Hi all, This is my first review for Boost, since I am not a great C++ expert. Nevertheless, Boost.Coroutine addresses a problem that I was recently facing so I could form a valid opinion. The problem that I was facing was the need to interrupt, and later resume, the dijkstra_shortest_path algorithm. Following the documentation for Boost.Coroutine is was straightforward to rewrite the original function to a coroutine, and the result works just as it should with minimal debugging and no hidden surprises. To me this means that the library is extraordinary user-friendly, and it gives me control over my algorithms in a way that I did not have before. I therefore vote *FOR* acceptance. On 03/09/2012 12:46, Hartmut Kaiser wrote:
- What is your evaluation of the design? Good. It is very easy to use. - What is your evaluation of the implementation? Can't tell. - What is your evaluation of the documentation? Looks good. However, for a casual user like me it seems more relevant to understand the control flow of the coroutines than details about stack space and exceptions. All the 'Notes' and 'Warnings' are very prominent at the cost of the main points to be understood.
- What is your evaluation of the potential usefulness of the library? Good - Did you try to use the library? With what compiler? Did you have any problems? I used it with VS2010, without any problems. - How much effort did you put into your evaluation? A glance? A quick reading? In-depth study? I just used the library to solve my particular case, and read the documentation for as far as necessary. I did not look at the implementation at all. I did not investigate the performance, but did verify the results. My experiment is at http://people.pwf.cam.ac.uk/ahh34/dijkstra.zip
- Are you knowledgeable about the problem domain? No. Regards Hartmut Review Manager Kind regards, Alex