Re: [Bug-wget] how to workaround corrupted --continue files?

[Matt Wilkie]

Thanks for that Keisal. I thought I had tried timestamping before and it 
didn't work; something about always downloading even when local file was 
current(?). Anyway, today it is working, even with --continue, so that 
will do. :)

best regards,

matt wilkie
--------------------------------------------
Geomatics Analyst
Information Management and Technology
Yukon Department of Environment
10 Burns Road * Whitehorse, Yukon * Y1A 4Y9
867-667-8133 Tel * 867-393-7003 Fax
http://environmentyukon.gov.yk.ca/geomatics/
-------------------------------------------- 



Keisial wrote:
> Matt Wilkie wrote:
>   
> > Dear wget,
> >
> > I have scripts which rely on wget to handle knowing whether a given
> > file needs to be downloaded or not. Normally this works quite well,
> > but in some circumstances (as I've learned the hard way):
> >
> >  wget -c foobar.zip  # 2009-march-01, 545kb
> >  wget -c foobar.zip  # 2009-march-15, 570kb
> >
> > In this sequence, the second archive version is 45kb larger than the
> > first.  wget just asks for the tailing bit of the second and dumps it
> > on the end of the local file: corrupt archive.
> >
> > It will also pass over an update where the second newer archive is
> > _smaller_.
> >
> > Sooo, what can we do instead? Rsync is my fav, but what do you do when
> > the data provider doesn't use it?
> >
> > thanks,
> >
> >     
> The -N option, also called --timestamping will take into account the
> timestamps, and download the file only if newer than local, so it can be
> used to replace -c in that case.
>
>
> .
>
>