames: # This is a shortcut that, in most cases, allows us to # eliminate this package as an owner without the need # to examine inode numbers of parent directories. return False # Use stat rather than lstat since we want to follow # any symlinks to the real parent directory. parent_path = os.path.dirname(destfile) try: parent_stat = os.stat(parent_path) except EnvironmentError, e: if e.errno != errno.ENOENT: raise del e return False if self._contents_inodes is None: self._contents_inodes = {} parent_paths = set() for x in pkgfiles: p_path = os.path.dirname(x) if p_path in parent_paths: continue parent_paths.add(p_path) try: s = os.stat(p_path) except OSError: pass else: inode_key = (s.st_dev, s.st_ino) # Use lists of paths in case multiple # paths reference the same inode. p_path_list = self._contents_inodes.get(inode_key) if p_path_list is None: p_path_list = [] self._contents_inodes[inode_key] = p_path_list if p_path not in p_path_list: p_path_list.append(p_path) p_path_list = self._contents_inodes.get( (parent_stat.st_dev, parent_stat.st_ino)) if p_path_list: for p_path in p_path_list: x = os.path.join(p_path, basename) if x in pkgfiles: return x return False def _linkmap_rebuild(self, **kwargs): if self._linkmap_broken: return try: self.vartree.dbapi.linkmap.rebuild(**kwargs) except CommandNotFound, e: self._linkmap_broken = True self._display_merge("!!! Disabling preserve-libs " + \ "due to error: Command Not Found: %s\n" % (e,), level=logging.ERROR, noiselevel=-1) def _find_libs_to_preserve(self): """ Get set of relative paths for libraries to be preserved. The file paths are selected from self._installed_instance.getcontents(). """ if self._linkmap_broken or not \ (self._installed_instance is not None and \ "preserve-libs" in self.settings.features): return None linkmap = self.vartree.dbapi.linkmap installed_instance = self._installed_instance old_contents = installed_instance.getcontents() root = self.myroot root_len = len(root) - 1 lib_graph = digraph() path_node_map = {} def path_to_node(path): node = path_node_map.get(path) if node is None: node = LinkageMap._LibGraphNode(path, root) alt_path_node = lib_graph.get(node) if alt_path_node is not None: node = alt_path_node node.alt_paths.add(path) path_node_map[path] = node return node consumer_map = {} provider_nodes = set() # Create provider nodes and add them to the graph. for f_abs in old_contents: f = f_abs[root_len:] if self.isowner(f, root): continue try: consumers = linkmap.findConsumers(f) except KeyError: continue if not consumers: continue provider_node = path_to_node(f) lib_graph.add(provider_node, None) provider_nodes.add(provider_node) consumer_map[provider_node] = consumers # Create consumer nodes and add them to the graph. # Note that consumers can also be providers. for provider_node, consumers in consumer_map.iteritems(): for c in consumers: if self.isowner(c, root): continue consumer_node = path_to_node(c) if installed_instance.isowner(c, root) and \ consumer_node not in provider_nodes: # This is not a provider, so it will be uninstalled. continue lib_graph.add(provider_node, consumer_node) # Locate nodes which should be preserved. They consist of all # providers that are reachable from consumers that are not # providers themselves. preserve_nodes = set() for consumer_node in lib_graph.root_nodes(): if consumer_node in provider_nodes: continue # Preserve all providers that are reachable from this consumer. node_stack = lib_graph.child_nodes(consumer_node) while node_stack: provider_node = node_stack.pop() if provider_node in preserve_nodes: continue preserve_nodes.add(provider_node) node_stack.extend(lib_graph.child_nodes(provider_node)) preserve_paths = set() for preserve_node in preserve_nodes: preserve_paths.update(preserve_node.alt_paths) return preserve_paths def _add_preserve_libs_to_contents(self, preserve_paths): """ Preserve libs returned from _find_libs_to_preserve(). """ if not preserve_paths: return showMessage = self._display_merge root = self.myroot # Copy contents entries from the old package to the new one. new_contents = self.getcontents().copy() old_contents = self._installed_instance.getcontents() for f in sorted(preserve_paths): f_abs = os.path.join(root, f.lstrip(os.sep)) contents_entry = old_contents.get(f_abs) if contents_entry is None: # This will probably never happen, but it might if one of the # paths returned from findConsumers() refers to one of the libs # that should be preserved yet the path is not listed in the # contents. Such a path might belong to some other package, so # it shouldn't be preserved here. showMessage(("!!! File '%s' will not be preserved " + \ "due to missing contents entry\n") % (f_abs,), level=logging.ERROR, noiselevel=-1) preserve_paths.remove(f) continue new_contents[f_abs] = contents_entry obj_type = contents_entry[0] showMessage(">>> needed %s %s\n" % (obj_type, f_abs), noiselevel=-1) # Add parent directories to contents if necessary. parent_dir = os.path.dirname(f_abs) while len(parent_dir) > len(root): new_contents[parent_dir] = ["dir"] prev = parent_dir parent_dir = os.path.dirname(parent_dir) if prev == parent_dir: break outfile = atomic_ofstream(os.path.join(self.dbtmpdir, "CONTENTS")) write_contents(new_contents, root, outfile) outfile.close() self._clear_contents_cache() def _find_unused_preserved_libs(self): """ Find preserved libraries that don't have any consumers left. """ if self._linkmap_broken: return {} # Since preserved libraries can be consumers of other preserved # libraries, use a graph to track consumer relationships. plib_dict = self.vartree.dbapi.plib_registry.getPreservedLibs() lib_graph = digraph() preserved_nodes = set() preserved_paths = set() path_cpv_map = {} path_node_map = {} root = self.myroot def path_to_node(path): node = path_node_map.get(path) if node is None: node = LinkageMap._LibGraphNode(path, root) alt_path_node = lib_graph.get(node) if alt_path_node is not None: node = alt_path_node node.alt_paths.add(path) path_node_map[path] = node return node linkmap = self.vartree.dbapi.linkmap for cpv, plibs in plib_dict.iteritems(): for f in plibs: path_cpv_map[f] = cpv preserved_node = path_to_node(f) if not preserved_node.file_exists(): continue lib_graph.add(preserved_node, None) preserved_paths.add(f) preserved_nodes.add(preserved_node) for c in self.vartree.dbapi.linkmap.findConsumers(f): consumer_node = path_to_node(c) if not consumer_node.file_exists(): continue # Note that consumers may also be providers. lib_graph.add(preserved_node, consumer_node) # Eliminate consumers having providers with the same soname as an # installed library that is not preserved. This eliminates # libraries that are erroneously preserved due to a move from one # directory to another. provider_cache = {} for preserved_node in preserved_nodes: soname = linkmap.getSoname(preserved_node) for consumer_node in lib_graph.parent_nodes(preserved_node): if consumer_node in preserved_nodes: continue providers = provider_cache.get(consumer_node) if providers is None: providers = linkmap.findProviders(consumer_node) provider_cache[consumer_node] = providers providers = providers.get(soname) if providers is None: continue for provider in providers: if provider in preserved_paths: continue provider_node = path_to_node(provider) if not provider_node.file_exists(): continue if provider_node in preserved_nodes: continue # An alternative provider seems to be # installed, so drop this edge. lib_graph.remove_edge(preserved_node, consumer_node) break cpv_lib_map = {} while not lib_graph.empty(): root_nodes = preserved_nodes.intersection(lib_graph.root_nodes()) if not root_nodes: break lib_graph.difference_update(root_nodes) unlink_list = set() for node in root_nodes: unlink_list.update(node.alt_paths) unlink_list = sorted(unlink_list) for obj in unlink_list: cpv = path_cpv_map[obj] removed = cpv_lib_map.get(cpv) if removed is None: removed = set() cpv_lib_map[cpv] = removed removed.add(obj) return cpv_lib_map def _remove_preserved_libs(self, cpv_lib_map): """ Remove files returned from _find_unused_preserved_libs(). """ files_to_remove = set() for files in cpv_lib_map.itervalues(): files_to_remove.update(files) files_to_remove = sorted(files_to_remove) showMessage = self._display_merge root = self.myroot parent_dirs = set() for obj in files_to_remove: obj = os.path.join(root, obj.lstrip(os.sep)) parent_dirs.add(os.path.dirname(obj)) if os.path.islink(obj): obj_type = "sym" else: obj_type = "obj" try: os.unlink(obj) except OSError, e: if e.errno != errno.ENOENT: raise del e else: showMessage("<<< !needed %s %s\n" % (obj_type, obj), noiselevel=-1) # Remove empty parent directories if possible. while parent_dirs: x = parent_dirs.pop() while True: try: os.rmdir(x) except OSError: break prev = x x = os.path.dirname(x) if x == prev: break self.vartree.dbapi.plib_registry.pruneNonExisting() def _collision_protect(self, srcroot, destroot, mypkglist, mycontents): collision_ignore = set([normalize_path(myignore) for myignore in \ shlex.split(self.settings.get("COLLISION_IGNORE", ""))]) # For collisions with preserved libraries, the current package # will assume ownership and the libraries will be unregistered. plib_dict = self.vartree.dbapi.plib_registry.getPreservedLibs() plib_cpv_map = {} plib_paths = set() for cpv, paths in plib_dict.iteritems(): plib_paths.update(paths) for f in paths: plib_cpv_map[f] = cpv plib_inodes = self._lstat_inode_map(plib_paths) plib_collisions = {} showMessage = self._display_merge scheduler = self._scheduler stopmerge = False collisions = [] destroot = normalize_path(destroot).rstrip(os.path.sep) + \ os.path.sep showMessage("%s checking %d files for package collisions\n" % \ (green("*"), len(mycontents))) for i, f in enumerate(mycontents): if i % 1000 == 0 and i != 0: showMessage("%d files checked ...\n" % i) if scheduler is not None and \ 0 == i % self._file_merge_yield_interval: scheduler.scheduleYield() dest_path = normalize_path( os.path.join(destroot, f.lstrip(os.path.sep))) try: dest_lstat = os.lstat(dest_path) except EnvironmentError, e: if e.errno == errno.ENOENT: del e continue elif e.errno == errno.ENOTDIR: del e # A non-directory is in a location where this package # expects to have a directory. dest_lstat = None parent_path = dest_path while len(parent_path) > len(destroot): parent_path = os.path.dirname(parent_path) try: dest_lstat = os.lstat(parent_path) break except EnvironmentError, e: if e.errno != errno.ENOTDIR: raise del e if not dest_lstat: raise AssertionError( "unable to find non-directory " + \ "parent for '%s'" % dest_path) dest_path = parent_path f = os.path.sep + dest_path[len(destroot):] if f in collisions: continue else: raise if f[0] != "/": f="/"+f plibs = plib_inodes.get((dest_lstat.st_dev, dest_lstat.st_ino)) if plibs: for path in plibs: cpv = plib_cpv_map[path] paths = plib_collisions.get(cpv) if paths is None: paths = set() plib_collisions[cpv] = paths paths.add(path) # The current package will assume ownership and the # libraries will be unregistered, so exclude this # path from the normal collisions. continue isowned = False full_path = os.path.join(destroot, f.lstrip(os.path.sep)) for ver in mypkglist: if ver.isowner(f, destroot): isowned = True break if not isowned and self.isprotected(full_path): isowned = True if not isowned: stopmerge = True if collision_ignore: if f in collision_ignore: stopmerge = False else: for myignore in collision_ignore: if f.startswith(myignore + os.path.sep): stopmerge = False break if stopmerge: collisions.append(f) return collisions, plib_collisions def _lstat_inode_map(self, path_iter): """ Use lstat to create a map of the form: {(st_dev, st_ino) : set([path1, path2, ...])