On 15 Aug 2001, at 19:22, Andreas Heck wrote:

> And here is the bug: after continuing the download from another ftp Server
> wget says that it has downloaded 337% of the file instead of 77%. It counts
> up to 430% when the file is finished.

I don't think the fact that the second half of the transfer was from a 
different server has anything to do with it, assuming they are running 
similar FTP servers.

This is due to an inconsistency in the way FTP servers report the size 
on resuming an FTP transfer. Some report the total file size and others 
report the remaining size. Wget currently compensates for this if it 
knows what the file size should be (e.g. when doing a recursive 
directory retrieval) but when it is asked to download a single file, it 
currently does not know how big it is beforehand. Wget could be 
modified to find this out easily enough but this would introduce an 
extra delay before the proper transfer was resumed. I suppose that is 
acceptable in this case, as it would only need to do this extra step 
for this special case.

Reply via email to