Re:Why your web framework should not adopt Rack API
by weavejester on Friday June 15, 2012 @11:59AM
(#4117124)
Ring supports streaming (via Java InputStreams), but not non-blocking I/O.
[ Parent
| on HN ]
Re:Why your web framework should not adopt Rack API
by andrewvc on Friday June 15, 2012 @12:09PM
(#4117173)
Ah, good to know James!
Of course non-blocking is what people want for stuff like websockets in most situations.
Re:Why your web framework should not adopt Rack API
by weavejester on Friday June 15, 2012 @12:12PM
(#4117191)
Right, websockets servers tend to need NIO otherwise you just wind up with too many threads to manage. However, streaming has uses outside of websockets; for instance, returning a response that's too large to want to hold in memory.
Re:Why your web framework should not adopt Rack API
by gerrit on Friday June 15, 2012 @12:54PM
(#4117406)
Or simply flushing out the <head> of a html document before the body is generated so the browser can start pre-fetching resources. That's the main use case that motivated Rails to implement streaming.