} Multiple paths may reference the same inode due to hardlinks. All lstat() calls are relative to self.myroot. """ root = self.myroot inode_map = {} for f in path_iter: path = os.path.join(root, f.lstrip(os.sep)) try: st = os.lstat(path) except OSError, e: if e.errno not in (errno.ENOENT, errno.ENOTDIR): raise del e continue key = (st.st_dev, st.st_ino) paths = inode_map.get(key) if paths is None: paths = set() inode_map[key] = paths paths.add(f) return inode_map def _security_check(self, installed_instances): if not installed_instances: return 0 showMessage = self._display_merge scheduler = self._scheduler file_paths = set() for dblnk in installed_instances: file_paths.update(dblnk.getcontents()) inode_map = {} real_paths = set() for i, path in enumerate(file_paths): if scheduler is not None and \ 0 == i % self._file_merge_yield_interval: scheduler.scheduleYield() try: s = os.lstat(path) except OSError, e: if e.errno not in (errno.ENOENT, errno.ENOTDIR): raise del e continue if not stat.S_ISREG(s.st_mode): continue path = os.path.realpath(path) if path in real_paths: continue real_paths.add(path) if s.st_nlink > 1 and \ s.st_mode & (stat.S_ISUID | stat.S_ISGID): k = (s.st_dev, s.st_ino) inode_map.setdefault(k, []).append((path, s)) suspicious_hardlinks = [] for path_list in inode_map.itervalues(): path, s = path_list[0] if len(path_list) == s.st_nlink: # All hardlinks seem to be owned by this package. continue suspicious_hardlinks.append(path_list) if not suspicious_hardlinks: return 0 msg = [] msg.append("suid/sgid file(s) " + \ "with suspicious hardlink(s):") msg.append("") for path_list in suspicious_hardlinks: for path, s in path_list: msg.append("\t%s" % path) msg.append("") msg.append("See the Gentoo Security Handbook " + \ "guide for advice on how to proceed.") self._eerror("preinst", msg) return 1 def _eerror(self, phase, lines): from portage.elog.messages import eerror as _eerror if self._scheduler is None: for l in lines: _eerror(l, phase=phase, key=self.settings.mycpv) else: self._scheduler.dblinkElog(self, phase, _eerror, lines) def treewalk(self, srcroot, destroot, inforoot, myebuild, cleanup=0, mydbapi=None, prev_mtimes=None): """ This function does the following: calls self._preserve_libs if FEATURES=preserve-libs calls self._collision_protect if FEATURES=collision-protect calls doebuild(mydo=pkg_preinst) Merges the package to the livefs unmerges old version (if required) calls doebuild(mydo=pkg_postinst) calls env_update calls elog_process @param srcroot: Typically this is ${D} @type srcroot: String (Path) @param destroot: Path to merge to (usually ${ROOT}) @type destroot: String (Path) @param inforoot: root of the vardb entry ? @type inforoot: String (Path) @param myebuild: path to the ebuild that we are processing @type myebuild: String (Path) @param mydbapi: dbapi which is handed to doebuild. @type mydbapi: portdbapi instance @param prev_mtimes: { Filename:mtime } mapping for env_update @type prev_mtimes: Dictionary @rtype: Boolean @returns: 1. 0 on success 2. 1 on failure secondhand is a list of symlinks that have been skipped due to their target not existing; we will merge these symlinks at a later time. """ showMessage = self._display_merge scheduler = self._scheduler srcroot = normalize_path(srcroot).rstrip(os.path.sep) + os.path.sep destroot = normalize_path(destroot).rstrip(os.path.sep) + os.path.sep if not os.path.isdir(srcroot): showMessage("!!! Directory Not Found: D='%s'\n" % srcroot, level=logging.ERROR, noiselevel=-1) return 1 inforoot_slot_file = os.path.join(inforoot, "SLOT") slot = None try: f = open(inforoot_slot_file) try: slot = f.read().strip() finally: f.close() except EnvironmentError, e: if e.errno != errno.ENOENT: raise del e if slot is None: slot = "" def eerror(lines): self._eerror("preinst", lines) if slot != self.settings["SLOT"]: showMessage("!!! WARNING: Expected SLOT='%s', got '%s'\n" % \ (self.settings["SLOT"], slot), level=logging.WARN) if not os.path.exists(self.dbcatdir): os.makedirs(self.dbcatdir) otherversions = [] for v in self.vartree.dbapi.cp_list(self.mysplit[0]): otherversions.append(v.split("/")[1]) # filter any old-style virtual matches slot_matches = [cpv for cpv in self.vartree.dbapi.match( "%s:%s" % (cpv_getkey(self.mycpv), slot)) \ if cpv_getkey(cpv) == cpv_getkey(self.mycpv)] if self.mycpv not in slot_matches and \ self.vartree.dbapi.cpv_exists(self.mycpv): # handle multislot or unapplied slotmove slot_matches.append(self.mycpv) others_in_slot = [] from portage import config for cur_cpv in slot_matches: # Clone the config in case one of these has to be unmerged since # we need it to have private ${T} etc... for things like elog. others_in_slot.append(dblink(self.cat, catsplit(cur_cpv)[1], self.vartree.root, config(clone=self.settings), vartree=self.vartree, treetype="vartree", scheduler=self._scheduler)) retval = self._security_check(others_in_slot) if retval: return retval if slot_matches: # Used by self.isprotected(). max_dblnk = None max_counter = -1 for dblnk in others_in_slot: cur_counter = self.vartree.dbapi.cpv_counter(dblnk.mycpv) if cur_counter > max_counter: max_counter = cur_counter max_dblnk = dblnk self._installed_instance = max_dblnk myfilelist = [] mylinklist = [] def onerror(e): raise for parent, dirs, files in os.walk(srcroot, onerror=onerror): for f in files: file_path = os.path.join(parent, f) file_mode = os.lstat(file_path).st_mode if