Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [cdt-dev] EDC and asynchronous operations

At 03:59 PM 10/12/2010, Pawel Piech wrote:
On 10/12/2010 01:41 PM, John Cortell wrote:
2. ACPM seems to me a good tool for the client of an asynchronous API. However, It's not clear to me that an asynchronous API should expose ACPM objects, which I think is what Pawel was hinting at.
Correct, it is what I was getting at.  I'm also not clear on how much value of such change would bring.  Personally, I'm not as convinced of the benefit of the data coherency story, simply because we haven't really had many data coherency problems in the UI integration use cases so far.  However, I do see a lot of potential value in simplifying the programming model away from using callbacks as we do now.  But even if we do show that there there is clear advantage to exposing cache objects in APIs, it may not be enough for existing DSF debuggers to decide to migrate to this API style.  Either way, DSF should be able to evolve to accommodate this programming model.  So I hope that we can collectively do some prototyping and start answering these questions.

Indeed. I very much like the idea of collecting a variety of information in bulk, via a single asynchronous call, then coding the consuming logic in a simplified, synchronous style. That to me is the single biggest advantage of ACPM, since the biggest disadvantage of DSF and TCF for me is the disjointed programming flow those APIs impose on the client. That flow not only makes it difficult to program to the API, but it makes debugging it equally, if not more, difficult.


Back to the top