pkg.py

import bz2
from collections import namedtuple
import contextlib
import gzip
import hashlib
import io
import lzma
import mmap
import os
from pathlib import Path, PurePath
import re
from shlex import quote
import shutil
import stat
import subprocess
import tempfile
import time
from urllib.parse import urljoin

try:
    import magic
    has_magic = True
except ImportError:
    has_magic = False
import rpm
from rpmlint.helpers import (byte_to_string, ENGLISH_ENVIRONMENT,
                             print_warning, pushd)
from rpmlint.pkgfile import PkgFile
import zstandard as zstd


DepInfo = namedtuple('DepInfo', ('name', 'flags', 'version'))

# 64: RPMSENSE_PREREQ is 0 with rpm 4.4..4.7, we want 64 here in order
# to do the right thing with those versions and packages built with other
# rpm versions
PREREQ_FLAG = (rpm.RPMSENSE_PREREQ or 64) | rpm.RPMSENSE_SCRIPT_PRE | \
    rpm.RPMSENSE_SCRIPT_POST | rpm.RPMSENSE_SCRIPT_PREUN | \
    rpm.RPMSENSE_SCRIPT_POSTUN

SCRIPT_TAGS = [
    (rpm.RPMTAG_PREIN, rpm.RPMTAG_PREINPROG, '%pre'),
    (rpm.RPMTAG_POSTIN, rpm.RPMTAG_POSTINPROG, '%post'),
    (rpm.RPMTAG_PREUN, rpm.RPMTAG_PREUNPROG, '%preun'),
    (rpm.RPMTAG_POSTUN, rpm.RPMTAG_POSTUNPROG, '%postun'),
    (rpm.RPMTAG_TRIGGERSCRIPTS, rpm.RPMTAG_TRIGGERSCRIPTPROG, '%trigger'),
    (rpm.RPMTAG_PRETRANS, rpm.RPMTAG_PRETRANSPROG, '%pretrans'),
    (rpm.RPMTAG_POSTTRANS, rpm.RPMTAG_POSTTRANSPROG, '%posttrans'),
    (rpm.RPMTAG_VERIFYSCRIPT, rpm.RPMTAG_VERIFYSCRIPTPROG, '%verifyscript'),
    # file triggers: rpm >= 4.12.90
    (getattr(rpm, 'RPMTAG_FILETRIGGERSCRIPTS', 5066),
     getattr(rpm, 'RPMTAG_FILETRIGGERSCRIPTPROG', 5067),
     '%filetrigger'),
    (getattr(rpm, 'RPMTAG_TRANSFILETRIGGERSCRIPTS', 5076),
     getattr(rpm, 'RPMTAG_TRANSFILETRIGGERSCRIPTPROG', 5077),
     '%transfiletrigger'),
]

RPM_SCRIPTLETS = ('pre', 'post', 'preun', 'postun', 'pretrans', 'posttrans',
                  'trigger', 'triggerin', 'triggerprein', 'triggerun',
                  'triggerpostun', 'verifyscript', 'filetriggerin',
                  'filetrigger', 'filetriggerun', 'filetriggerpostun',
                  'transfiletriggerin', 'transfiletrigger',
                  'transfiletriggerun', 'transfiletriggerun',
                  'transfiletriggerpostun')


gzip_regex = re.compile(r'\.t?gz?$')
bz2_regex = re.compile(r'\.t?bz2?$')
xz_regex = re.compile(r'\.(t[xl]z|xz|lzma)$')
zst_regex = re.compile(r'\.zst$')


def catcmd(fname):
    """Get a 'cat' command that handles possibly compressed files."""
    fname = str(fname)
    cat = 'gzip -dcf'
    if bz2_regex.search(fname):
        cat = 'bzip2 -dcf'
    elif xz_regex.search(fname):
        cat = 'xz -dc'
    elif zst_regex.search(fname):
        cat = 'zstd -dc'
    return cat


def compression_algorithm(fname):
    """Return compression algorithm based on filename if known, None otherwise."""
    fname = str(fname)
    if gzip_regex.search(fname):
        return gzip
    elif bz2_regex.search(fname):
        return bz2
    elif xz_regex.search(fname):
        return lzma
    elif zst_regex.search(fname):
        return zstd
    else:
        return None


def is_utf8(fname):
    compression = compression_algorithm(fname)
    if compression is None:
        with open(fname, 'rb') as f:
            return is_utf8_bytestr(f.read())

    with compression.open(fname, 'rb') as f:
        try:
            return is_utf8_bytestr(f.read())
        except OSError:
            return True


