Eric, thanks for your positive review. Eric Niebler wrote:
John Torjo wrote:
The formal review of the Boost.Functional/Forward library, proposed by Tobias Schwinger, begins today :
<snip>
* What is your evaluation of the design?
Good. I can think of a couple of missing features.
- An option to handle const and non-const arguments up to N, and only [non-]const arguments from N to M.
That sentence is about a single scenario (not two or three), correct? What use cases do you have in mind?
- Automatically detect if the compiler supports variadic templates and rvalue references and use those instead. This can be a useful portability utility when such compilers become more common.
Yes, that seems a good idea.
* What is your evaluation of the documentation?
Short and to the point. Would be nice if the reference section for each of the adaptors had a little example usage.
IIRC, I just failed to come up with small but meaningful samples. Maybe I should loosen the criteria, figuring a stupid example might still be better than no example ;-). Regards, Tobias