On Jun 20, 2019, at 4:20 PM, Glenn Knickerbocker wrote:
>
> On 2019-06-19 2:29 PM, Paul Gilmartin wrote:
>>>> { stage1; stage2; stage3; } | stage4
>> Stages run in order, sharing input. So what isn’t consumed by
>> the first stage is available to the second, etc. But if a stage buffers
>> reads and doesn’t process the data, the effect is unpredictable.
>
> So in Pipelines it's:
>
> (end /) ... | in: fillup | stage1 | out: faninany | stage4
> / in: | stage2 | out:
> / in: | stage3 | out:
>
> with the same caveat about buffering reads.
>
I’m somewhat surprised. I had understood that Pipelines (usually)
doesn’t consume a record until something explicitly eats it.
In your example, what causes stage1, stage2, and stage3 to
run consecutively rather than concurrently? Would FANIN rather
than FANINANY make a difference?
— gil