def is_utf8_bytestr(s):
    """Returns True whether the given text is UTF-8.
    Due to changes in rpm, needs to handle both bytes and unicode."""
    if not isinstance(s, (bytes, str)):
        unexpected = type(s).__name__
        raise TypeError(f'Expected str/bytes, not {unexpected}')

    try:
        if isinstance(s, bytes):
            s.decode('utf-8')
    except UnicodeError:
        return False

    return True


def has_forbidden_controlchars(val):
    if isinstance(val, (str, bytes)):
        string = val
        if isinstance(val, bytes):
            val = memoryview(val)
        for c in val:
            if isinstance(c, str):
                c = ord(c)
            if c < 32 and (c not in (9, 10, 13)):
                return string
    if isinstance(val, (tuple, list)):
        for item in val:
            return has_forbidden_controlchars(item)
    return False


# from yum 3.2.27, rpmUtils.miscutils, with rpmlint modifications
def compareEVR(evr1, evr2):
    (e1, v1, r1) = evr1
    (e2, v2, r2) = evr2
    # return 1: a is newer than b
    # 0: a and b are the same version
    # -1: b is newer than a
    # rpmlint mod: don't stringify None epochs to 'None' strings
    if e1 is not None:
        e1 = str(e1)
    v1 = str(v1)
    r1 = str(r1)
    if e2 is not None:
        e2 = str(e2)
    v2 = str(v2)
    r2 = str(r2)
    rc = rpm.labelCompare((e1, v1, r1), (e2, v2, r2))
    return rc


# from yum 3.2.27, rpmUtils.miscutils, with rpmlint modifications
def rangeCompare(reqtuple, provtuple):
    """returns true if provtuple satisfies reqtuple"""
    (reqn, reqf, (reqe, reqv, reqr)) = reqtuple
    (n, f, (e, v, r)) = provtuple
    if reqn != n:
        return 0

    # unversioned satisfies everything
    if not f or not reqf:
        return 1

    # and you thought we were done having fun
    # if the requested release is left out then we have
    # to remove release from the package prco to make sure the match
    # is a success - ie: if the request is EQ foo 1:3.0.0 and we have
    # foo 1:3.0.0-15 then we have to drop the 15 so we can match
    if reqr is None:
        r = None
    # rpmlint mod: don't mess with provided Epoch, doing so breaks e.g.
    # 'Requires: foo < 1.0' should not be satisfied by 'Provides: foo = 1:0.5'
    # if reqe is None:
    #    e = None
    if reqv is None:  # just for the record if ver is None then we're going to segfault
        v = None

    # if we just require foo-version, then foo-version-* will match
    if r is None:
        reqr = None

    rc = compareEVR((e, v, r), (reqe, reqv, reqr))

    # does not match unless
    if rc >= 1:
        if reqf in ['GT', 'GE', 4, 12]:
            return 1
        if reqf in ['EQ', 8] and f in ['LE', 10, 'LT', 2]:
            return 1
        if reqf in ['LE', 'LT', 'EQ', 10, 2, 8] and f in ['LE', 'LT', 10, 2]:
            return 1

    if rc == 0:
        if reqf in ['GT', 4] and f in ['GT', 'GE', 4, 12]:
            return 1
        if reqf in ['GE', 12] and f in ['GT', 'GE', 'EQ', 'LE', 4, 12, 8, 10]:
            return 1
        if reqf in ['EQ', 8] and f in ['EQ', 'GE', 'LE', 8, 12, 10]:
            return 1
        if reqf in ['LE', 10] and f in ['EQ', 'LE', 'LT', 'GE', 8, 10, 2, 12]:
            return 1
        if reqf in ['LT', 2] and f in ['LE', 'LT', 10, 2]:
            return 1
    if rc <= -1:
        if reqf in ['GT', 'GE', 'EQ', 4, 12, 8] and f in ['GT', 'GE', 4, 12]:
            return 1
        if reqf in ['LE', 'LT', 10, 2]:
            return 1
#                if rc >= 1:
#                    if reqf in ['GT', 'GE', 4, 12]:
#                        return 1
#                if rc == 0:
#                    if reqf in ['GE', 'LE', 'EQ', 8, 10, 12]:
#                        return 1
#                if rc <= -1:
#                    if reqf in ['LT', 'LE', 2, 10]:
#                        return 1