stat.S_ISREG(file_mode): myfilelist.append(file_path[len(srcroot):]) elif stat.S_ISLNK(file_mode): # Note: os.walk puts symlinks to directories in the "dirs" # list and it does not traverse them since that could lead # to an infinite recursion loop. mylinklist.append(file_path[len(srcroot):]) # If there are no files to merge, and an installed package in the same # slot has files, it probably means that something went wrong. if self.settings.get("PORTAGE_PACKAGE_EMPTY_ABORT") == "1" and \ not myfilelist and not mylinklist and others_in_slot: installed_files = None for other_dblink in others_in_slot: installed_files = other_dblink.getcontents() if not installed_files: continue from textwrap import wrap wrap_width = 72 msg = [] d = ( self.mycpv, other_dblink.mycpv ) msg.extend(wrap(("The '%s' package will not install " + \ "any files, but the currently installed '%s'" + \ " package has the following files: ") % d, wrap_width)) msg.append("") msg.extend(sorted(installed_files)) msg.append("") msg.append("package %s NOT merged" % self.mycpv) msg.append("") msg.extend(wrap( ("Manually run `emerge --unmerge =%s` " % \ other_dblink.mycpv) + "if you really want to " + \ "remove the above files. Set " + \ "PORTAGE_PACKAGE_EMPTY_ABORT=\"0\" in " + \ "/etc/make.conf if you do not want to " + \ "abort in cases like this.", wrap_width)) eerror(msg) if installed_files: return 1 # check for package collisions blockers = None if self._blockers is not None: # This is only supposed to be called when # the vdb is locked, like it is here. blockers = self._blockers() if blockers is None: blockers = [] collisions, plib_collisions = \ self._collision_protect(srcroot, destroot, others_in_slot + blockers, myfilelist + mylinklist) # Make sure the ebuild environment is initialized and that ${T}/elog # exists for logging of collision-protect eerror messages. if myebuild is None: myebuild = os.path.join(inforoot, self.pkg + ".ebuild") doebuild_environment(myebuild, "preinst", destroot, self.settings, 0, 0, mydbapi) prepare_build_dirs(destroot, self.settings, cleanup) if collisions: collision_protect = "collision-protect" in self.settings.features protect_owned = "protect-owned" in self.settings.features msg = "This package will overwrite one or more files that" + \ " may belong to other packages (see list below)." if not (collision_protect or protect_owned): msg += " Add either \"collision-protect\" or" + \ " \"protect-owned\" to FEATURES in" + \ " make.conf if you would like the merge to abort" + \ " in cases like this. See the make.conf man page for" + \ " more information about these features." if self.settings.get("PORTAGE_QUIET") != "1": msg += " You can use a command such as" + \ " `portageq owners / ` to identify the" + \ " installed package that owns a file. If portageq" + \ " reports that only one package owns a file then do NOT" + \ " file a bug report. A bug report is only useful if it" + \ " identifies at least two or more packages that are known" + \ " to install the same file(s)." + \ " If a collision occurs and you" + \ " can not explain where the file came from then you" + \ " should simply ignore the collision since there is not" + \ " enough information to determine if a real problem" + \ " exists. Please do NOT file a bug report at" + \ " http://bugs.gentoo.org unless you report exactly which" + \ " two packages install the same file(s). Once again," + \ " please do NOT file a bug report unless you have" + \ " completely understood the above message." self.settings["EBUILD_PHASE"] = "preinst" from textwrap import wrap msg = wrap(msg, 70) if collision_protect: msg.append("") msg.append("package %s NOT merged" % self.settings.mycpv) msg.append("") msg.append("Detected file collision(s):") msg.append("") for f in collisions: msg.append("\t%s" % \ os.path.join(destroot, f.lstrip(os.path.sep))) eerror(msg) msg = [] msg.append("") msg.append("Searching all installed" + \ " packages for file collisions...") msg.append("") msg.append("Press Ctrl-C to Stop") msg.append("") eerror(msg) owners = self.vartree.dbapi._owners.get_owners(collisions) self.vartree.dbapi.flush_cache() for pkg, owned_files in owners.iteritems(): cpv = pkg.mycpv msg = [] msg.append("%s" % cpv) for f in sorted(owned_files): msg.append("\t%s" % os.path.join(destroot, f.lstrip(os.path.sep))) msg.append("") eerror(msg) if not owners: eerror(["None of the installed" + \ " packages claim the file(s).", ""]) # The explanation about the collision and how to solve # it may not be visible via a scrollback buffer, especially # if the number of file collisions is large. Therefore, # show a summary at the end. if collision_protect: msg = "Package '%s' NOT merged due to file collisions." % \ self.settings.mycpv elif protect_owned and owners: msg = "Package '%s' NOT merged due to file collisions." % \ self.settings.mycpv else: msg = "Package '%s' merged despite file collisions." % \ self.settings.mycpv msg += " If necessary, refer to your elog " + \ "messages for the whole content of the above message." eerror(wrap(msg, 70)) if collision_protect or (protect_owned and owners): return 1 # The merge process may move files out of the image directory, # which causes invalidation of the .installed flag. try: os.unlink(os.path.join( os.path.dirname(normalize_path(srcroot)), ".installed")) except OSError, e: if e.errno != errno.ENOENT: raise del e self.dbdir = self.dbtmpdir self.delete() ensure_dirs(self.dbtmpdir) # run preinst script if scheduler is None: showMessage(">>> Merging %s to %s\n" % (self.mycpv, destroot)) a = doebuild(myebuild, "preinst", destroot, self.settings, use_cache=0, tree=self.treetype, mydbapi=mydbapi, vartree=self.vartree) else: a = scheduler.dblinkEbuildPhase( self, mydbapi, myebuild, "preinst") # XXX: