src/KnownFileList.cpp

   1 //
   2 // This file is part of the aMule Project.
   3 //
   4 // Copyright (c) 2003-2011 aMule Team ( admin@amule.org / http://www.amule.org )
   5 // Copyright (c) 2002-2011 Merkur ( devs@emule-project.net / http://www.emule-project.net )
   6 //
   7 // Any parts of this program derived from the xMule, lMule or eMule project,
   8 // or contributed by third-party developers are copyrighted by their
   9 // respective authors.
  10 //
  11 // This program is free software; you can redistribute it and/or modify
  12 // it under the terms of the GNU General Public License as published by
  13 // the Free Software Foundation; either version 2 of the License, or
  14 // (at your option) any later version.
  15 //
  16 // This program is distributed in the hope that it will be useful,
  17 // but WITHOUT ANY WARRANTY; without even the implied warranty of
  18 // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  19 // GNU General Public License for more details.
  20 //
  21 // You should have received a copy of the GNU General Public License
  22 // along with this program; if not, write to the Free Software
  23 // Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA
  24 //
  25
  26
  27 #include "KnownFileList.h"      // Interface declarations
  28
  29 #include <common/DataFileVersion.h>
  30
  31 #include <memory>               // Do_not_auto_remove (lionel's Mac, 10.3)
  32 #include "PartFile.h"           // Needed for CPartFile
  33 #include "amule.h"
  34 #include "Logger.h"
  35 #include "MemFile.h"
  36 #include "ScopedPtr.h"
  37 #include "SearchList.h"         // Needed for UpdateSearchFileByHash
  38 #include <common/Format.h>
  39
  40
  41 // This function is inlined for performance
  42 inline bool CKnownFileList::KnownFileMatches(
  43         CKnownFile *knownFile,
  44         const CPath& filename,
  45         uint32 in_date,
  46         uint64 in_size) const
  47 {
  48         return
  49                 (knownFile->GetLastChangeDatetime() == (time_t)in_date) &&
  50                 (knownFile->GetFileSize() == in_size) &&
  51                 (knownFile->GetFileName() == filename);
  52 }
  53
  54
  55 CKnownFileList::CKnownFileList()
  56 {
  57         accepted = 0;
  58         requested = 0;
  59         transferred = 0;
  60         m_filename = wxT("known.met");
  61         m_knownSizeMap = NULL;
  62         m_duplicateSizeMap = NULL;
  63         Init();
  64 }
  65
  66
  67 CKnownFileList::~CKnownFileList()
  68 {
  69         Clear();
  70 }
  71
  72
  73 bool CKnownFileList::Init()
  74 {
  75         CFile file;
  76
  77         CPath fullpath = CPath(theApp->ConfigDir + m_filename);
  78         if (!fullpath.FileExists()) {
  79                 // This is perfectly normal. The file was probably either
  80                 // deleted, or this is the first time running aMule.
  81                 return false;
  82         }
  83
  84         if (!file.Open(fullpath)) {
  85                 AddLogLineC(CFormat(_("WARNING: %s cannot be opened.")) % m_filename);
  86                 return false;
  87         }
  88
  89         try {
  90                 uint8 version = file.ReadUInt8();
  91                 if ((version != MET_HEADER) && (version != MET_HEADER_WITH_LARGEFILES)) {
  92                         AddLogLineC(_("WARNING: Known file list corrupted, contains invalid header."));
  93                         return false;
  94                 }
  95
  96                 wxMutexLocker sLock(list_mut);
  97                 uint32 RecordsNumber = file.ReadUInt32();
  98                 AddDebugLogLineN(logKnownFiles, CFormat(wxT("Reading %i known files from file format 0x%2.2x."))
  99                         % RecordsNumber % version);
 100                 for (uint32 i = 0; i < RecordsNumber; i++) {
 101                         CScopedPtr<CKnownFile> record;
 102                         if (record->LoadFromFile(&file)) {
 103                                 AddDebugLogLineN(logKnownFiles,
 104                                         CFormat(wxT("Known file read: %s")) % record->GetFileName());
 105                                 Append(record.release());
 106                         } else {
 107                                 AddLogLineC(_("Failed to load entry in known file list, file may be corrupt"));
 108                         }
 109                 }
 110                 AddDebugLogLineN(logKnownFiles, wxT("Finished reading known files"));
 111
 112                 return true;
 113         } catch (const CInvalidPacket& e) {
 114                 AddLogLineC(_("Invalid entry in known file list, file may be corrupt: ") + e.what());
 115         } catch (const CSafeIOException& e) {
 116                 AddLogLineC(CFormat(_("IO error while reading %s file: %s")) % m_filename % e.what());
 117         }
 118
 119         return false;
 120 }
 121
 122
 123 void CKnownFileList::Save()
 124 {
 125         CFile file(theApp->ConfigDir + m_filename, CFile::write_safe);
 126         if (!file.IsOpened()) {
 127                 return;
 128         }
 129
 130         wxMutexLocker sLock(list_mut);
 131         AddDebugLogLineN(logKnownFiles, CFormat(wxT("start saving %s")) % m_filename);
 132
 133         try {
 134                 // Kry - This is the version, but we don't know it till
 135                 // we know if any largefile is saved. This allows the list
 136                 // to be compatible with previous versions.
