When mirroring, download can be interrupted and later resumed thanks to the --timestamping option contained in --mirror. However, wget still accesses all files to check whether they have been updated.
When mirroring large sites with lots of small files, this results in a much wasted time, especially on a slow connection and if the download was interrupted just some minutes or hours before (thus no files are likely to have been updated). Is there a way to prevent wget from even checking files for update, and to really pick up the downloading from where it had been interrupted, i.e. without having to redescend the tree from the original starting document again? Maybe this could be faked by using wget's log? Massimiliano Mirra