Decide how to handle failures here. if a != os.EX_OK: showMessage("!!! FAILED preinst: "+str(a)+"\n", level=logging.ERROR, noiselevel=-1) return a # copy "info" files (like SLOT, CFLAGS, etc.) into the database for x in os.listdir(inforoot): self.copyfile(inforoot+"/"+x) # write local package counter for recording counter = self.vartree.dbapi.counter_tick(self.myroot, mycpv=self.mycpv) lcfile = open(os.path.join(self.dbtmpdir, "COUNTER"),"w") lcfile.write(str(counter)) lcfile.close() # open CONTENTS file (possibly overwriting old one) for recording outfile = open(os.path.join(self.dbtmpdir, "CONTENTS"),"w") self.updateprotect() #if we have a file containing previously-merged config file md5sums, grab it. conf_mem_file = os.path.join(destroot, CONFIG_MEMORY_FILE) cfgfiledict = grabdict(conf_mem_file) cfgfiledict_orig = cfgfiledict.copy() if "NOCONFMEM" in self.settings: cfgfiledict["IGNORE"]=1 else: cfgfiledict["IGNORE"]=0 # Always behave like --noconfmem is enabled for downgrades # so that people who don't know about this option are less # likely to get confused when doing upgrade/downgrade cycles. pv_split = catpkgsplit(self.mycpv)[1:] for other in others_in_slot: if pkgcmp(pv_split, catpkgsplit(other.mycpv)[1:]) < 0: cfgfiledict["IGNORE"] = 1 break # Don't bump mtimes on merge since some application require # preservation of timestamps. This means that the unmerge phase must # check to see if file belongs to an installed instance in the same # slot. mymtime = None # set umask to 0 for merging; back up umask, save old one in prevmask (since this is a global change) prevmask = os.umask(0) secondhand = [] # we do a first merge; this will recurse through all files in our srcroot but also build up a # "second hand" of symlinks to merge later if self.mergeme(srcroot, destroot, outfile, secondhand, "", cfgfiledict, mymtime): return 1 # now, it's time for dealing our second hand; we'll loop until we can't merge anymore. The rest are # broken symlinks. We'll merge them too. lastlen = 0 while len(secondhand) and len(secondhand)!=lastlen: # clear the thirdhand. Anything from our second hand that # couldn't get merged will be added to thirdhand. thirdhand = [] if self.mergeme(srcroot, destroot, outfile, thirdhand, secondhand, cfgfiledict, mymtime): return 1 #swap hands lastlen = len(secondhand) # our thirdhand now becomes our secondhand. It's ok to throw # away secondhand since thirdhand contains all the stuff that # couldn't be merged. secondhand = thirdhand if len(secondhand): # force merge of remaining symlinks (broken or circular; oh well) if self.mergeme(srcroot, destroot, outfile, None, secondhand, cfgfiledict, mymtime): return 1 #restore umask os.umask(prevmask) #if we opened it, close it outfile.flush() outfile.close() # write out our collection of md5sums cfgfiledict.pop("IGNORE", None) if cfgfiledict != cfgfiledict_orig: ensure_dirs(os.path.dirname(conf_mem_file), gid=portage_gid, mode=02750, mask=02) writedict(cfgfiledict, conf_mem_file) # These caches are populated during collision-protect and the data # they contain is now invalid. It's very important to invalidate # the contents_inodes cache so that FEATURES=unmerge-orphans # doesn't unmerge anything that belongs to this package that has # just been merged. for dblnk in others_in_slot: dblnk._clear_contents_cache() self._clear_contents_cache() linkmap = self.vartree.dbapi.linkmap self._linkmap_rebuild(include_file=os.path.join(inforoot, linkmap._needed_aux_key)) # Preserve old libs if they are still in use preserve_paths = self._find_libs_to_preserve() if preserve_paths: self._add_preserve_libs_to_contents(preserve_paths) # If portage is reinstalling itself, remove the old # version now since we want to use the temporary # PORTAGE_BIN_PATH that will be removed when we return. reinstall_self = False if self.myroot == "/" and \ match_from_list(PORTAGE_PACKAGE_ATOM, [self.mycpv]): reinstall_self = True autoclean = self.settings.get("AUTOCLEAN", "yes") == "yes" others_in_slot.append(self) # self has just been merged for dblnk in list(others_in_slot): if dblnk is self: continue if not (autoclean or dblnk.mycpv == self.mycpv or reinstall_self): continue showMessage(">>> Safely unmerging already-installed instance...\n") others_in_slot.remove(dblnk) # dblnk will unmerge itself now dblnk._linkmap_broken = self._linkmap_broken dblnk.unmerge(trimworld=0, ldpath_mtimes=prev_mtimes, others_in_slot=others_in_slot) # TODO: Check status and abort if necessary. dblnk.delete() showMessage(">>> Original instance of package unmerged safely.\n") if len(others_in_slot) > 1: from portage.output import colorize showMessage(colorize("WARN", "WARNING:") + " AUTOCLEAN is disabled. This can cause serious" + " problems due to overlapping packages.\n", level=logging.WARN, noiselevel=-1) # We hold both directory locks. self.dbdir = self.dbpkgdir self.delete() _movefile(self.dbtmpdir, self.dbpkgdir, mysettings=self.settings) # keep track of the libs we preserved if preserve_paths: self.vartree.dbapi.plib_registry.register(self.mycpv, slot, counter, sorted(preserve_paths)) # Check for file collisions with blocking packages # and remove any colliding files from their CONTENTS # since they now belong to this package. self._clear_contents_cache() contents = self.getcontents() destroot_len = len(destroot) - 1 for blocker in blockers: self.vartree.dbapi.removeFromContents(blocker, iter(contents), relative_paths=False) # Unregister any preserved libs that this package has overwritten # and update the contents of the packages that owned them. plib_registry = self.vartree.dbapi.plib_registry plib_dict = plib_registry.getPreservedLibs() for cpv, paths in plib_collisions.iteritems(): if