urpm-tools/rpm5utils/synthesis.py

import rpm
import re
from urllib2 import urlopen, HTTPError, URLError
import subprocess
import platform
import sys

ARCH = platform.machine()

def get_command_output(command, fatal_fails=True):
    '''Execute command using subprocess.Popen and return its stdout output string. If return code is not 0, print error message and exit'''
    #vprint("Executing command: " + str(command))
    res = subprocess.Popen(command, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
    output = list(res.communicate())
    if sys.stdout.encoding:
      output[0] = output[0].decode(sys.stdout.encoding).encode("UTF-8")
      output[1] = output[1].decode(sys.stdout.encoding).encode("UTF-8")
    if(res.returncode != 0 and fatal_fails):
        print(_("Error while calling command") + " '" + " ".join(command) + "'")
        if(output[1] != None or output[0] != None):
            print(_("Error message: \n")+  ((output[0].strip() + "\n") if output[0]!=None else "") +
                  (output[1].strip() if output[1]!=None else "") )
        exit(1)
    return [output[0], output[1], res.returncode]

class MediaSet(object):   
    @staticmethod
    def from_list(media):
        ms = MediaSet()
        ms.urls = []
        ms.media = {}
        ms.by_url = {}
        for medium in media:
            med, url = medium
            ms.media[med] = url
            ms.by_url[url] = med
            ms.urls.append(url)
        return ms
            
    @staticmethod
    def from_system(cmd):
        ms = MediaSet()
        ms.urls = []
        ms.media = {}
        ms.by_url = {}
        
        lines = get_command_output(cmd + ["--list-url", "--list-media", 'active'])[0].strip().split("\n")
        
        for line in lines:
            parts = line.split(" ")
            medium = ' '.join(parts[:-1])
            url = parts[-1]
            if(url.endswith("/")):
                url = url[:-1]
            if(url.find('/') != -1):
                ms.media[medium] = url
                ms.by_url[parts[-1]] = medium
                ms.urls.append(url)
        return ms
        

class NEVR(object):
    EQUAL = rpm.RPMSENSE_EQUAL #8
    GREATER = rpm.RPMSENSE_GREATER #4
    LESS = rpm.RPMSENSE_LESS #2
    #re_ver = re.compile('^([\d\.]+:)?([\w\d\.\-\[\]]+)(:[\d\.]+)?$')
    
    re_dep_ver = re.compile('^([^ \[\]]+)\[([\>\<\=\!]*) ([^ ]+)\]$')    
    re_dep = re.compile('^([^ \[\]]+)$')
    types = {None: 0,
             '==' : EQUAL,
             '' : EQUAL,
             '=' : EQUAL,
             '>=' : EQUAL|GREATER,
             '<=' : EQUAL|LESS,
             '>' : GREATER,
             '<' : LESS,
             '!=' : LESS|GREATER,
             '<>' : LESS|GREATER}  
    __slots__=['N', 'EVR', 'DE', 'DT', 'FL', 'E', 'VR']
    def __init__(self, N, EVR, DE=None, DT=None, FL=None, E=None):      
        self.N = N
        self.EVR = EVR
        self.DE = DE
        self.DT = DT
        self.FL = FL
        self.E = E
        self.VR = EVR
        
        if E:
            if EVR.startswith(E + ':'):
                self.VR = EVR[len(E)+1:]
            else:
                self.EVR = E + ':' + self.EVR
   
        #try to get E
        if not self.E and self.EVR and self.EVR.find(':') != -1:
            items = self.EVR.split(':')
            if items[0].find('.') == -1 and items[0].find('-') == -1:
                self.E = items[0]
        if not self.E and self.EVR:
            self.E = '0'
            self.EVR = '0:' + self.EVR
        
        if self.DE == 'None':
            self.DE = None
        
    def __str__(self):
        if self.FL:
            for t in NEVR.types:
                if not t:
                    continue
                if NEVR.types[t] == self.FL:
                    return "%s %s %s" % (self.N, t, self.EVR)
        if self.EVR:
            return "%s == %s" % (self.N, self.EVR)
        
        return "%s" % (self.N)

    def __repr__(self):
        return self.__str__()
    
    def __eq__(self, val):
        if not isinstance(val, NEVR):
            raise Exception("Internal error: comparing between NEVR and " + str(type(val)))
        return str(self) == str(val)
    
    def __ne__(self, val):
        return not (self == val)
    