    return 0


# from yum 3.2.23, rpmUtils.miscutils, with rpmlint modifications
def formatRequire(name, flags, evr):
    s = name

    if flags and flags & (rpm.RPMSENSE_LESS | rpm.RPMSENSE_GREATER |
                          rpm.RPMSENSE_EQUAL):
        s = s + ' '
        if flags & rpm.RPMSENSE_LESS:
            s = s + '<'
        if flags & rpm.RPMSENSE_GREATER:
            s = s + '>'
        if flags & rpm.RPMSENSE_EQUAL:
            s = s + '='
        s = f'{s} {versionToString(evr)}'
    return s


def versionToString(evr):
    if not isinstance(evr, (list, tuple)):
        # assume string
        return evr
    ret = ''
    if evr[0] is not None and evr[0] != '':
        ret += str(evr[0]) + ':'
    if evr[1] is not None:
        ret += evr[1]
        if evr[2] is not None and evr[2] != '':
            ret += '-' + evr[2]
    return ret


# from yum 3.2.23, rpmUtils.miscutils, with some rpmlint modifications
def stringToVersion(verstring):
    if verstring in (None, ''):
        return (None, None, None)
    epoch = None
    i = verstring.find(':')
    if i != -1:
        with contextlib.suppress(ValueError):
            # garbage in epoch, ignore it
            epoch = int(verstring[:i])
    i += 1
    j = verstring.find('-', i)
    if j != -1:
        if verstring[i:j] == '':
            version = None
        else:
            version = verstring[i:j]
        release = verstring[j + 1:]
    else:
        if verstring[i:] == '':
            version = None
        else:
            version = verstring[i:]
        release = None
    return (epoch, version, release)


def parse_deps(line):
    """
    Parse provides/requires/conflicts/obsoletes line to list of
    (name, flags, (epoch, version, release)) tuples.
    """

    prcos = []
    tokens = re.split(r'[\s,]+', line.strip())

    # Drop line continuation backslash in multiline macro definition (for
    # spec file parsing), e.g.
    # [...] \
    # Obsoletes: foo-%1 <= 1.0.0 \
    # [...] \
    # (yes, this is an ugly hack and we probably have other problems with
    #  multiline macro definitions elsewhere...)
    if tokens[-1] == '\\':
        del tokens[-1]

    prco = []
    while tokens:
        token = tokens.pop(0)
        if not token:
            # skip empty tokens
            continue

        plen = len(prco)

        if plen == 0:
            prco.append(token)

        elif plen == 1:
            flags = 0
            if token[0] in ('=', '<', '<=', '>', '>='):
                # versioned, flags
                if '=' in token:
                    flags |= rpm.RPMSENSE_EQUAL
                if '<' in token:
                    flags |= rpm.RPMSENSE_LESS
                if '>' in token:
                    flags |= rpm.RPMSENSE_GREATER
                prco.append(flags)
            else:
                # no flags following name, treat as unversioned, add and reset
                prco.extend((flags, (None, None, None)))
                prcos.append(tuple(prco))
                prco = [token]

        elif plen == 2:
            # last token of versioned one, add and reset
            prco.append(stringToVersion(token))
            prcos.append(tuple(prco))
            prco = []

    plen = len(prco)
    if plen:
        if plen == 1:
            prco.extend((0, (None, None, None)))
        elif plen == 2:
            prco.append((None, None, None))
        prcos.append(tuple(prco))

    return prcos


def _get_magic_libmagic(path):
    return magic.detect_from_filename(path).name


def _get_magic_python_magic(path):
    return magic.from_file(path)


def get_magic(path):
    # python-magic & libmagic compatibility code
    # https://github.com/ahupp/python-magic/blob/master/COMPAT.md
    detect_magic = _get_magic_python_magic
    if not hasattr(magic, 'from_file'):
        # libmagic python bindings
        detect_magic = _get_magic_libmagic

    try:
        return detect_magic(path)
    except (ValueError, FileNotFoundError):
        return ''


# classes representing package

class AbstractPkg:
    def cleanup(self):
        pass

    def _calc_magic(self, pkgfile):
        magic = pkgfile.magic
        if not magic:
            if stat.S_ISDIR(pkgfile.mode):
                magic = 'directory'
            elif stat.S_ISLNK(pkgfile.mode):
                magic = "symbolic link to `%s'" % pkgfile.linkto
            elif not pkgfile.size:
                magic = 'empty'
        if not magic and not pkgfile.is_ghost and has_magic:
            start = time.monotonic()
            magic = get_magic(pkgfile.path)
            self.timers['libmagic'] += time.monotonic() - start
        if magic is None or Pkg._magic_from_compressed_re.search(magic):
            # Discard magic from inside compressed files ('file -z')
            # until PkgFile gets decompression support.  We may get
            # such magic strings from package headers already now;
            # for example Fedora's rpmbuild as of F-11's 4.7.1 is
            # patched so it generates them.
            magic = ''
        return magic