cpv not in plib_dict: continue if cpv == self.mycpv: continue try: slot, counter = self.vartree.dbapi.aux_get( cpv, ["SLOT", "COUNTER"]) except KeyError: continue remaining = [f for f in plib_dict[cpv] if f not in paths] plib_registry.register(cpv, slot, counter, remaining) self.vartree.dbapi.removeFromContents(cpv, paths) self.vartree.dbapi._add(self) contents = self.getcontents() #do postinst script self.settings["PORTAGE_UPDATE_ENV"] = \ os.path.join(self.dbpkgdir, "environment.bz2") self.settings.backup_changes("PORTAGE_UPDATE_ENV") try: if scheduler is None: a = doebuild(myebuild, "postinst", destroot, self.settings, use_cache=0, tree=self.treetype, mydbapi=mydbapi, vartree=self.vartree) if a == os.EX_OK: showMessage(">>> %s %s\n" % (self.mycpv, "merged.")) else: a = scheduler.dblinkEbuildPhase( self, mydbapi, myebuild, "postinst") finally: self.settings.pop("PORTAGE_UPDATE_ENV", None) # XXX: Decide how to handle failures here. if a != os.EX_OK: showMessage("!!! FAILED postinst: "+str(a)+"\n", level=logging.ERROR, noiselevel=-1) return a downgrade = False for v in otherversions: if pkgcmp(catpkgsplit(self.pkg)[1:], catpkgsplit(v)[1:]) < 0: downgrade = True #update environment settings, library paths. DO NOT change symlinks. env_update(makelinks=(not downgrade), target_root=self.settings["ROOT"], prev_mtimes=prev_mtimes, contents=contents, env=self.settings.environ(), writemsg_level=self._display_merge) # For gcc upgrades, preserved libs have to be removed after the # the library path has been updated. self._linkmap_rebuild() cpv_lib_map = self._find_unused_preserved_libs() if cpv_lib_map: self._remove_preserved_libs(cpv_lib_map) for cpv, removed in cpv_lib_map.iteritems(): if not self.vartree.dbapi.cpv_exists(cpv): continue self.vartree.dbapi.removeFromContents(cpv, removed) return os.EX_OK def mergeme(self, srcroot, destroot, outfile, secondhand, stufftomerge, cfgfiledict, thismtime): """ This function handles actual merging of the package contents to the livefs. It also handles config protection. @param srcroot: Where are we copying files from (usually ${D}) @type srcroot: String (Path) @param destroot: Typically ${ROOT} @type destroot: String (Path) @param outfile: File to log operations to @type outfile: File Object @param secondhand: A set of items to merge in pass two (usually or symlinks that point to non-existing files that may get merged later) @type secondhand: List @param stufftomerge: Either a diretory to merge, or a list of items. @type stufftomerge: String or List @param cfgfiledict: { File:mtime } mapping for config_protected files @type cfgfiledict: Dictionary @param thismtime: The current time (typically long(time.time()) @type thismtime: Long @rtype: None or Boolean @returns: 1. True on failure 2. None otherwise """ showMessage = self._display_merge writemsg = self._display_merge scheduler = self._scheduler from os.path import sep, join srcroot = normalize_path(srcroot).rstrip(sep) + sep destroot = normalize_path(destroot).rstrip(sep) + sep # this is supposed to merge a list of files. There will be 2 forms of argument passing. if isinstance(stufftomerge, basestring): #A directory is specified. Figure out protection paths, listdir() it and process it. mergelist = os.listdir(join(srcroot, stufftomerge)) offset = stufftomerge else: mergelist = stufftomerge offset = "" for i, x in enumerate(mergelist): if scheduler is not None and \ 0 == i % self._file_merge_yield_interval: scheduler.scheduleYield() mysrc = join(srcroot, offset, x) mydest = join(destroot, offset, x) # myrealdest is mydest without the $ROOT prefix (makes a difference if ROOT!="/") myrealdest = join(sep, offset, x) # stat file once, test using S_* macros many times (faster that way) try: mystat = os.lstat(mysrc) except OSError, e: writemsg("\n") writemsg(red("!!! ERROR: There appears to be ")+bold("FILE SYSTEM CORRUPTION.")+red(" A file that is listed\n")) writemsg(red("!!! as existing is not capable of being stat'd. If you are using an\n")) writemsg(red("!!! experimental kernel, please boot into a stable one, force an fsck,\n")) writemsg(red("!!! and ensure your filesystem is in a sane state. ")+bold("'shutdown -Fr now'\n")) writemsg(red("!!! File: ")+str(mysrc)+"\n", noiselevel=-1) writemsg(red("!!! Error: ")+str(e)+"\n", noiselevel=-1) return 1 mymode = mystat[stat.ST_MODE] # handy variables; mydest is the target object on the live filesystems; # mysrc is the source object in the temporary install dir try: mydstat = os.lstat(mydest) mydmode = mydstat.st_mode except OSError, e: if e.errno != errno.ENOENT: raise del e #dest file doesn't exist mydstat = None mydmode = None if stat.S_ISLNK(mymode): # we are merging a symbolic link myabsto = abssymlink(mysrc) if myabsto.startswith(srcroot): myabsto = myabsto[len(srcroot):] myabsto = myabsto.lstrip(sep) myto = os.readlink(mysrc) if self.settings and self.settings["D"]: if myto.startswith(self.settings["D"]): myto = myto[len(self.settings["D"]):] # myrealto contains the path of the real file to which this symlink points. # we can simply test for existence of this file to see if the target has been merged yet myrealto = normalize_path(os.path.join(destroot, myabsto)) if mydmode!