 137                 bool bContainsAnyLargeFiles = false;
 138                 file.WriteUInt8(0);
 139
 140                 file.WriteUInt32(m_knownFileMap.size() + m_duplicateFileList.size());
 141
 142                 // Duplicates handling. Duplicates needs to be saved first,
 143                 // since it is the last entry that gets used.
 144                 KnownFileList::iterator itDup = m_duplicateFileList.begin();
 145                 for ( ; itDup != m_duplicateFileList.end(); ++itDup ) {
 146                         (*itDup)->WriteToFile(&file);
 147                         if ((*itDup)->IsLargeFile()) {
 148                                 bContainsAnyLargeFiles = true;
 149                         }
 150                 }
 151
 152                 CKnownFileMap::iterator it = m_knownFileMap.begin();
 153                 for (; it != m_knownFileMap.end(); ++it) {
 154                         it->second->WriteToFile(&file);
 155                         if (it->second->IsLargeFile()) {
 156                                 bContainsAnyLargeFiles = true;
 157                         }
 158                 }
 159
 160                 file.Seek(0);
 161                 file.WriteUInt8(bContainsAnyLargeFiles ? MET_HEADER_WITH_LARGEFILES : MET_HEADER);
 162                 file.Close();
 163         } catch (const CIOFailureException& e) {
 164                 AddLogLineC(CFormat(_("Error while saving %s file: %s")) % m_filename % e.what());
 165         }
 166         AddDebugLogLineN(logKnownFiles, CFormat(wxT("finished saving %s")) % m_filename);
 167 }
 168
 169
 170 void CKnownFileList::Clear()
 171 {
 172         wxMutexLocker sLock(list_mut);
 173
 174         DeleteContents(m_knownFileMap);
 175         DeleteContents(m_duplicateFileList);
 176         ReleaseIndex();
 177 }
 178
 179
 180 CKnownFile* CKnownFileList::FindKnownFile(
 181         const CPath& filename,
 182         time_t in_date,
 183         uint64 in_size)
 184 {
 185         wxMutexLocker sLock(list_mut);
 186
 187         if (m_knownSizeMap) {
 188                 std::pair<KnownFileSizeMap::const_iterator, KnownFileSizeMap::const_iterator> p;
 189                 p = m_knownSizeMap->equal_range((uint32) in_size);
 190                 for (KnownFileSizeMap::const_iterator it = p.first; it != p.second; it++) {
 191                         CKnownFile *cur_file = it->second;
 192                         if (KnownFileMatches(cur_file, filename, in_date, in_size)) {
 193                                 return cur_file;
 194                         }
 195                 }
 196         } else {
 197                 for (CKnownFileMap::const_iterator it = m_knownFileMap.begin();
 198                          it != m_knownFileMap.end(); ++it) {
 199                         CKnownFile *cur_file = it->second;
 200                         if (KnownFileMatches(cur_file, filename, in_date, in_size)) {
 201                                 return cur_file;
 202                         }
 203                 }
 204         }
 205
 206         return IsOnDuplicates(filename, in_date, in_size);
 207 }
 208
 209
 210 CKnownFile *CKnownFileList::IsOnDuplicates(
 211         const CPath& filename,
 212         uint32 in_date,
 213         uint64 in_size) const
 214 {
 215         if (m_duplicateSizeMap) {
 216                 std::pair<KnownFileSizeMap::const_iterator, KnownFileSizeMap::const_iterator> p;
 217                 p = m_duplicateSizeMap->equal_range((uint32) in_size);
 218                 for (KnownFileSizeMap::const_iterator it = p.first; it != p.second; it++) {
 219                         CKnownFile *cur_file = it->second;
 220                         if (KnownFileMatches(cur_file, filename, in_date, in_size)) {
 221                                 return cur_file;
 222                         }
 223                 }
 224         } else {
 225                 for (KnownFileList::const_iterator it = m_duplicateFileList.begin();
 226                          it != m_duplicateFileList.end(); ++it) {
 227                         CKnownFile *cur_file = *it;
 228                         if (KnownFileMatches(cur_file, filename, in_date, in_size)) {
 229                                 return cur_file;
 230                         }
 231                 }
 232         }
 233         return NULL;
 234 }
 235
 236
 237 CKnownFile* CKnownFileList::FindKnownFileByID(const CMD4Hash& hash)
 238 {
 239         wxMutexLocker sLock(list_mut);
 240
 241         if (!hash.IsEmpty()) {
 242                 if (m_knownFileMap.find(hash) != m_knownFileMap.end()) {
 243                         return m_knownFileMap[hash];
 244                 } else {
 245                         return NULL;
 246                 }
 247         }
 248         return NULL;
 249
 250 }
 251
 252
 253 bool CKnownFileList::SafeAddKFile(CKnownFile* toadd, bool afterHashing)
 254 {
 255         bool ret;
 256         {
 257                 wxMutexLocker sLock(list_mut);
