On 17.03.2010, at 22:01, manolo gouy wrote:

>>>> We have now basically two names for each graphics function, e.g.,
>>>> fl_rect() and Fl_Device::rect()
>>>> and only one (generally, but not always, the second) is documented,
>>>> so this creates a warning for the undocumented one.
>>>> I believe the solution would be to hide from Doxygen the name
>>>> Fl_Device::<i>graphics_function</i>() because we don't want
>>>> end-users to use it, and instead document the
>>>> <i>fl_graphics_function</i>() name.
>>>> =20
>>>> Does that sound correct ?
>>> =20
>>> Yes, that sounds correct. But that would involve much moving of
>>> Doxygen docs to the right places. And maybe we should consider
>>> to use Matt's proposal \internal (or similar, don't remember
>>> exactly) to hide internals rather than using #if[n]def FL_DOXYGEN.
>> 
>> I suggest that we keep the documentation of the fl_xxx calls as they =
>> are. But Fl_Device and all its methods should be documented nevertheless =
>> (and references (\see) the corresponding fl_xxx function), because a =
>> developer may want to write a new Fl_Device for some exotic embedded =
>> display.
>> 
>> Marking the Fl_Device class members \internal may be a good idea.
>> 
>> - Matthias=
>> 
> 
> The trouble is that the documentation is in the .cxx files next
> to what used to be the fl_xxx() functions. Most (but not all)
> of these are now the Fl_Device::xxx() functions. Thus,
> most of fl_xxx() functions are now undocumented.

Oh, I see. We'd need to copy all docs to the header, right? Annoying :-(

_______________________________________________
fltk-dev mailing list
[email protected]
http://lists.easysw.com/mailman/listinfo/fltk-dev

Reply via email to