[Live-devel] How to implement non-blocking fashion in custom doGetNextFrame()

[Cristiano Belloni]

xcsmith at rockwellcollins.com wrote:
> 1. Are you inheriting from FramedSource, from DeviceSource, or 
> modifying DeviceSource?
Inheriting from FramedSource.
> 2. Is "isAwaitingData()" function an abbreviation for 
> "isCurrentlyAwaitingData()", or are these separate functions?
It's a separate function. Returns true if I have a new frame ready.
> 3. When you copy your data to fTo in deliverFrame(), are you using 
> "memcpy()" or similar function to write the data?
> 4. Does your "deliverFrame()" call "afterGetting()"?  You have to call 
> "afterGetting()" so that the downstream filter / sink can begin 
> handling your data. (Ross said this on the last thread post)
memcpy.  And yes, it calls afterGetting()
> 5. Does your "afterGetting() function set fIsCurrentlyAwaitingData? If 
> you don't set fIsCurrentlyAwaitingData to False, then as you see in 
> FramedSource::getNextFrame(), the source->sink chain will quit. 
> FramedSource.cpp line 64.
Err.. I don't define any afterGetting
>
>
> Ross provided a general solution email earlier.  If you have more 
> questions, I'd be interested in looking at what you've written.  I 
> enjoy working with the filter code.  The source you've posted isn't 
> enough for me to help more specifically.  Can you post all / more of 
> your file or just send it to me?
> ~xo
Currently I'm experimenting modifying taskScheduler, and I'm rewriting 
the code. A soon as I have some readable code, I'll gladly send it to you.

Many thanks,
Regards,

Cristiano
-- 
Belloni Cristiano
Imavis Srl.
www.imavis.com <http://www.imavis.com>
belloni at imavis.com <mailto://belloni@imavis.com>