Chris McDonough wrote:
> Actually, "stream iterator" support is independent of blobs.

Right. Also, dealing with long-living requests or dribbling data piece
by piece is substantially something different than streaming large data
to the user agent without holding application resources forever. I'm not
sure Zope 2 actually differentiates between those two.

> There was some talk a while back about Zope 3 implementing some form of
> "IResult" interface that allowed for streaming.  I'm not sure where that
> ended up, though.  I seem to remember Jim checking something in.

Yes, there is support for streaming large data without holding
application resources. What you do is write the data to a (temp)file and
return the file object. The zope.publisher will know what to do. The
"IResult" interface was made private and nowadays just an implementation

Zope3-users mailing list

Reply via email to