Re: http retriever (sign me up)

[Joey Hess <joeyh@debian.org>]

David Whedon wrote:
> > How does this compare to BusyBox wget?
> 
> Probably not favorably.  Actually after playing with it more the only thing
> going for it is that it is tiny.  It isn't a drop in solution.  I actually
> wasn't aware that wget was in busybox.

I'd love to have a http retreiver, it would make the demo system I threw
together a lot more impressive. :-)

I think that the actual http transport program is just a small corner of
the http retreiver. If we assume some wget-ish program, that supports
"<program> <url>" style command lines and outputs to a standard location,
we still need the glue code around it to make it conform to the
retreiver interface, and provide the base of the url to it. The base of
the url has to be picked somehow, probably we want to have a mirror
list that is presented with debconf, with sensible defaults. I think all
this can be written more or less independant of whatever wget-like
program turns out to be smallest.

-- 
see shy jo