    @staticmethod
    def from_depstring(s, DE_toremove=None):
        s = s.replace('[*]', '')
        
        if DE_toremove:
            res = NEVR.re_dep_ver.match(s)
            if res:
                (name, t, val) = res.groups()
            
                if val.endswith(':' + DE_toremove):
                    val = val[:-(len(DE_toremove) + 1)]
                    EVR = '%s[%s %s]' % (name, t, val)
        
        res = NEVR.re_dep.match(s)
        if res:
            return NEVR(res.group(1), None)
        
        res = NEVR.re_dep_ver.match(s)
            
        if not res:
            raise Exception('Incorrect requirement string: ' + s)
        (name, t, val) = res.groups()
        return NEVR(name, val, FL=NEVR.types[t])
    
    
    re_version =  re.compile("(\.)?((alpha)|(cvs)|(svn)|(r))?\d+((mdv)|(mdk)|(mnb))")
    @staticmethod
    def from_filename(rpmname, E=None):
            ''' Returns [name, version] for given rpm file or package name '''
            suffix = ['.x86_64', '.noarch'] + ['.i%s86' % i for i in range(3,6)]
            for s in suffix:
                if(rpmname.endswith(s)):
                    rpmname = rpmname[:-len(s)]
            
            sections = rpmname.split("-")
            if(NEVR.re_version.search(sections[-1]) == None):
                name = sections[:-3]
                version = sections[-3:-1]
            else:
                name = sections[:-2]
                version = sections[-2:]
            return NEVR("-".join(name), "-".join(version), FL=NEVR.EQUAL, E=E)
    
    def satisfies(self, val):
        if self.N != val.N:
            return False
                    
        if self.EVR == None or val.EVR == None:
            return True
        
        (pname, pt, pval) = (self.N, self.FL, self.EVR)
        (rname, rt, rval) = (val.N, val.FL, val.EVR)

        def cut_part(seperator, val1, val2):
            if val1 and val2 and val1.count(seperator) != val2.count(seperator):
                n = max(val1.count(seperator), val2.count(seperator))
                val1 = seperator.join(val1.split(seperator)[:n])
                val2 = seperator.join(val2.split(seperator)[:n])
            return (val1, val2)
            
        (rval, pval) = cut_part(':', rval, pval)
        (rval, pval) = cut_part('-', rval, pval)
              
        res = rpm.evrCompare(rval, pval)
        
        if res == 1: # >
            if pt & NEVR.GREATER:
                return True
            elif pt & NEVR.LESS:
                if rt & NEVR.LESS:
                    return True
                else:
                    return False
            else: 
                if rt & NEVR.LESS:
                    return True
                else:
                    return False
            
        elif res == 0:
            if rt & NEVR.EQUAL and pt & NEVR.EQUAL:
                return True
            if rt & NEVR.LESS and pt & NEVR.LESS:
                return True
            if rt & NEVR.GREATER and pt & NEVR.GREATER:
                return True
            return False
            
        else: # <
            if rt & NEVR.GREATER:
                return True
            elif rt & NEVR.LESS:
                if pt & NEVR.LESS:
                    return True
                else:
                    return False
            else: 
                if pt & NEVR.LESS:
                    return True
                else:
                    return False
          
        
class PackageSet:
    tags = ['provides','requires','obsoletes','suggests', 'conflicts']
    alltags = tags + ['nevr', 'arch']
    def __init__(self):
        self.what = {}
        self.packages = {}
            
    def load_from_system(self):
        print(_("Loading the list of installed packages..."))
        ts = rpm.TransactionSet() 
        mi = ts.dbMatch() 
   
        for tag in PackageSet.tags:
            self.what[tag] = {}
            
        for h in mi: 
            name = h['name']
            if(name == 'gpg-pubkey'):
                continue
            if(name not in self.packages):
                self.packages[h['name']] = {}
            else:
                print(_("Duplicating ") + name + '-' + h['version'] + '-' + h['release'])
                print(_("Already found: ") + name + '-' + self.packages[name]["nevr"].EVR)
            