    # internal function to gather dependency info used by the above ones
    def _gather_aux(self, header, xs, nametag, flagstag, versiontag,
                    prereq=None):
        names = header[nametag]
        flags = header[flagstag]
        versions = header[versiontag]

        if versions:
            for loop in range(len(versions)):
                name = byte_to_string(names[loop])
                evr = stringToVersion(byte_to_string(versions[loop]))
                if prereq is not None and flags[loop] & PREREQ_FLAG:
                    prereq.append((name, flags[loop] & (~PREREQ_FLAG), evr))
                else:
                    xs.append(DepInfo(name, flags[loop], evr))
        return xs, prereq

    def _gather_dep_info(self):
        _requires = []
        _prereq = []
        _provides = []
        _conflicts = []
        _obsoletes = []
        _recommends = []
        _suggests = []
        _enhances = []
        _supplements = []

        _requires, _prereq = self._gather_aux(self.header, _requires,
                                              rpm.RPMTAG_REQUIRENAME,
                                              rpm.RPMTAG_REQUIREFLAGS,
                                              rpm.RPMTAG_REQUIREVERSION,
                                              _prereq)
        _conflits, _ = self._gather_aux(self.header, _conflicts,
                                        rpm.RPMTAG_CONFLICTNAME,
                                        rpm.RPMTAG_CONFLICTFLAGS,
                                        rpm.RPMTAG_CONFLICTVERSION)
        _provides, _ = self._gather_aux(self.header, _provides,
                                        rpm.RPMTAG_PROVIDENAME,
                                        rpm.RPMTAG_PROVIDEFLAGS,
                                        rpm.RPMTAG_PROVIDEVERSION)
        _obsoletes, _ = self._gather_aux(self.header, _obsoletes,
                                         rpm.RPMTAG_OBSOLETENAME,
                                         rpm.RPMTAG_OBSOLETEFLAGS,
                                         rpm.RPMTAG_OBSOLETEVERSION)
        _recommends, _ = self._gather_aux(self.header, _recommends,
                                          rpm.RPMTAG_RECOMMENDNAME,
                                          rpm.RPMTAG_RECOMMENDFLAGS,
                                          rpm.RPMTAG_RECOMMENDVERSION)
        _suggests, _ = self._gather_aux(self.header, _suggests,
                                        rpm.RPMTAG_SUGGESTNAME,
                                        rpm.RPMTAG_SUGGESTFLAGS,
                                        rpm.RPMTAG_SUGGESTVERSION)
        _enhances, _ = self._gather_aux(self.header, _enhances,
                                        rpm.RPMTAG_ENHANCENAME,
                                        rpm.RPMTAG_ENHANCEFLAGS,
                                        rpm.RPMTAG_ENHANCEVERSION)
        _supplements, _ = self._gather_aux(self.header, _supplements,
                                           rpm.RPMTAG_SUPPLEMENTNAME,
                                           rpm.RPMTAG_SUPPLEMENTFLAGS,
                                           rpm.RPMTAG_SUPPLEMENTVERSION)

        return (_requires, _prereq, _provides, _conflicts, _obsoletes, _recommends,
                _suggests, _enhances, _supplements)

    def scriptprog(self, which):
        """
        Get the specified script interpreter as a string.
        Depending on rpm-python version, the string may or may not include
        interpreter arguments, if any.
        """
        if which is None:
            return ''
        prog = self[which]
        if prog is None:
            prog = ''
        elif isinstance(prog, (list, tuple)):
            # http://rpm.org/ticket/847#comment:2
            prog = ''.join(prog)
        return prog

    def __enter__(self):
        return self

    def __exit__(self, exc_type, exc_val, exc_tb):
        self.cleanup()

    def check_versioned_dep(self, name, version):
        # try to match name%_isa as well (e.g. 'foo(x86-64)', 'foo(x86-32)')
        name_re = re.compile(r'^%s(\(\w+-\d+\))?$' % re.escape(name))
        for d in self.requires + self.prereq:
            if name_re.match(d[0]):
                if d[1] & rpm.RPMSENSE_EQUAL != rpm.RPMSENSE_EQUAL \
                        or d[2][1] != version:
                    return False
                return True
        return False

    def read_with_mmap(self, filename):
        """Mmap a file, return it's content decoded."""
        try:
            with open(Path(self.dir_name() or '/', filename.lstrip('/'))) as in_file:
                return mmap.mmap(in_file.fileno(), 0, mmap.MAP_SHARED, mmap.PROT_READ).read().decode()
        except Exception:
            return ''

    def grep(self, regex, filename):
        """Grep regex from a file, return first matching line number (starting with 1)."""
        data = self.read_with_mmap(filename)
        match = regex.search(data)
        if match:
            return data.count('\n', 0, match.start()) + 1
        else:
            return None


class Pkg(AbstractPkg):
    _magic_from_compressed_re = re.compile(r'\([^)]+\s+compressed\s+data\b')

    def __init__(self, filename, dirname, header=None, is_source=False, extracted=False, verbose=False):
        self.filename = filename
        self.extracted = extracted

