So how can a programmer use described features? Possibilities are endless:

1. Download a filelist and get a specified file from zip file only
2. Multi-threaded file downloads
3. Allow end-users to pause/resume downloads of large files

Here is simple code to implement first example (downloading a specified file from zip):

  LMDWebHTTPGet1.URL := 'http://www.lmd.de/downloads/lmd2010vcl/setupse10d14.zip';
  // "end of central directory" size.
  // Minus shows that we read it from end of file
  LMDWebHTTPGet1.RangeEnd := -17;
  // Downloads last 17 bytes of file
  if LMDWebHTTPGet1.Process(False, False) then
  begin
    LEndOfCentralRecord.LoadFromStream(LMDWebHTTPGet1.Data);
    // Fills RangeStart by "Central Directory" structure begin offset
    LMDWebHTTPGet1.RangeStart := LEndOfCentralRecord.OffsetOfCentralDirectory;
    LMDWebHTTPGet1.RangeEnd := LMDWeHTTPGet1.RangeStart + LEndOfCentralRecord.SizeOfCentralDirectory;
    // Downloads "Central Directory" structure
    if LMDWebHTTPGet1.Process(False, False) then
    begin
      LCentralDirectory.LoadFromStream(LMDWebHTTPGet1.Data);
      // Fills RangeStart by offset of compressed file in ZIP
      LMDWebHTTPGet1.RangeStart := LCentralDirectory.Files['readme.txt'].FileOffset;
      // Compressed size
      LMDWebHTTPGet1.RangeEnd := LMDWebHTTPGet1.RangeStart + LCentralDirectory.Files['readme.txt'].CompressedSize;
      // Downloads compressed file "readme.txt"
      if LMDWebHTTPGet1.Process(False, False) then
      begin
        // Processing of compressed file
      end;
    end;
  end;

LEndOfCentralRecord, LCentralDirectory are placeholders that schematically shows corresponding zip format structures – “Central directory structure” and “End of central directory record”. The full specification is described here.

Besides that WebPack includes a demo project in demos\lmdweb\LMDWebHTTPGetResume folder.