On Sun, Jun 12, 2005 at 09:45:39PM -0400, seth vidal wrote: > On Sun, 2005-06-12 at 21:27 -0400, Brian Long wrote: > > Would it make sense to add a feature / work-around which would be a > > fall-back mechanism when HTTP byte-range requests fail? The only > > work-around I see would be to download the entire RPM and strip off the > > correct byte-range from the file instead of over HTTP. Is this a > > ridiculous request? > > not really. Though I would recommend: > 1. supporting the http 1.1 spec ;) > 2. seeing if you can switch them to ftp, briefly. > > Michael, Ryan, > Do you remember what we had discussed/decided on wrt a fallback mode > for systems not supporting byte-ranges? I do not remember. I think at the time it was basically "well, they must support byteranges". urlgrabber does byteranges for local files pretty reliably, so it would not actually be that hard to: a) download the whole file b) fetch the byterange from the local file That's obviously very very painful. There are a couple of other options. One could do a stupid byterange implementation in which you start downloading the whole file and then stop when you get the part you're interested in. That's a bit hackish, but not all that different from how we do byteranges in ftp (except we use reget to start in the right place). This is getting strange enough that it's probably not the sort of thing that should go in urlgrabber. It can be implemented outside of urlgrabber with grabber.urlopen() -Michael -- Michael D. Stenner mstenner@xxxxxxxxxxxxxxx ECE Department, the University of Arizona 520-626-1619 1230 E. Speedway Blvd., Tucson, AZ 85721-0104 ECE 524G