=None: #destination exists if not stat.S_ISLNK(mydmode): if stat.S_ISDIR(mydmode): # directory in the way: we can't merge a symlink over a directory # we won't merge this, continue with next file... continue if os.path.exists(mysrc) and stat.S_ISDIR(os.stat(mysrc)[stat.ST_MODE]): # Kill file blocking installation of symlink to dir #71787 pass elif self.isprotected(mydest): # Use md5 of the target in ${D} if it exists... try: newmd5 = perform_md5(join(srcroot, myabsto)) except FileNotFound: # Maybe the target is merged already. try: newmd5 = perform_md5(myrealto) except FileNotFound: newmd5 = None mydest = new_protect_filename(mydest, newmd5=newmd5) # if secondhand is None it means we're operating in "force" mode and should not create a second hand. if (secondhand != None) and (not os.path.exists(myrealto)): # either the target directory doesn't exist yet or the target file doesn't exist -- or # the target is a broken symlink. We will add this file to our "second hand" and merge # it later. secondhand.append(mysrc[len(srcroot):]) continue # unlinking no longer necessary; "movefile" will overwrite symlinks atomically and correctly mymtime = movefile(mysrc, mydest, newmtime=thismtime, sstat=mystat, mysettings=self.settings) if mymtime != None: showMessage(">>> %s -> %s\n" % (mydest, myto)) outfile.write("sym "+myrealdest+" -> "+myto+" "+str(mymtime)+"\n") else: showMessage("!!! Failed to move file.\n", level=logging.ERROR, noiselevel=-1) showMessage("!!! %s -> %s\n" % (mydest, myto), level=logging.ERROR, noiselevel=-1) return 1 elif stat.S_ISDIR(mymode): # we are merging a directory if mydmode != None: # destination exists if bsd_chflags: # Save then clear flags on dest. dflags = mydstat.st_flags if dflags != 0: bsd_chflags.lchflags(mydest, 0) if not os.access(mydest, os.W_OK): pkgstuff = pkgsplit(self.pkg) writemsg("\n!!! Cannot write to '"+mydest+"'.\n", noiselevel=-1) writemsg("!!! Please check permissions and directories for broken symlinks.\n") writemsg("!!! You may start the merge process again by using ebuild:\n") writemsg("!!! ebuild "+self.settings["PORTDIR"]+"/"+self.cat+"/"+pkgstuff[0]+"/"+self.pkg+".ebuild merge\n") writemsg("!!! And finish by running this: env-update\n\n") return 1 if stat.S_ISLNK(mydmode) or stat.S_ISDIR(mydmode): # a symlink to an existing directory will work for us; keep it: showMessage("--- %s/\n" % mydest) if bsd_chflags: bsd_chflags.lchflags(mydest, dflags) else: # a non-directory and non-symlink-to-directory. Won't work for us. Move out of the way. if movefile(mydest, mydest+".backup", mysettings=self.settings) is None: return 1 showMessage("bak %s %s.backup\n" % (mydest, mydest), level=logging.ERROR, noiselevel=-1) #now create our directory if self.settings.selinux_enabled(): import selinux sid = selinux.get_sid(mysrc) selinux.secure_mkdir(mydest,sid) else: os.mkdir(mydest) if bsd_chflags: bsd_chflags.lchflags(mydest, dflags) os.chmod(mydest, mystat[0]) os.chown(mydest, mystat[4], mystat[5]) showMessage(">>> %s/\n" % mydest) else: #destination doesn't exist if self.settings.selinux_enabled(): import selinux sid = selinux.get_sid(mysrc) selinux.secure_mkdir(mydest, sid) else: os.mkdir(mydest) os.chmod(mydest, mystat[0]) os.chown(mydest, mystat[4], mystat[5]) showMessage(">>> %s/\n" % mydest) outfile.write("dir "+myrealdest+"\n") # recurse and merge this directory if self.mergeme(srcroot, destroot, outfile, secondhand, join(offset, x), cfgfiledict, thismtime): return 1 elif stat.S_ISREG(mymode): # we are merging a regular file mymd5 = perform_md5(mysrc, calc_prelink=1) # calculate config file protection stuff mydestdir = os.path.dirname(mydest) moveme = 1 zing = "!!!" mymtime = None if mydmode != None: # destination file exists if stat.S_ISDIR(mydmode): # install of destination is blocked by an existing directory with the same name moveme = 0 showMessage("!!! %s\n" % mydest, level=logging.ERROR, noiselevel=-1) elif stat.S_ISREG(mydmode) or (stat.S_ISLNK(mydmode) and os.path.exists(mydest) and stat.S_ISREG(os.stat(mydest)[stat.ST_MODE])): cfgprot = 0 # install of destination is blocked by an existing regular file, # or by a symlink to an existing regular file; # now, config file management may come into play. # we only need to tweak mydest if cfg file management is in play. if self.isprotected(mydest): # we have a protection path; enable config file management. destmd5 = perform_md5(mydest, calc_prelink=1) if mymd5 == destmd5: #file already in place; simply update mtimes of destination moveme = 1 else: if mymd5 == cfgfiledict.get(myrealdest, [None])[0]: """ An identical update has previously been merged. Skip it unless the user has chosen --noconfmem.""" moveme = cfgfiledict["IGNORE"] cfgprot = cfgfiledict["IGNORE"] if not moveme: zing = "---" mymtime = long(mystat.st_mtime) else: moveme = 1 cfgprot = 1 if moveme: # Merging a new file, so update confmem. cfgfiledict[myrealdest] = [mymd5] elif destmd5 == cfgfiledict.get(myrealdest, [None])[0]: """A previously remembered update has been accepted, so it is removed from confmem.""" del cfgfiledict[myrealdest] if cfgprot: mydest = new_protect_filename(mydest, newmd5=mymd5) # whether config protection or not, we merge the new file the # same way. Unless moveme=0 (blocking directory) if moveme: mymtime = movefile(mysrc, mydest, newmtime=thismtime, sstat=mystat, mysettings=self.settings) if mymtime is None: return 1 zing = ">>>" if mymtime != None: outfile.write("obj "+myrealdest+" "+mymd5+" "+str(mymtime)+"\n") showMessage("%s %s\n" % (zing,mydest)) else: # we are merging a fifo or device node zing = "!!!" if mydmode is None: # destination doesn't exist if movefile(mysrc, mydest, newmtime=thismtime, sstat=mystat, mysettings=self.settings) != None: zing = ">>>" else: return 1 if stat.S_ISFIFO(mymode): outfile.write("fif %s\n" % myrealdest) else: outfile.write("dev %s\n" % myrealdest) showMessage(zing + " " + mydest + "\n") def merge(self, mergeroot, inforoot, myroot, myebuild=None, cleanup=0, mydbapi=None, prev_mtimes=None): """ If portage is reinstalling itself, create temporary copies of PORTAGE_BIN_PATH and PORTAGE_PYM_PATH in order to avoid relying on the new versions which may be incompatible. Register an atexit hook to