Jochen Roderburg <[email protected]> writes: > I have to admit of course, that the combination of a usable > Modify-Date and a Content-Disposition filename will be very rare "in > the wild", but nevertheless possible. I have seen the > Content-Disposition headers mostly when the reply data is dynamically > generated by some web application. Currently I have not even a test > case at hand, a few that I knew have all disappeared from the net. > > I would still like to know how the flow of events is now in the cases > that you have changed. > > In my understanding the "old" logic was that wget issues a HEAD > request in *all* cases where it needs additional info from the headers > before the real data download starts or to decide that it does not > need to download anything. > > And now it gets this info directly from the reply headers of a GET > request? And after having read all these headers it pauses and decides > how to continue? And it stops reading further when it finds out that > no real download is necessary?
Previously, wget *always* sent a HEAD request when --content-disposition was specified. Now wget sends the additional HEAD request only if it can really help us to save a GET, in other words: only when -N is specified. Cheers, Giuseppe
