On Monday, September 28, 2015 at 3:35:37 PM UTC-5, Edward Diener wrote:

...

Hi All,

Boost.Fit is a header-only function utility library for C++11. It supports both gcc, clang and visual studio. More info can be found here:

https://github.com/pfultz2/Fit

I would like to take this time to ask for a review. Vicente Botet has mentioned he could be a review manager. So I would like to move forward in scheduling a formal review. I plan to boostify the library for the

On 9/28/2015 10:45 AM, paul Fultz wrote: formal

...

review. In addition, I would like to ask for an informal review at this time. If anyone has some feedback or questions, that would be greatly appreciated.

Would I be right in saying that Fit's library functionality is based on creating global variables which represent function objects ?

It does provide a lot of capabilities for this, but it is not limited to this.

If so this seems a negative for me as I try, like many other C++ programmers, to avoid using global variables in my modules.

It is considered bad to share global state in the program, which is why global mutable variables and shared pointers is not recommended. However, it is not considered bad practice to define gloabl consts in the program(in fact it is recommended over magic numbers), which is what these global function objects are. Furthermore, this is common practice in a lot of libraries, such as Boost.Spirit, Boost.Proto, Boost.Phoenix, Boost.Hana and ranges-v3.

OTOH if Fit can work with function objects which are member variables or local variables its usefulness would be much greater IMO.

Yes, the adaptors can work with any function object(local or global). It just requires a little more boilerplate to wrap it in a function for global consumption. For example, you can write the print function in the quick start guide as a function as well: template<class T> void print(const T& elem) { fit::conditional( fit::if_(std::is_fundamental<T>())([](auto& x) { std::cout << x << std::endl; }); [](const std::string& x) { std::cout << x << std::endl; }, [](auto& range) { for(const auto& x:range) print(x); } )(elem); }

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

On Monday, September 28, 2015 at 7:34:44 PM UTC-5, Vicente J. Botet Escriba wrote:

...

Hi All,

Boost.Fit is a header-only function utility library for C++11. It supports both gcc, clang and visual studio. More info can be found here:

https://github.com/pfultz2/Fit

I would like to take this time to ask for a review. Vicente Botet has mentioned he could be a review manager. So I would like to move forward in scheduling a formal review. I plan to boostify the library for the

Le 28/09/15 16:45, paul Fultz a écrit : formal

...

review. In addition, I would like to ask for an informal review at this time. If anyone has some feedback or questions, that would be greatly appreciated.

Thanks, Paul

Hi Paul,

thanks for pushing your library up to a Boost review ready state. Count with me as review manager. I will do a first inspection and share with you my first impressions.

Thanks.

Best, Vicente

P.S.Is Fit independent from Tick?

Yes, it is. Fit doesn't require Tick.

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

Le 03/10/15 22:28, Ron Garcia a écrit :

Hello Paul,

Thank you for your contribution to Boost. And thank you to Vicente for volunteering to be a review manager and providing early feedback. Regarding scheduling for review, please contact myself and/or John Phillip once the library is ready for review and we will go forward with scheduling it.

Best, Ron

On 2015-09-28, at 7:45 AM, paul Fultz wrote:

...

Hi All,

Boost.Fit is a header-only function utility library for C++11. It supports both gcc, clang and visual studio. More info can be found here:

https://github.com/pfultz2/Fit

I would like to take this time to ask for a review. Vicente Botet has mentioned he could be a review manager. So I would like to move forward in scheduling a formal review. I plan to boostify the library for the formal review. In addition, I would like to ask for an informal review at this time. If anyone has some feedback or questions, that would be greatly appreciated.

Hi Ron,

I have started to inspect the documentation which is BTW quite large as often on this kind of libraries, as there are a lot of little functions. I have opened some issues on the GitHub repository, and I suggest to those interested in the library to do the same. This will help us to know what needs to be done before going to a Boost review. For the time being you can add the library on the review schedule if not already done. I hope Paul F. will have free time to follow up the issues and make the library ready for review not too far. Once we will see that the most of the issues are removed and that we don't have too much news ones we will fix a date. It is really awesome how C++11 helps to write readable code following the functional programming paradigm. Best, Vicente P.S. Paul, I have not reached to tag the issues, e.g. as a question, a suggestion, a comment, ..... Are the user able to do that or it is only the right of the repository owner?

P.S. Paul, I have not reached to tag the issues, e.g. as a question, a suggestion, a comment, ..... Are the user able to do that or it is only the right of the repository owner?

Thats a good question. Let me find out.

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

On 10/5/2015 7:16 PM, Vicente J. Botet Escriba wrote:

Le 28/09/15 16:45, paul Fultz a écrit :

...

Hi All,

Boost.Fit is a header-only function utility library for C++11. It supports both gcc, clang and visual studio. More info can be found here:

https://github.com/pfultz2/Fit

I have two questions to this list as review manager of this library:

In his current state the library is in namespace fit the files are under directory fit and the macros are prefixed by FIT_. Do we need that Paul moves to the namespace boost::fit, directory boost/fit and prefix by BOOST_FIT_ the macros before the review?

It would be easier testing the library if the directory structure matched that of Boost source. I don't think the two other things are required for the review.

The documentation doesn't contain hypertext helping to navigate from the usage of a function to its reference documentation. This make more difficult the review even if there is a explorer for the table of contents. Do we need hypertext links before the review?

It would be nice but it shouldn't hold up a review.