        # record decompression and extraction time
        start = time.monotonic()
        self.dirname = self._extract_rpm(dirname, verbose)
        self.timers = {'ExtractRpm': time.monotonic() - start, 'libmagic': 0}
        self.current_linenum = None

        self._req_names = -1

        if header:
            self.header = header
            self.is_source = is_source
        else:
            # Create a package object from the file name
            ts = rpm.TransactionSet()
            # Don't check signatures here...
            ts.setVSFlags(rpm._RPMVSF_NOSIGNATURES)
            fd = os.open(filename, os.O_RDONLY)
            try:
                self.header = ts.hdrFromFdno(fd)
            finally:
                os.close(fd)
            self.is_source = not self.header[rpm.RPMTAG_SOURCERPM]

        self.name = self[rpm.RPMTAG_NAME]

        (self.requires, self.prereq, self.provides, self.conflicts,
         self.obsoletes, self.recommends, self.suggests, self.enhances,
         self.supplements) = self._gather_dep_info()

        self.req_names = [x[0] for x in self.requires + self.prereq]

        self.files = self._gather_files_info()
        self.config_files = [x.name for x in self.files.values() if x.is_config]
        self.doc_files = [x.name for x in self.files.values() if x.is_doc]
        self.ghost_files = [x.name for x in self.files.values() if x.is_ghost]
        self.noreplace_files = [x.name for x in self.files.values() if x.is_noreplace]
        self.missingok_files = [x.name for x in self.files.values() if x.is_missingok]

        if self.is_no_source:
            self.arch = 'nosrc'
        elif self.is_source:
            self.arch = 'src'
        else:
            self.arch = self.header.format('%{ARCH}')

    # Return true if the package is a nosource package.
    # NoSource files are ghosts in source packages.
    @property
    def is_no_source(self):
        return self.is_source and self.ghost_files

    # access the tags like an array
    def __getitem__(self, key):
        try:
            val = self.header[key]
        except KeyError:
            val = []
        if val == []:
            return None
        else:
            # Note that text tags we want to try decoding for real in TagsCheck
            # such as summary, description and changelog are not here.
            if key in (rpm.RPMTAG_NAME, rpm.RPMTAG_VERSION, rpm.RPMTAG_RELEASE,
                       rpm.RPMTAG_ARCH, rpm.RPMTAG_GROUP, rpm.RPMTAG_BUILDHOST,
                       rpm.RPMTAG_LICENSE, rpm.RPMTAG_HEADERI18NTABLE,
                       rpm.RPMTAG_PACKAGER, rpm.RPMTAG_SOURCERPM,
                       rpm.RPMTAG_DISTRIBUTION, rpm.RPMTAG_VENDOR) \
            or key in (x[0] for x in SCRIPT_TAGS) \
            or key in (x[1] for x in SCRIPT_TAGS):
                val = byte_to_string(val)
                if key == rpm.RPMTAG_GROUP and val == 'Unspecified':
                    val = None
            return val

    # return the name of the directory where the package is extracted
    def dir_name(self):
        return self.dirname

    def _extract_rpm(self, dirname, verbose):
        if not Path(dirname).is_dir():
            print_warning('Unable to access dir %s' % dirname)
        elif dirname == '/':
            # it is an InstalledPkg
            pass
        else:
            self.__tmpdir = tempfile.TemporaryDirectory(
                prefix='rpmlint.%s.' % Path(self.filename).name, dir=dirname
            )
            dirname = self.__tmpdir.name

