Re: wwwoffle make GET from HEAD

Joerg Sommer <[email protected]> writes:

> if I connect to my wwwoffle and make "HEAD http://www.test.de/ HTTP/1.1"
> wwwoffle makes a get on the outside.
> 
> This is a problem, because my browser makes a HEAD before downloading a
> file and makes in a second call the GET. I use "lock-file = yes". On big
> files the second call is denied from getting the file, because its locked
> by the first HEAD call.

What browser are you using that makes a HEAD request followed by a GET
request?

> Why wwwoffle makes a get, when I ask for head? If I check my bookmarks, I
> take a "wget -i bookmarks.html --spider". This cause wwwoffle to download
> all pages named in my bookmark file.

WWWOFFLE makes a GET request when you ask for a HEAD because it does
not know that you are going to ask for a GET later and the HEAD data
would be useless or worse.

If you ask for the HEAD of page that you don't have cached already
then WWWOFFLE can do nothing with the data so it would have to throw
it away (WWWOFFLE will cache the whole thing or nothing, it won't
cache just the HEAD).  This is a waste of time.

If you ask for a HEAD of a page that is already cached and nothing has
changed then WWWOFFLE will have made a conditional GET anyway so you
get approximately the same behaviour.  Nothing lost here.

If you ask for the HEAD of a page that is already cached and it has
changed then WWWOFFLE will have to throw away either the HEAD or the
cached file. (it caches all or nothing, either it keeps what it has or
it throws it all away).  This has now destroyed the cached data which
is worse than the other cases.

-- 
Andrew.
----------------------------------------------------------------------
Andrew M. Bishop                             amb-Uxr6IM1mbv2TY6FTCsQk+9Bc4/[email protected]
                                      http://www.gedanken.demon.co.uk/

WWWOFFLE users page:
        http://www.gedanken.demon.co.uk/wwwoffle/version-2.8/user.html