            E = str(h['epoch'])
            V = h['version']
            R = h['release']
            DE = h['distepoch']
            DT = h['disttag']
            
            if E == None or E == 'None':
                E = '0'
                            
            EVR = "%s:%s-%s" % (E, V, R)
            
            nevr = NEVR(name, EVR, FL=NEVR.EQUAL, DE=DE, DT=DT, E=E)
            self.packages[name]['nevr'] = nevr
            self.packages[name]['arch'] = h['arch']

            for tag in PackageSet.tags:
                if tag not in self.packages[name]:
                    self.packages[name][tag] = []
                dss = h.dsFromHeader(tag[:-1] + 'name')
                for s in dss:                  
                    fl = s.Flags()
                    #undocumented flag for special dependencies
                    if fl & 16777216:
                        continue
                    fl = fl % 16
    
                    _evr = s.EVR()
                    
                    if _evr == '':
                        evr = NEVR(s.N(), None, FL=fl)
                    else:
                        evr = NEVR(s.N(), _evr, FL=fl)
    
                    self.packages[name][tag].append(evr)
                    
                    if evr.N not in self.what[tag]:
                        self.what[tag][evr.N] = []
                    self.what[tag][evr.N].append((name, evr))
                    
    def load_from_repository(self, ms):
        url_by_synthesis_url = {}   
        global fields
        
        def get_synthesis_by_url(url):
            if url.startswith('file://'):
                url = url[6:]
            if url.startswith('/'):
                medium = ms.by_url[url]
                return '/var/lib/urpmi/%s/synthesis.hdlist.cz' % medium
            else:
                return url + "/media_info/synthesis.hdlist.cz"
                
        medium_by_synth = {}
        synthesis_lists = []
        for url in ms.urls:
            synth = get_synthesis_by_url(url)
            synthesis_lists.append(synth)
            url_by_synthesis_url[synth] = url
            medium_by_synth[synth] = ms.by_url[url]
    
        def clear_data():
            '''Clears the data of the current package from 'fields' dictionary'''
            global fields
            fields = {"provides":[], "requires":[], "obsoletes":[], "suggests":[], 
                      "conflicts":[], "info":[], "summary":[]}
        arches32 = ['i%d86' for i in range(3,6)]
        for tag in PackageSet.tags:
            self.what[tag] = {}
            
        #the following code is awful, I know. But it's easy-to-understand and clear.
        # don't like it - write better and send me :)
        for synthesis_list in synthesis_lists:
            try:
                #print synthesis_list
                print(_("Processing medium ") + medium_by_synth[synthesis_list] + "...")
                if(synthesis_list.startswith("http://") or synthesis_list.startswith("ftp://")):
                    r = urlopen(synthesis_list)
                    s = r.read()
                    r.close()
                elif(synthesis_list.startswith("rsync://")):
                    tmppath = '/tmp/urpm-reposync.synthesis_lists'
                    if (not os.path.exists(tmppath)):
                        os.mkdir(tmppath)
                    filename = tmppath + '/' + os.path.basename(synthesis_list)
                    os.system("rsync --copy-links %s %s 1>/dev/null 2>&1" % (synthesis_list, filename))
                    r = open(filename)
                    s = r.read()
                    r.close()
                    shutil.rmtree(tmppath)
                elif(synthesis_list.startswith("/")): #local file
                    if not os.path.exists(synthesis_list):
                        eprint(_('Could not read synthesis file. (File %s not found)') % synthesis_list)
                        continue
                    r = open(synthesis_list)
                    s = r.read()
                    r.close()
                res = subprocess.Popen(['gzip', '-d'], stdin=subprocess.PIPE, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
                output = res.communicate(s)
                clear_data()
                for line in output[0].split('\n'):
                    if(line == ''):  # there can be empty lines
                        continue
                    
                    items = line.split("@")
                    data = [x.strip() for x in items[2:]]
                    fields[items[1]] = data
                    
                    if(items[1] == "info"):
                        rpmname = items[2]
                        size = int(items[4])
                        nevr = NEVR.from_filename(items[2], E=items[3])
                        nevr.E = items[3]
                        