 258                 ret = Append(toadd, afterHashing);
 259         }
 260         if (ret) {
 261                 theApp->searchlist->UpdateSearchFileByHash(toadd->GetFileHash());
 262         }
 263         return ret;
 264 }
 265
 266
 267 bool CKnownFileList::Append(CKnownFile *Record, bool afterHashing)
 268 {
 269         if (Record->GetFileSize() > 0) {
 270                 const CMD4Hash& tkey = Record->GetFileHash();
 271                 CKnownFileMap::iterator it = m_knownFileMap.find(tkey);
 272                 if (it == m_knownFileMap.end()) {
 273                         m_knownFileMap[tkey] = Record;
 274                         return true;
 275                 } else {
 276                         CKnownFile *existing = it->second;
 277                         if (KnownFileMatches(Record, existing->GetFileName(), existing->GetLastChangeDatetime(), existing->GetFileSize())) {
 278                                 // The file is already on the list, ignore it.
 279                                 AddDebugLogLineN(logKnownFiles, CFormat(wxT("%s is already on the list")) % Record->GetFileName().GetPrintable());
 280                                 return false;
 281                         } else if (IsOnDuplicates(Record->GetFileName(), Record->GetLastChangeDatetime(), Record->GetFileSize())) {
 282                                 // The file is on the duplicates list, ignore it.
 283                                 // Should not happen, at least not after hashing. Or why did it get hashed in the first place then?
 284                                 AddDebugLogLineN(logKnownFiles, CFormat(wxT("%s is on the duplicates list")) % Record->GetFileName().GetPrintable());
 285                                 return false;
 286                         } else {
 287                                 if (afterHashing && existing->GetFileSize() == Record->GetFileSize()) {
 288                                         // We just hashed a "new" shared file and find it's already known under a different name or date.
 289                                         // Guess what - it was probably renamed or touched.
 290                                         // So copy over all properties from the existing known file and just keep name/date.
 291                                         time_t newDate = Record->GetLastChangeDatetime();
 292                                         CPath newName = Record->GetFileName();
 293                                         CMemFile f;
 294                                         existing->WriteToFile(&f);
 295                                         f.Reset();
 296                                         Record->LoadFromFile(&f);
 297                                         Record->SetLastChangeDatetime(newDate);
 298                                         Record->SetFileName(newName);
 299                                 }
 300                                 // The file is a duplicated hash. Add THE OLD ONE to the duplicates list.
 301                                 // (This is used when reading the known file list where the duplicates are stored in front.)
 302                                 m_duplicateFileList.push_back(existing);
 303                                 if (theApp->sharedfiles) {
 304                                         // Removing the old kad keywords created with the old filename
 305                                         theApp->sharedfiles->RemoveKeywords(existing);
 306                                 }
 307                                 m_knownFileMap[tkey] = Record;
 308                                 return true;
 309                         }
 310                 }
 311         } else {
 312                 AddDebugLogLineN(logGeneral,
 313                         CFormat(wxT("%s is 0-size, not added")) %
 314                         Record->GetFileName());
 315
 316                 return false;
 317         }
 318 }
 319
 320 // Make an index by size to speed up FindKnownFile
 321 // Size modulo 2^32 is enough here
 322 void CKnownFileList::PrepareIndex()
 323 {
 324         ReleaseIndex();
 325         m_knownSizeMap = new KnownFileSizeMap;
 326         for (CKnownFileMap::const_iterator it = m_knownFileMap.begin(); it != m_knownFileMap.end(); it++) {
 327                 m_knownSizeMap->insert(std::pair<uint32, CKnownFile*>((uint32) it->second->GetFileSize(), it->second));
 328         }
 329         m_duplicateSizeMap = new KnownFileSizeMap;
 330         for (KnownFileList::const_iterator it = m_duplicateFileList.begin(); it != m_duplicateFileList.end(); it++) {
 331                 m_duplicateSizeMap->insert(std::pair<uint32, CKnownFile*>((uint32) (*it)->GetFileSize(), *it));
 332         }
 333 }
 334
 335
 336 void CKnownFileList::ReleaseIndex()
 337 {
 338         delete m_knownSizeMap;
 339         delete m_duplicateSizeMap;
 340         m_knownSizeMap = NULL;
 341         m_duplicateSizeMap = NULL;
 342 }
 343
 344 // File_checked_for_headers