1. 技术文章
  2. Using unrar library - extracting files into a filestream buffer

Using unrar library - extracting files into a filestream buffer

由 自闭症网瘾萝莉.ら 提交于 2020-01-11 20:04:43

问题


What I need is to be able to extract the files in a .rar file into streams. I'm creating a test case to get a sense of how to use unrar source. I've been searching and tinkering for a while, but I can't figure out how to use the library. I'm surprised I can't even find documentation or a tutorial for it, considering how common .rar archives are.

I've made a bit of progress on my own, but it doesn't always work. Certain files are extracted properly. Other files are jumbled up for some reason (but not completely "garbage" binary data). All I know so far is, usually (but not always):

  • not working files have fileInfo.Method = 48. They appear to be files that have a compression ratio of 100% - i.e. no compression

  • working files have fileInfo.Method = 49, 50, 51, 52, or 53, which correspond to the compression speeds, Fastest, Fast, Normal, Good, Best

But I have no idea why that is. Still can't find documentation or a working example.

Below is the test case source I have so far and an example rar archive that, when extracted with this program, has both working and not working files.

/* put in the same directory as the unrar source files
 * compiling with:
 *   make clean
 *   make lib
 *   g++ rartest.cpp -o rartest libunrar.so -lboost_filesystem
 */

#include  <cstring>
#include  <iostream>
#include  <fstream>

#include  <boost/filesystem.hpp>

#define _UNIX
#define  RARDLL
#include  "dll.hpp"

using namespace std;
namespace fs = boost::filesystem;

//char fileName[100] = "testout0.jpg\0";
//
//// doens't work
//int PASCAL ProcessDataProc(unsigned char* buffer, int buffLen) {
//  cout  << "writing..." << endl;
//  ofstream outFile(fileName);
//  cout << buffLen << endl;
//  cout << outFile.write((const char*)buffer, buffLen) << endl;
//  cout  << "done writing..." << endl;
//  fileName[7]++;
//}

int CALLBACK CallbackProc(unsigned int msg, long myBuffer, long rarBuffer, long bufferLen) {
  switch(msg) {
    case UCM_CHANGEVOLUME:
      break;
    case UCM_PROCESSDATA:
      memcpy((char*)myBuffer, (char*)rarBuffer, bufferLen);
      break;
    case UCM_NEEDPASSWORD:
      break;
  }
  return 1;
}

int main(int argc, char* argv[]) {
  if (argc != 2)
    return 0;
  ifstream archiveStream(argv[1]);
  if (!archiveStream.is_open())
    cout << "fstream couldn't open file\n";

  // declare and set parameters
  HANDLE rarFile;
  RARHeaderDataEx fileInfo;
  RAROpenArchiveDataEx archiveInfo;
  memset(&archiveInfo, 0, sizeof(archiveInfo));
  archiveInfo.CmtBuf = NULL;
  //archiveInfo.OpenMode = RAR_OM_LIST;
  archiveInfo.OpenMode = RAR_OM_EXTRACT;
  archiveInfo.ArcName = argv[1];

  // Open file
  rarFile = RAROpenArchiveEx(&archiveInfo);
  if (archiveInfo.OpenResult != 0) {
    RARCloseArchive(rarFile);
    cout  << "unrar couldn't open" << endl;
    exit(1);
  }
  fileInfo.CmtBuf = NULL;

  cout  << archiveInfo.Flags << endl;

  // loop through archive
  int numFiles = 0;
  int fileSize;
  int RHCode;
  int PFCode;
  while(true) {
    RHCode = RARReadHeaderEx(rarFile, &fileInfo);
    if (RHCode != 0) break;

    numFiles++;
    fs::path path(fileInfo.FileName);
    fileSize = fileInfo.UnpSize;

    cout << fileInfo.Method << " " << fileInfo.FileName << " (" << fileInfo.UnpSize << ")" << endl;

    char fileBuffer[fileInfo.UnpSize];