            # BusyBox' cpio does not support '-D' argument and the only safe
            # usage is doing chdir before invocation.
            filename = Path(self.filename).resolve()
            with pushd(dirname):
                stderr = None if verbose else subprocess.DEVNULL
                if shutil.which('rpm2archive'):
                    with open(filename, 'rb') as rpm_data:
                        subprocess.check_output('rpm2archive - | tar -xz && chmod -R +rX .', shell=True, env=ENGLISH_ENVIRONMENT,
                                                stderr=stderr, stdin=rpm_data)
                else:
                    command_str = f'rpm2cpio {quote(str(filename))} | cpio -id && chmod -R +rX .'
                    subprocess.check_output(command_str, shell=True, env=ENGLISH_ENVIRONMENT, stderr=stderr)
            self.extracted = True
        return dirname

    def check_signature(self):
        ret = subprocess.run(('rpm', '-Kv', self.filename),
                             stdout=subprocess.PIPE, stderr=subprocess.STDOUT,
                             env=ENGLISH_ENVIRONMENT, text=True)
        text = ret.stdout
        if text.endswith('\n'):
            text = text[:-1]
        return ret.returncode, text

    # remove the extracted files from the package
    def cleanup(self):
        if self.extracted and self.dirname:
            self.__tmpdir.cleanup()

    def langtag(self, tag, lang):
        """Get value of tag in the given language."""
        # LANGUAGE trumps other env vars per GNU gettext docs, see also #166
        orig = os.environ.get('LANGUAGE')
        os.environ['LANGUAGE'] = lang
        ret = self[tag]
        if orig is not None:
            os.environ['LANGUAGE'] = orig
        return ret

    # extract information about the files
    def _gather_files_info(self):
        ret = {}
        flags = self.header[rpm.RPMTAG_FILEFLAGS]
        modes = self.header[rpm.RPMTAG_FILEMODES]
        users = self.header[rpm.RPMTAG_FILEUSERNAME]
        groups = self.header[rpm.RPMTAG_FILEGROUPNAME]
        links = [byte_to_string(x) for x in self.header[rpm.RPMTAG_FILELINKTOS]]
        sizes = self.header[rpm.RPMTAG_FILESIZES]
        if len(sizes) != len(flags):
            sizes = self.header[rpm.RPMTAG_LONGFILESIZES]
        md5s = self.header[rpm.RPMTAG_FILEMD5S]
        mtimes = self.header[rpm.RPMTAG_FILEMTIMES]
        rdevs = self.header[rpm.RPMTAG_FILERDEVS]
        langs = self.header[rpm.RPMTAG_FILELANGS]
        inodes = self.header[rpm.RPMTAG_FILEINODES]
        requires = [byte_to_string(x) for x in self.header[rpm.RPMTAG_FILEREQUIRE]]
        provides = [byte_to_string(x) for x in self.header[rpm.RPMTAG_FILEPROVIDE]]
        files = [byte_to_string(x) for x in self.header[rpm.RPMTAG_FILENAMES]]
        magics = [byte_to_string(x) for x in self.header[rpm.RPMTAG_FILECLASS]]
        try:  # rpm >= 4.7.0
            filecaps = self.header[rpm.RPMTAG_FILECAPS]
        except AttributeError:
            filecaps = None

        # rpm-python < 4.6 does not return a list for this (or FILEDEVICES,
        # FWIW) for packages containing exactly one file
        if not isinstance(inodes, list):
            inodes = [inodes]

        if files:
            for idx, file in enumerate(files):
                pkgfile = PkgFile(file)
                pkgfile.path = os.path.normpath(os.path.join(
                    self.dir_name() or '/', pkgfile.name.lstrip('/')))
                pkgfile.flags = flags[idx]
                pkgfile.mode = modes[idx]
                pkgfile.user = byte_to_string(users[idx])
                pkgfile.group = byte_to_string(groups[idx])
                pkgfile.linkto = links[idx] and os.path.normpath(links[idx])
                pkgfile.size = sizes[idx]
                pkgfile.md5 = md5s[idx]
                pkgfile.mtime = mtimes[idx]
                pkgfile.rdev = rdevs[idx]
                pkgfile.inode = inodes[idx]
                pkgfile.requires = parse_deps(requires[idx])
                pkgfile.provides = parse_deps(provides[idx])
                pkgfile.lang = byte_to_string(langs[idx])
                pkgfile.magic = magics[idx]
                pkgfile.magic = self._calc_magic(pkgfile)
                if filecaps:
                    pkgfile.filecaps = byte_to_string(filecaps[idx])
                ret[pkgfile.name] = pkgfile
        return ret

    def readlink(self, pkgfile):
        """
        Resolve symlinks for the given PkgFile, return the dereferenced
        PkgFile if it is found in this package, None if not.
        """
        result = pkgfile
        while result and result.linkto:
            linkpath = urljoin(result.name, result.linkto)
            linkpath = os.path.normpath(linkpath)
            result = self.files.get(linkpath)
        return result

    def get_core_reqs(self):
        """
        Return the list of dependencies that are not found by find-requires
        withouth the flag RPM
        """
        core_reqs = []

        for dep in rpm.ds(self.header, 'requires'):
            # skip deps which were found by find-requires
            if dep.Flags() & rpm.RPMSENSE_FIND_REQUIRES != 0:
                continue
            core_reqs.append(dep.N())

        return core_reqs


def get_installed_pkgs(name):
    """Get list of installed package objects by name."""

