🚀 go-pugleaf

"CJ" <cluis@yahoo.com> wrote in message 
news:472a0628@newsgroups.borland.com...

> Lets say you download half of a file, what can I store maybe
> in an .ini file that will show me where I stopped, this way I
> can resume the file again.  I don't want to use the Wininet
> cache system for the files, I would like to do it all by setting
> the pointer somewhere myself.

To do what you are asking, you will have to use your own file handle.  That 
means using InternetReadFile() and WriteFile() to download the data yourself 
manually insted of letting WinInet do it for you.  Since you would be using 
your own file handle, then you can store the file whereever you want and use 
whatever filenme you want.


Gambit 

Am I still going to use the following function in this order?

InternetOpen()
InternetConnect()
HttpOpenRequest()
HttpSendRequest()
InternetReadFile()
FileStream->Write()

Thanks




> To do what you are asking, you will have to use your own file handle. 
> That means using InternetReadFile() and WriteFile() to download the data 
> yourself manually insted of letting WinInet do it for you.  Since you 
> would be using

Forget about this code, I saw an example of something you did online
before.  There is one part in the code that it's cashing in, I am not sure
what kind of value is supposed to be in there.  Question, isn't new 
FileStream
supposed to be called before you check the FileStream->Size  ?????

TFileStream* FileStream;

1.  if(FileName != "")
2. {
3.   if(FileExists(FileName))
4.   {
5.    if(AcceptRanges.Pos("bytes") != 0)
6.    {
7.     AnsiString Range = ("Range: bytes=" + AnsiString(FileStream->Size) + 
"-\r\n");
8.     if(!HttpAddRequestHeaders(hFile, Range.c_str(), -1, 
HTTP_ADDREQ_FLAG_ADD))
9.     {
      throw Exception("unable to add range header to file request");
     }
     bIsResuming = true;
    }

------------  The full part of the code that --------------
  SaveDialog = new TSaveDialog(NULL);
  try
  {
   if(SaveDialog->Execute())
    FileName = SaveDialog->FileName;
  }
  __finally
  {
   delete SaveDialog;
  }

  if(FileName != "")
  {
   if(FileExists(FileName))
   {
    if(AcceptRanges.Pos("bytes") != 0)
    {
     AnsiString Range = ("Range: bytes=" + AnsiString(FileStream->Size) + 
"-\r\n");
     if(!HttpAddRequestHeaders(hFile, Range.c_str(), -1, 
HTTP_ADDREQ_FLAG_ADD))
     {
      throw Exception("unable to add range header to file request");
     }
     bIsResuming = true;
    }
   }

   if(!HttpSendRequest(hFile, NULL, 0, NULL, 0))
   {
    throw Exception("unable to send file request to the server");
   }
   if(bIsResuming)
    FileStream = new TFileStream(FileName, fmOpenReadWrite | 
fmShareDenyWrite);
   else
    FileStream = new TFileStream(FileName, fmCreate);

   try
   {
    if(bIsResuming)
    {
     FileStream->Seek(0, soFromEnd);
     dwBytesRead = FileStream->Position;
    }
    else
     dwBytesRead;

    ProgressBar->Position = (dwBytesRead * 100 / dwBufLen);
    ProgressBar->Max = dwContentLen;

    while(InternetReadFile(hFile, Data, 1024, &dwBytesRead))
    {
     if(dwBytesRead == 0)
      break;

     iWritten = 0;
     DataPtr = Data;

     do
     {
      iWritten = FileStream->Write(DataPtr, dwBytesRead);
      if(iWritten < 1)
       throw Exception("unable to write to local file");

      DataPtr += iWritten;
      dwBytesRead -= iWritten;

      ProgressBar->StepBy(iWritten);
      Application->ProcessMessages();
     }
     while(dwBytesRead != 0);
    }

    if(FileStream->Size != dwContentLen)
     throw Exception("File saved is incomplete");
   }
   __finally
   {
    delete FileStream;
   }
  }
 }
 __finally
 {
  if(hFile) InternetCloseHandle(hFile);
  if(hConnect) InternetCloseHandle(hConnect);
  if(hSession) InternetCloseHandle(hSession);
 }

"CJ" <cluis@yahoo.com> wrote in message 
news:472a33c4$1@newsgroups.borland.com...

> There is one part in the code that it's cashing in, I am not
> sure what kind of value is supposed to be in there.

