On Apr 19, 2013 4:58 PM, "Daniel Stenberg" <daniel@haxx.se> wrote: > > On Fri, 19 Apr 2013, Carsten Bormann wrote: > >> I'd like to understand the failure model a bit better. > > > I'd just add that a resumed upload does not necessarily imply a previous upload failure. It could also be that you're appending parts over time as they "come in" or similar. Oh, maybe you'd like O_APPEND semantics? (Not that those are tightly defined, but you know what I mean.) Should HTTP have random byte range I/O and byte range locking? Should it be a filesystem or object storage protocol? No, well, maybe; HTTP is every protocol :) Nico --
Received on Saturday, 20 April 2013 05:08:10 UTC