    ts = rpm.TransactionSet()
    if re.search(r'[?*]|\[.+\]', name):
        mi = ts.dbMatch()
        mi.pattern('name', rpm.RPMMIRE_GLOB, name)
    else:
        mi = ts.dbMatch('name', name)

    return [InstalledPkg(name, hdr) for hdr in mi]


# Class to provide an API to an installed package
class InstalledPkg(Pkg):
    def __init__(self, name, hdr=None):
        if not hdr:
            ts = rpm.TransactionSet()
            mi = ts.dbMatch('name', name)
            if not mi:
                raise KeyError(name)
            try:
                hdr = next(mi)
            except StopIteration:
                raise KeyError(name)

        super().__init__(name, '/', hdr, extracted=True)
        # create a fake filename to satisfy some checks on the filename
        self.filename = '%s-%s-%s.%s.rpm' % \
            (self.name, self[rpm.RPMTAG_VERSION], self[rpm.RPMTAG_RELEASE],
             self[rpm.RPMTAG_ARCH])

    def cleanup(self):
        pass

    def check_signature(self):
        return (0, 'fake: pgp md5 OK')


class FakeHeader(dict):
    def sprintf(self, expr):
        """
        Replaces expressions like %{} with actual package
        """

        tagre = re.compile(r'%{([^}]*)}')
        for tag in tagre.findall(expr):
            expr = expr.replace(f'%{tag}', self[f'RPMTAG_{tag}'])
        return expr

    def __missing__(self, key):
        try:
            key = getattr(rpm, key)
        except (TypeError, KeyError):
            raise KeyError

        if key not in self:
            raise KeyError

        return self[key]


# Class to provide an API to a 'fake' package, eg. for specfile-only checks
class FakePkg(AbstractPkg):
    _autoheaders = [
        'requires',
        'conflicts',
        'provides',
        'obsoletes',
        'recommends',
        'suggests',
        'enhances',
        'supplements',
    ]

    def __init__(self, name, is_source=False):
        self.timers = {'ExtractRpm': 0, 'libmagic': 0}
        self.name = str(name)
        self.filename = f'{name}.rpm'
        self.arch = None
        self.current_linenum = None
        self.dirname = None
        self.is_source = False

        # files are dictionary where key is name of a file
        self.files = {}
        self.ghost_files = {}

        # header is a dictionary to mock rpm metadata
        self.header = FakeHeader()
        for i in self._autoheaders:
            # the header name wihtout the ending 's'
            tagname = i[:-1].upper()
            self.header[getattr(rpm, f'RPMTAG_{tagname}NAME')] = []
            self.header[getattr(rpm, f'RPMTAG_{tagname}FLAGS')] = []
            self.header[getattr(rpm, f'RPMTAG_{tagname}VERSION')] = []
        self.header[rpm.RPMTAG_FILENAMES] = []

    def add_file(self, path, name):
        pkgfile = PkgFile(name)
        pkgfile.path = path
        self.files[name] = pkgfile
        return pkgfile

    def _mock_file(self, path, attrs):
        metadata = None
        if attrs.get('create_dirs', False):
            for i in PurePath(path).parents[:attrs.get('include_dirs', -1)]:
                self.add_dir(str(i))
        metadata = attrs.get('metadata', None)

        if attrs.get('is_dir', False):
            self.add_dir(path, metadata=metadata)
            return

        content = ''
        if 'content-path' in attrs:
            content = open(attrs['content-path'], 'rb')
        elif 'content' in attrs:
            content = attrs['content']

        if 'linkto' in attrs:
            self.add_symlink_to(path, attrs['linkto'])
        else:
            self.add_file_with_content(path, content, metadata=metadata)
        self.header[rpm.RPMTAG_FILENAMES].append(path)

        if 'content-path' in attrs:
            content.close()

    def create_files(self, files):
        """
        This is a helper method to create files(real files); not PkgFile
        objects.
        """

