Subject: Re: Streaming best practice: no user-defined functions with node parameters ... Do you agree?
From: Wendell Piez <wapiez@xxxxxxxxxxxxxxx>
Date: Fri, 15 Nov 2013 10:10:01 -0500
|
On Fri, Nov 15, 2013 at 9:58 AM, Michael Kay <mike@xxxxxxxxxxxx> wrote:
>> But does it mean we won't be able (when
>> streaming) to call functions even with temporary trees or snapshots?
>
> In the current draft (which we're working hard to get the final bugs out
of...) we've re-introduced a kind of data-flow analysis that enables us to
distinguish statically whether the node being passed to a function is a
streamed node or not. There's no problem with passing "ordinary" nodes to a
function (or template) parameter, it's only streamed nodes that are
problematic.
Yes, excellent. Distinguishing between streamed nodes and nodes in
trees that are entirely accessible makes sense.
Cheers, Wendell
|