Hi,

I was testing a big download with 2.3.3 on Linux using j2sdk-1.4.0 and see a strange problem with a file called "******". This file is (according to Limewire) 374,802K big. I have repeatedly tried to download it but for some reason it always starts over when it has downloaded the 374,802K. Another strange thing I see is this: right now I am downloading it again (previous stuff deleted from incomplete dir) and currently it says that the progress is 11%. So the file in the incomplete directory should show 11% of 374,802K = approx 41,228K. But, if I press refresh in the incomplete dir I see the size of the file is about 215 Megs. Either this is a problem with the file or Limewire is messing up somewhere. I had the same problem on Win98 also. Any ideas?

Greets,
Puzzled

The download starting over may be to do with how you specified the download... If you selected 'Download All' on a group of the same file - then that is what it will do... download all the files in the group. You should choose 'Download Any' if you just want the one copy of the file.

The file size thing relates to swarming, I think. When LimeWire can get a file from more than one source then it gets different bits of the file from the different sources. It will ask for an early chunk and some later chunks. LimeWire isn't very clever about the way it stores the chunks - It just extends the size of the file so that it is big enough to include the last chunk. The bit between the chunks gets filled in eventually. In the mean time, you have a large file but only a small percentage of it actually downloaded.

This is fairly wasteful of space. Hopefully this is one of the things that will be improved along with Hashing.

Mark

the problem with this is that these spots of blank space sometimes don't get filled in. i've dled movies before where they're missing huge chunks from the middle when they're officially "done" downloading. since these are movies, it gets a bit annoying and crashes the player. i've tried to reprocess the file & get limewire to fill in the missing chunks, but it doesn't work. the only solution i've found was to redownload the whole movie (which also always has missing chunks) and manually go through both of them with a video editor and pull out parts of one to patch into the other. as you can probably guess, this gets rather annoying after the 4th or 5th time. hopefully this will get fixed once we get hashing working in limewire, but until then, that's the way things go.