        # files can be just a list
        if isinstance(files, list) or isinstance(files, tuple):
            for path in files:
                self._mock_file(path, {})
        # list of files with attributes and content
        elif isinstance(files, dict):
            for path, file in files.items():
                self._mock_file(path, file)

    def add_dir(self, path, metadata=None):
        name = path
        pkgdir = PkgFile(name)
        pkgdir.magic = 'directory'

        path = os.path.join(self.dir_name(), path.lstrip('/'))
        os.makedirs(Path(path), exist_ok=True)
        pkgdir.inode = os.stat(Path(path)).st_ino

        pkgdir.path = path
        self.files[name] = pkgdir

        if metadata:
            for k, v in metadata.items():
                setattr(pkgdir, k, v)

        return pkgdir

    def add_file_with_content(self, name, content, metadata=None, **flags):
        """
        Add file to the FakePkg and fill the file with provided
        string content.
        """
        path = os.path.join(self.dir_name(), name.lstrip('/'))
        pkg_file = PkgFile(name)
        pkg_file.path = path
        pkg_file.mode = stat.S_IFREG | 0o0644
        pkg_file.user = 'root'
        pkg_file.group = 'root'
        self.files[name] = pkg_file

        # create files in filesystem
        os.makedirs(Path(path).parent, exist_ok=True)

        if isinstance(content, str):
            content = content.encode('utf-8', errors='ignore')

        with open(Path(path), 'wb') as out:
            # file like content
            if isinstance(content, io.IOBase):
                shutil.copyfileobj(content, out)
            else:
                out.write(content)

        # Generating md5 hash values for real files:
        pkg_file.md5 = self.md5_checksum(Path(path))
        pkg_file.size = os.path.getsize(Path(path))
        pkg_file.inode = os.stat(Path(path)).st_ino
        pkg_file.magic = self._calc_magic(pkg_file)

        if metadata:
            for k, v in metadata.items():
                setattr(pkg_file, k, v)
        for key, value in flags.items():
            setattr(pkg_file, key, value)

    def initiate_files_base_data(self):
        """ This method is called after adding metadata of each file """
        self.config_files = [x.name for x in self.files.values() if x.is_config]
        self.doc_files = [x.name for x in self.files.values() if x.is_doc]
        self.ghost_files = [x.name for x in self.files.values() if x.is_ghost]
        self.noreplace_files = [x.name for x in self.files.values() if x.is_noreplace]
        self.missingok_files = [x.name for x in self.files.values() if x.is_missingok]

    def add_header(self, header):
        for k, v in header.items():
            if k in self._autoheaders:
                # the header name wihtout the ending 's'
                tagname = k[:-1].upper()
                for i in v:
                    name, flags, version = parse_deps(i)[0]
                    version = versionToString(version)
                    self.header[getattr(rpm, f'RPMTAG_{tagname}NAME')].append(name)
                    self.header[getattr(rpm, f'RPMTAG_{tagname}FLAGS')].append(flags)
                    self.header[getattr(rpm, f'RPMTAG_{tagname}VERSION')].append(version)
                continue

            key = getattr(rpm, f'RPMTAG_{k}'.upper())
            self.header[key] = v

            if key == rpm.RPMTAG_ARCH:
                self.arch = v

        (self.requires, self.prereq, self.provides, self.conflicts,
         self.obsoletes, self.recommends, self.suggests, self.enhances,
         self.supplements) = self._gather_dep_info()

        self.req_names = [x[0] for x in self.requires + self.prereq]

    def add_symlink_to(self, name, target):
        """
        Add symlink to name file which path is related to name.
        Eg. name == '/etc/foo' and target == '../bar' creates a symlink file
        /etc/bar that points to /etc/foo.
        """
        pkg_file = PkgFile(name)
        pkg_file.mode = stat.S_IFLNK
        pkg_file.linkto = target
        pkg_file.user = 'root'
        pkg_file.group = 'root'
        self.files[name] = pkg_file

    def readlink(self, pkgfile):
        # HACK: reuse the real Pkg's logic
        return Pkg.readlink(self, pkgfile)

    def dir_name(self):
        if not self.dirname:
            self.__tmpdir = tempfile.TemporaryDirectory(prefix='rpmlint.%s.' % Path(self.name).name)
            self.dirname = self.__tmpdir.name
        return self.dirname

    def md5_checksum(self, file_name):
        md5_hash = hashlib.md5()
        with open(file_name, 'rb') as f:
            for byte_block in iter(lambda: f.read(4096), b''):
                md5_hash.update(byte_block)
        return md5_hash.hexdigest()

    def cleanup(self):
        if self.dirname:
            self.__tmpdir.cleanup()

    def get_core_reqs(self):
        core_reqs = []
        return core_reqs

    # access the tags like an array
    def __getitem__(self, key):
        return self.header.get(key, None)