The size of the file that was previously downloaded.

> isn't new FileStream supposed to be called before you
> check the FileStream->Size  ?????

Yes.  Here is an updated example:

    SaveDialog = new TSaveDialog(NULL);
    try
    {
        if( SaveDialog->Execute() )
            FileName = SaveDialog->FileName;
    }
    __finally
    {
        delete SaveDialog;
    }

    if( FileName != "" )
    {
        TSearchRec sr;
        if( FindFirstFile(FileName, faAnyFile, sr) )
        {
            FindClose(sr);

            if( (AcceptRanges.Pos("bytes") != 0) && (sr.Size > 0) )
            {
                AnsiString Range = ("Range: bytes=" + AnsiString(sr.Size) + 
"-\r\n");
                if( !HttpAddRequestHeaders(hFile, Range.c_str(), -1, 
HTTP_ADDREQ_FLAG_ADD) )
                    throw Exception("unable to add range header to file 
request");
                bIsResuming = true;
            }
        }

        if( !HttpSendRequest(hFile, NULL, 0, NULL, 0) )
            throw Exception("unable to send file request to the server");

        if( bIsResuming )
            FileStream = new TFileStream(FileName, fmOpenReadWrite | 
fmShareDenyWrite);
        else
            FileStream = new TFileStream(FileName, fmCreate);

        try
        {
            if( bIsResuming )
                FileStream->Position = sr.Size;

            dwBytesRead = FileStream->Position;

            ProgressBar->Position = (dwBytesRead * 100 / dwBufLen);
            ProgressBar->Max = dwContentLen;

            while( InternetReadFile(hFile, Data, 1024, &dwBytesRead) )
            {
                if( dwBytesRead == 0 )
                    break;

                iWritten = 0;
                DataPtr = Data;

                do
                {
                    iWritten = FileStream->Write(DataPtr, dwBytesRead);
                    if( iWritten < 1 )
                        throw Exception("unable to write to local file");

                    DataPtr += iWritten;
                    dwBytesRead -= iWritten;

                    ProgressBar->StepBy(iWritten);
                    Application->ProcessMessages();
                }
                while( dwBytesRead != 0 );
            }

            if( FileStream->Size != dwContentLen )
                throw Exception("File saved is incomplete");
        }
        __finally
        {
            delete FileStream;
        }
    }


Gambit 

I really do appreciate your help, this works great.

I have another question if you don't mind.

How do you download the file using two threads?  Lets
say I want to download a 9MB file without taking too
long.  I want to download I guess half of the file using
the first thread, then the next half using the second thread.

Where do I start?  I know absolutely nothing about threads.
All I have ever read about it is you better be careful or the
program will crash.  So I never messed with it at all.


Thanks again.
CJ

"CJ" <cluis@yahoo.com> wrote in message 
news:472a7a8a$1@newsgroups.borland.com...

> How do you download the file using two threads?

If the server supports the "Range" request header, then you can have one 
thread send a request for one range of bytes, and the other thread can 
request a different range of bytes.  Then simply have them download to 
different areas of the same file, or download to separate files that you can 
concatenate together later on.

> Where do I start?  I know absolutely nothing about threads.

I suggest you look at the VCL's TThread class, and look at 
Borland/CodeGear's examples that ship with the IDE.


Gambit 

> If the server supports the "Range" request header, then you can have one 
> thread send a request for one range of bytes, and the other thread can 
> request a different range of bytes.  Then simply have them download to

I have seen program do this before when copying files from one location
to another.  I don't remember the name of the program, but I have seen it
use up to four threads to copy files from one pc to another.

I still can't fathom the idea of two or more threads writing to the same 
file.

Do you know of any code examples out there that does this?  I would like
to look and see if I can understand any of it at all.

Thanks
CJ

"CJ" <cluis@yahoo.com> wrote in message 
news:472b6ef8@newsgroups.borland.com...

> I still can't fathom the idea of two or more threads writing
> to the same file.

Simply create two separate streams to the same file, position them within 
the file accordingdly, and then make sure each thread does not cross over 
into another thread's area of the file when writing to it.


Gambit 

> Simply create two separate streams to the same file, position them within 
> the file accordingdly, and then make sure each thread does not cross over 
> into another thread's area of the file when writing to it.

Thanks, I will have to experiment and try that.  Thanks for all of your
help once again.

CJ