commons-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Aquator <>
Subject [PIPELINE] Questions about pipeline
Date Fri, 26 Sep 2008 13:18:11 GMT

I am playing with pipeline, and I have some questions regarding to usage.

What is the suggested method of avoiding "traffic-jam" in the pipe? I mean, when a stage produces
results fast, followed by a long-running stage. I will run out of stack space in large amount
of input data.

Currently, my solution uses the context raise/registerListener methods. The slow stage notifies
it's "feeder" stage, that new data can be processed. Is there any better ideas for this problem?

My other issue is about branches. Is there a way to attach separated branches together? For
example, a stage needs input from two different branches. Is there any solution to apply a
synchronized data flow? (Lets say I have two branches, one produces A-s, and the other produces
B-s. I want a stage, that is being fed by those two branches, and produces a sequence of ABABAB...)
Is there an implementation for such behaviour?

Finally, I am interested about the various StageDrivers. I'd like some more detailed informations
then the API. Especially usage advices, samples, to help choose the best stagedriver for the
certain stages.

Thanks in advance for your time,
Istvan Cseh 

Olcsó repülőjegyet mindenkinek!
Repjegyek a legjobb napi áron akár BBP és sztornó biztosítással is.

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message