Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [paho-dev] Sorry to be a pain!

I think looking forward rather than back is the right approach for a C++ interface. There is always the C interface to fall back on.


On 21/05/13 18:12, Frank Pagliughi wrote:
Yes, there is no problem for me to contribute it. I'm working on it on my own and am self-employed, so there are no company entanglements either. The Eclipse IP & licensing should not present a problem.

Currently, most of the basic async functionality is working (using the C library/API, as is), and the example apps are ported and working. I literally cut and pasted the Javadoc API as a starting point, so the coverage and match to Java should be pretty good.

As I mentioned, I used C++11 extensions (shared_ptr, mutex, & condition_variable) to get something working quickly for my own use. I will post it in this form, and then we can discuss the API, and how to pull the implementation back for older compilers. I'm thinking to write drop-in replacements for the missing std classes to keep the API intact. Look forward rather than back. Otherwise user apps would need to perform memory management.


On 05/21/2013 11:25 AM, Ian Craggs wrote:
Hey Frank,

no problem. I appreciate the feedback - it's good to know that the code is being used! And I realize that I have some thinking to do about the tokens, and also explaining more about how the APIs were intended to be used.

I'd like to get a C++ layer into Eclipse Paho too, so if you are in a position to contribute your code (Eclipse has rules about contributions related to IP, licensing, etc) that would be a big help. If not, for any reason, then I'll go ahead with my own, and take comments of course.


Back to the top