                        disttagepoch = '-'
                        if(len(items)>6):
                            disttagepoch = items[6]
                            nevr.DT = items[6]
                        if(len(items)>7):
                            disttagepoch += items[7]
                            nevr.DE = items[7]
    
                        arch = items[2].split('.')[-1]
                        if arch in arches32 and ARCH in arches:
                            arch = ARCH
                        
                        in_repo = nevr.N in self.packages
                        new_arch_correct = arch == ARCH
                        
                        if in_repo:
                            if nevr.DE == self.packages[nevr.N]['nevr'].DE:
                                ver_newer =  rpm.evrCompare(nevr.EVR, self.packages[nevr.N]['nevr'].EVR) == 1
                            else:
                                ver_newer = (nevr.DE > self.packages[nevr.N]['nevr'].DE)
                            
                            old_arch_correct = self.packages[nevr.N]['arch'] == ARCH
                        else:
                            ver_newer = None
                            old_arch_correct = None
                                                
                        toinst = not in_repo or (not old_arch_correct and new_arch_correct) or \
                                (ver_newer and old_arch_correct == new_arch_correct) 

                        if toinst:
                            #remove old data
                            if nevr.N in self.packages:
                                for tag in PackageSet.tags:
                                    for dep in self.packages[nevr.N][tag]:
                                        self.what[tag][dep.N].remove((nevr.N, dep))
                            else:
                                self.packages[nevr.N] = {}
                                
                            self.packages[nevr.N]['nevr'] = nevr
                            self.packages[nevr.N]["arch"] = arch
                            self.packages[nevr.N]["synthesis_list"] = synthesis_list
                            self.packages[nevr.N]["filename"] = rpmname
                            self.packages[nevr.N]["size"] = size
                            for tag in PackageSet.tags:
                                self.packages[nevr.N][tag] = []
                                for item in fields[tag]:
                                    if item == '':
                                        continue
                                    dep = NEVR.from_depstring(item, DE_toremove=nevr.DE)
                                    self.packages[nevr.N][tag].append(dep)
                                    if dep.N not in self.what[tag]:
                                        self.what[tag][dep.N] = []
                                    self.what[tag][dep.N].append((nevr.N, dep))
                                
                            self.packages[nevr.N]['medium'] = medium_by_synth[synthesis_list]
                        clear_data()
            except (HTTPError,URLError):
                eprint(_("File can not be processed! Url: ") + synthesis_list)


    def whattag(self, tag, val):
        if val.N not in self.what[tag]:
            return []
        found = []
        for (pkg, dep) in self.what[tag][val.N]:
            if dep.satisfies(val):
                found.append(pkg)
        return found
    
    def whattag_revert(self, tag, val):
        if val.N not in self.what[tag]:
            return []
        found = []
        for (pkg, dep) in self.what[tag][val.N]:
            if val.satisfies(dep):
                found.append(pkg)
        return found
    
    def whatprovides(self, val):
        return self.whattag('provides', val)

    def whatobsoletes(self, val):
        return self.whattag_revert('obsoletes', val)

    def whatrequires(self, val):
        return self.whattag_revert('requires', val)
    
    def whatconflicts(self, val):
        return self.whattag_revert('conflicts', val)
    
    def whatrequires_pkg(self, pkg):
        found = []
        for req in self.packages[pkg]['provides']:
            found += [(d, req) for d in self.whatrequires(req)]
        return found