clean up the temporary directories. Pre-load elog modules here since we won't be able to later if they get unmerged (happens when namespace changes). """ if self.vartree.dbapi._categories is not None: self.vartree.dbapi._categories = None if self.myroot == "/" and \ match_from_list(PORTAGE_PACKAGE_ATOM, [self.mycpv]) and \ not self.vartree.dbapi.cpv_exists(self.mycpv): settings = self.settings base_path_orig = os.path.dirname(settings["PORTAGE_BIN_PATH"]) from tempfile import mkdtemp # Make the temp directory inside PORTAGE_TMPDIR since, unlike # /tmp, it can't be mounted with the "noexec" option. base_path_tmp = mkdtemp("", "._portage_reinstall_.", settings["PORTAGE_TMPDIR"]) from portage.process import atexit_register atexit_register(shutil.rmtree, base_path_tmp) dir_perms = 0755 for subdir in "bin", "pym": var_name = "PORTAGE_%s_PATH" % subdir.upper() var_orig = settings[var_name] var_new = os.path.join(base_path_tmp, subdir) settings[var_name] = var_new settings.backup_changes(var_name) shutil.copytree(var_orig, var_new, symlinks=True) os.chmod(var_new, dir_perms) os.chmod(base_path_tmp, dir_perms) # This serves so pre-load the modules. elog_process(self.mycpv, self.settings, phasefilter=filter_mergephases) return self._merge(mergeroot, inforoot, myroot, myebuild=myebuild, cleanup=cleanup, mydbapi=mydbapi, prev_mtimes=prev_mtimes) def _merge(self, mergeroot, inforoot, myroot, myebuild=None, cleanup=0, mydbapi=None, prev_mtimes=None): retval = -1 self.lockdb() try: self.vartree.dbapi.plib_registry.load() self.vartree.dbapi.plib_registry.pruneNonExisting() retval = self.treewalk(mergeroot, myroot, inforoot, myebuild, cleanup=cleanup, mydbapi=mydbapi, prev_mtimes=prev_mtimes) # Process ebuild logfiles elog_process(self.mycpv, self.settings, phasefilter=filter_mergephases) if retval == os.EX_OK and "noclean" not in self.settings.features: if myebuild is None: myebuild = os.path.join(inforoot, self.pkg + ".ebuild") if self._scheduler is None: doebuild(myebuild, "clean", myroot, self.settings, tree=self.treetype, mydbapi=mydbapi, vartree=self.vartree) else: self._scheduler.dblinkEbuildPhase( self, mydbapi, myebuild, "clean") finally: self.vartree.dbapi.linkmap._clear_cache() self.unlockdb() return retval def getstring(self,name): "returns contents of a file with whitespace converted to spaces" if not os.path.exists(self.dbdir+"/"+name): return "" myfile = open(self.dbdir+"/"+name,"r") mydata = myfile.read().split() myfile.close() return " ".join(mydata) def copyfile(self,fname): shutil.copyfile(fname,self.dbdir+"/"+os.path.basename(fname)) def getfile(self,fname): if not os.path.exists(self.dbdir+"/"+fname): return "" myfile = open(self.dbdir+"/"+fname,"r") mydata = myfile.read() myfile.close() return mydata def setfile(self,fname,data): write_atomic(os.path.join(self.dbdir, fname), data) def getelements(self,ename): if not os.path.exists(self.dbdir+"/"+ename): return [] myelement = open(self.dbdir+"/"+ename,"r") mylines = myelement.readlines() myreturn = [] for x in mylines: for y in x[:-1].split(): myreturn.append(y) myelement.close() return myreturn def setelements(self,mylist,ename): myelement = open(self.dbdir+"/"+ename,"w") for x in mylist: myelement.write(x+"\n") myelement.close() def isregular(self): "Is this a regular package (does it have a CATEGORY file? A dblink can be virtual *and* regular)" return os.path.exists(os.path.join(self.dbdir, "CATEGORY")) def write_contents(contents, root, f): """ Write contents to any file like object. The file will be left open. """ root_len = len(root) - 1 for filename in sorted(contents): entry_data = contents[filename] entry_type = entry_data[0] relative_filename = filename[root_len:] if entry_type == "obj": entry_type, mtime, md5sum = entry_data line = "%s %s %s %s\n" % \ (entry_type, relative_filename, md5sum, mtime) elif entry_type == "sym": entry_type, mtime, link = entry_data line = "%s %s -> %s %s\n" % \ (entry_type, relative_filename, link, mtime) else: # dir, dev, fif line = "%s %s\n" % (entry_type, relative_filename) f.write(line) def tar_contents(contents, root, tar, protect=None, onProgress=None): from portage.util import normalize_path import tarfile root = normalize_path(root).rstrip(os.path.sep) + os.path.sep id_strings = {} maxval = len(contents) curval = 0 if onProgress: onProgress(maxval, 0) paths = contents.keys() paths.sort() for path in paths: curval += 1 try: lst = os.lstat(path) except OSError, e: if e.errno != errno.ENOENT: raise del e if onProgress: onProgress(maxval, curval) continue contents_type = contents[path][0] if path.startswith(root): arcname = path[len(root):] else: raise ValueError("invalid root argument: '%s'" % root) live_path = path if 'dir' == contents_type and \ not stat.S_ISDIR(lst.st_mode) and \ os.path.isdir(live_path): # Even though this was a directory in the original ${D}, it exists # as a symlink to a directory in the live filesystem. It must be # recorded as a real directory in the tar file to ensure that tar # can properly extract it's children. live_path = os.path.realpath(live_path) tarinfo = tar.gettarinfo(live_path, arcname) # store numbers instead of real names like tar's --numeric-owner tarinfo.uname = id_strings.setdefault(tarinfo.uid, str(tarinfo.uid)) tarinfo.gname = id_strings.setdefault(tarinfo.gid, str(tarinfo.gid)) if stat.S_ISREG(lst.st_mode): # break hardlinks due to bug #185305 tarinfo.type = tarfile.REGTYPE if protect and protect(path): # Create an empty file as a place holder in order to avoid # potential collision-protect issues. tarinfo.size = 0 tar.addfile(tarinfo) else: f = open(path) try: tar.addfile(tarinfo, f) finally: f.close() else: tar.addfile(tarinfo) if onProgress: onProgress(maxval, curval)