    // not sure what this does
    //RARSetProcessDataProc(rarFile, ProcessDataProc);

    // works for some files, but not for others
    RARSetCallback(rarFile, CallbackProc, (long) &fileBuffer);
    PFCode = RARProcessFile(rarFile, RAR_TEST, NULL, NULL);

    // properly extracts to a directory... but I need a stream
    // and I don't want to write to disk, read it, and delete from disk
    //PFCode = RARProcessFile(rarFile, RAR_EXTRACT, ".", fileInfo.FileName);

    // just skips
    //PFCode = RARProcessFile(rarFile, RAR_SKIP, NULL, NULL);

    if (PFCode != 0) {
      RARCloseArchive(rarFile);
      cout  << "error processing this file\n" << endl;
      exit(1);
    }
    ofstream outFile(path.filename().c_str());
    outFile.write(fileBuffer, fileSize);
  }
  if (RHCode != ERAR_END_ARCHIVE)
    cout  << "error traversing through archive: " << RHCode << endl;
  RARCloseArchive(rarFile);

  cout  << "num files: " << numFiles << endl;

}

update:

I've found a file that appears to be (claims to be?) the documentation, but according to the file, I'm not doing anything wrong. I think I might be forced to resort to CRC checking the buffers and implementing a workaround if it fails.

solution source (thanks, Denis Krjuchkov!):

/* put in the same directory as the unrar source files
 * compiling with:
 *   make clean
 *   make lib
 *   g++ rartest.cpp -o rartest libunrar.so -lboost_filesystem
 */

#include  <cstring>
#include  <iostream>
#include  <fstream>

#include  <boost/filesystem.hpp>
#include    <boost/crc.hpp>

#define _UNIX
#define  RARDLL
#include  "dll.hpp"

using namespace std;
namespace fs = boost::filesystem;

//char fileName[100] = "testout0.jpg\0";
//
//// doens't work
//int PASCAL ProcessDataProc(unsigned char* buffer, int buffLen) {
//  cout  << "writing..." << endl;
//  ofstream outFile(fileName);
//  cout << buffLen << endl;
//  cout << outFile.write((const char*)buffer, buffLen) << endl;
//  cout  << "done writing..." << endl;
//  fileName[7]++;
//}

int CALLBACK CallbackProc(unsigned int msg, long myBufferPtr, long rarBuffer, long bytesProcessed) {
  switch(msg) {
    case UCM_CHANGEVOLUME:
      return -1;
      break;
    case UCM_PROCESSDATA:
      memcpy(*(char**)myBufferPtr, (char*)rarBuffer, bytesProcessed);
      *(char**)myBufferPtr += bytesProcessed;
      return 1;
      break;
    case UCM_NEEDPASSWORD:
      return -1;
      break;
  }
}

int main(int argc, char* argv[]) {
  if (argc != 2)
    return 0;
  ifstream archiveStream(argv[1]);
  if (!archiveStream.is_open())
    cout << "fstream couldn't open file\n";

  // declare and set parameters
  RARHANDLE rarFile;  // I renamed this macro in dll.hpp for my own purposes
  RARHANDLE rarFile2;
  RARHeaderDataEx fileInfo;
  RAROpenArchiveDataEx archiveInfo;
  memset(&archiveInfo, 0, sizeof(archiveInfo));
  archiveInfo.CmtBuf = NULL;
  //archiveInfo.OpenMode = RAR_OM_LIST;
  archiveInfo.OpenMode = RAR_OM_EXTRACT;
  archiveInfo.ArcName = argv[1];

  // Open file
  rarFile = RAROpenArchiveEx(&archiveInfo);
  rarFile2 = RAROpenArchiveEx(&archiveInfo);
  if (archiveInfo.OpenResult != 0) {
    RARCloseArchive(rarFile);
    cout  << "unrar couldn't open" << endl;
    exit(1);
  }
  fileInfo.CmtBuf = NULL;

