Question

Sometimes I need to deal with huge PDF files, in such case, I use some new operator in a loop, and this is a so huge loop so that more than 10 GB memory will be used. As we all know, I must use some WIN API functions to use hard disk. But I don’t know how to implement it. Anyone can help me?

Here is a code snippet:

CMyObject gg_data;
for(__int64 i = 0; i < up_limit(__int64); ++i)
{
   //add new data to the PDF file
   CCAry pData = new CCAry("myData");
   gg_data.AddRef(pData);
}

//after the loop, save file and free memory so that no memory leaks occur...
//Caution: the "save" is a very expensive work, it will take several hours!!!
gg_data.save();

Answer 1

Memory mapped files are what you want.
This maps a file into a ‘buffer’ of memory and the OS manages it for you so that you can manipulate the file as if it was a huge chunk of RAM.