On 20 Apr 2012, at 17:47, Igor Stasenko wrote: > Yes, you can look at my hacky implementation of SCouchDB, where i > partially implemented > the HTTP protocol and use streams for content. See > http://www.squeaksource.com/SCouchDB > > there i use a stream to directly parse the input without waiting all > data to arrive. > This is also useful for up/downloading huge files, where content size > can easily surpass the amount of RAM available.
OK, I will have a look, but I don't think that will answer my questions. Zn already nicely handles streaming through its ZnStreamingEntity object: it captures all metadata and a reference to the stream and leaves it up to its user to drain it or to fill it. And indeed, that is the way to handle things like large uploads or downloads while not taking everything in memory at once. Sven