//  cout  << archiveInfo.Flags << endl;

  // loop through archive
  int numFiles = 0;
  int fileSize;
  int RHCode;
  int PFCode;
  int crcVal;
  bool workaroundUsed = false;
    char currDir[2] = ".";
    char tmpFile[11] = "buffer.tmp";
  while(true) {
    RHCode = RARReadHeaderEx(rarFile, &fileInfo);
    if (RHCode != 0) break;
    RARReadHeaderEx(rarFile2, &fileInfo);

    numFiles++;
    fs::path path(fileInfo.FileName);
    fileSize = fileInfo.UnpSize;
    crcVal = fileInfo.FileCRC;

    cout << dec << fileInfo.Method << " " << fileInfo.FileName << " (" << fileInfo.UnpSize << ")" << endl;
    cout << " " << hex << uppercase << crcVal << endl;

    char fileBuffer[fileSize];
    char* bufferPtr = fileBuffer;

    // not sure what this does
    //RARSetProcessDataProc(rarFile, ProcessDataProc);

    // works for some files, but not for others
    RARSetCallback(rarFile, CallbackProc, (long) &bufferPtr);
    PFCode = RARProcessFile(rarFile, RAR_TEST, NULL, NULL);

    // properly extracts to a directory... but I need a stream
    // and I don't want to write to disk, read it, and delete from disk
//    PFCode = RARProcessFile(rarFile, RAR_EXTRACT, currDir, fileInfo.FileName);

    // just skips
    //PFCode = RARProcessFile(rarFile, RAR_SKIP, NULL, NULL);

    if (PFCode != 0) {
      RARCloseArchive(rarFile);
      cout  << "error processing this file\n" << endl;
      exit(1);
    }

    // crc check
    boost::crc_32_type crc32result;
    crc32result.process_bytes(&fileBuffer, fileSize);
    cout << " " << hex << uppercase << crc32result.checksum() << endl;

    // old workaround - crc check always succeeds now!
    if (crcVal == crc32result.checksum()) {
      RARProcessFile(rarFile2, RAR_SKIP, NULL, NULL);
    }
    else {
      workaroundUsed = true;
      RARProcessFile(rarFile2, RAR_EXTRACT, currDir, tmpFile);
      ifstream inFile(tmpFile);
      inFile.read(fileBuffer, fileSize);
    }

    ofstream outFile(path.filename().c_str());
    outFile.write(fileBuffer, fileSize);
  }
  if (workaroundUsed) remove(tmpFile);
  if (RHCode != ERAR_END_ARCHIVE)
    cout  << "error traversing through archive: " << RHCode << endl;
  RARCloseArchive(rarFile);

  cout  << dec << "num files: " << numFiles << endl;

}

回答1:


I'm not familar with unrar, after quick reading of a documentation I think you are assuming that CallbackProc is called exactly once per file. However, I think unrar may call it multiple times. It unpacks some data then calls CallbackProc, then unpacks next data chunk and again calls CallbackProc, the process is iterated until all data is processed. You should remember how many bytes are actually written to buffer, and append new data at corresponding offset.




回答2:


I can't find any docs online either, but there are examples you can use:

Go to http://www.krugle.com, and in the lower left corner of the page, enter a keyword like RAROpenArchiveEx. You'll see header and source files from various open source projects that make use of unrar library.

That should get you going.




回答3:


You seem to have posted some source-code, but no actual question.

Have you considered looking at Rarlabs Feedback Page (which points to their forums

Also, see: This Article



来源：https://stackoverflow.com/questions/2327784/using-unrar-library-extracting-files-into-a-filestream-buffer

标签
c++
compression
unrar
易学教程内所有资源均来自网络或用户发布的内容，如有违反法律规定的内容欢迎反馈！
该文章没有解决你所遇到的问题?点击提问,说说你的问题,让更多的人一起探讨吧!