I process archive:
///////////////
import std.stdio, std.zip, std.file;

int main()
{
  auto zip = new ZipArchive(read("c:/test.zip"));
  foreach (item; zip.directory) {
    writeln("processing ", item.name, " ...");
    // processing item...
  }
  return 0;
}
/////////////////
it works well for normal archives.
but how to process zip archive ~1GB ?
it takes a long of RAM.

On Saturday, 19 July 2014 at 07:55:10 UTC, AntonSotov wrote:
> I process archive:
> ///////////////
> import std.stdio, std.zip, std.file;
>
> int main()
> {
>   auto zip = new ZipArchive(read("c:/test.zip"));
>   foreach (item; zip.directory) {
>     writeln("processing ", item.name, " ...");
>     // processing item...
>   }
>   return 0;
> }
> /////////////////
> it works well for normal archives.
> but how to process zip archive ~1GB ?
> it takes a long of RAM.

Hmm... it's unfortunate that ZipArchive doesn't take a file descriptor. As a workaround, you can use memory mapping:

import std.stdio, std.zip, std.file, std.mmfile;

int main()
{
  auto mmfile = new MmFile(File("c:/test.zip", "rb"));
  auto zip = new ZipArchive(mmfile[]);
  foreach (item; zip.directory) {
    writeln("processing ", item.name, " ...");
    // processing item...
  }
  return 0;
}

On Saturday, 19 July 2014 at 10:46:31 UTC, Marc Schütz wrote:

> Hmm... it's unfortunate that ZipArchive doesn't take a file descriptor. As a workaround, you can use memory mapping:

auto mmfile = new MmFile("c:/test.zip");
Thank you. it works!

Marc Schütz:

> import std.stdio, std.zip, std.file, std.mmfile;
>
> int main()
> {
>   auto mmfile = new MmFile(File("c:/test.zip", "rb"));
>   auto zip = new ZipArchive(mmfile[]);
>   foreach (item; zip.directory) {
>     writeln("processing ", item.name, " ...");
>     // processing item...
>   }
>   return 0;
> }

This could be added to the ZipArchive online docs.

Bye,
bearophile

It seems that ti only works for zip files with less than 65000 entries, zip64 that allow read more than that do seem to be implemented.

It seems that this only works for zip files with less than 65000
entries, zip64 that allow read more than that do not seem to be
implemented.