Ole Christian Eidheim

Happy to hear you are getting closer to a solution.

Thank you for starting this discussion, it is very interesting. First off, `get_future().wait()` is blocking and should not be run inside an io_service task. Secondly, the PPL Tasks are interesting,...

JavaScript's Promise is actually my favourite way of chaining async calls, but to my knowledge this is currently not possible to do with c++'s (boost::)asio. One problem I guess would...

In http://www.boost.org/doc/libs/1_64_0/doc/html/thread/synchronization.html#thread.synchronization.futures.reference.unique_future.then you have the line (under Notes): "The returned futures behave as the ones returned from boost::async, the destructor of the future object returned from then will block. This...

In your above link, synchronous_job::run is blocking. Our case is a bit different than the one described in the link, since both the promise/future and the callback should run in...

I was a bit unclear above, what I mean by running in the same thread, is that `future::wait()` should never be called, and `future::get` should only be called within the...

Thank you for reporting this. Do you have the full backtrace? Also, if you have a simple test case reproducing this access violation it would be great. I failed to...

I'm no help here sadly since I do not use Windows, and to me it seems like a boost on Windows issue. Maybe upgrading boost will help? Hopefully, someone with...

Feel free to create a pull request on this. I'm a bit busy in the following weeks, but I agree that this should be added.

When using strings as keys, the default map type goto is `std::unordered_map` I would say. The nice thing with constructors is that there can be potentially no unnecessary copies of...