On Tue, Nov 12, 2013 at 8:34 PM, Niall Douglas
Boost community feedback is requested for the formal peer review of the TypeIndex library by Antony Polukhin. Feedback requested includes:
1. Should this library be accepted into Boost?
YES. First, because the documentation clarified issues I do have but didn't realized I have yet (I guess I would have found out once I start porting my project to another compiler). I do use a lot of type_index as keys in different containers (including tbb concurrent associative containers) and a lot of the ids will come from runtime shared libraries loaded and unloaded when needed. (mostly I use them in different type-based message dispatching systems) So the documentation was a surprise and I'm happy to see that there are potential solutions. Second: I'm ok with either way of doing, to make boost::type_id and boost::type_index similar to std:: ones or to have a more consistent behaviour across platforms for boost ones. I'm not sure which one is preferable, but I would make sure to use in my own code the way that would allow me to have the exact (or almost exact) same behaviour whatever the platform (if possible). Whatever the way this library ends, I think I should used it as, as I was saying, I'm using a lot type index. One question though, I'm not totally clear if there is a remarkable difference in performance and binary size between with RTTI using std and without RTTI using this library. I understand that performance wasn't the concern but I might have missed useful indication on potential consequences of using this library. I read most of the documentation though I think I should re-read some parts as I was interrupted frequently through my reading.
2. Any conditions which should be attached to acceptance into Boost e.g. fixes, additional testing, changes to documentation. Please be as specific as possible here (bullet points are good!)
I don't have a particular condition, I'm not knowledgeable enough in the details of the library to even participate to the discussion.