[Mldonkey-users] spam files (not files with misleading names)

[Goswin Brederlow]

Hi,

I noticed something realy strange happening to a download. Suddenly I
got a average rate of 400K/s for the file with only 30K/s incoming for
all files. Within seconds several megabyte of the file arrived.

First I though I connected to myself and downloaded my own file. But
that wasn't the case.

The next thing I thought of was compression. But an avi file should
not benefit from it.

Heres the entry from files.ini:

  {     file_network = Donkey
     file_md4 = "719B25EE689D8B6444E913B2858CD1F6"
     file_size = 824213504
     file_all_chunks = 
"0000000000000000000000000000000000000000000000000000000000000000000000000001111000000"
     file_state = Downloading
     file_absent_chunks = [
      (0, 729600000);
      (768512000, 824213504)]
     file_filename = "Ghost in the Shell.avi"
     file_filenames = [
      "719B25EE689D8B6444E913B2858CD1F6";
      "Ghost in the Shell.avi";
      "Ghost in the shell.avi"]
     file_age = 1043196645.021374
     file_md4s = [
      F082F4CECEC2FD1F4E50210E12BC7A47;
      "398566920EDB9154D63730EF0CBD6C65";
      CC6F8C9D02B6316706DAA54B8848EF6F;
...
      "4C91CF2C132F85A05B8B9BD25D34CC0E";
      D7DEF262A127CD79096A108E7A9FC138;
      D7DEF262A127CD79096A108E7A9FC138;
      D7DEF262A127CD79096A108E7A9FC138;
      D7DEF262A127CD79096A108E7A9FC138;
      B5E88193E15ECE688E28E4F9D5F9A6CA;
      A934CC969973E466D80DB789257DD29F;
      "52E6AD5A9C385BC7B72249ADDE63FF3E";
      "396E5DC43E564FB07B1D89261AFDE98F";
      B93F13A8AF610A74CECE0FEC4ECE59D4;
      "1B87815D74125C4FB8659292ADF79ED1"]
     file_downloaded = 38912000
     file_chunks_age = [
      1043197469.032782;
...
     file_locations = [
      {         client_addr = ("80.24.219.82", 4662)
         client_md4 = "00000000000000000000000000000000"
         client_name = ""
         client_age = 1043195147.999739
         client_last_filereqs = 0.000000
         client_checked = false
         client_overnet = false
};


Did you notice the "D7DEF262A127CD79096A108E7A9FC138;" repeating?
That appears to be the md4sum of a block of complete 0's.

Now I see three things happening:
1. someone is sharing a file that wasn't completly downloaded (or
sharing his temp directory)
2. created a spam file with a big chunk of zeros in it so people will
download the file, get to see some of it (preview doesn't show the
error) and get anoyed by broken files.


I think its pretty unlikely that anyone would want to download a file
where two 9 MB blocks are identical. Files with two identical md4sums
for two chunks should be considered broken and the user should be
informed of this. If the block is all 0s that should probably count
more.

Any reservations about that? Ideas?

MfG
        Goswin