On 15.03.2010, at 20:48, Matthias Melcher wrote:
>
> On 15.03.2010, at 20:02, manolo gouy wrote:
>
>>>
>>> Wild, wild hacking ... ;-)
>>>
>>> Manolo, it's fine that you could fix this "id" problem, but I
>>> strongly suggest that we should find another solution - even if
>>> this is only needed for "old" gcc versions. Such hacks tend to
>>> stay in the code forever(TM).
>>>
>>> Fiddling around with #defines looks like a really bad solution.
>>>
>>> I just did a quick rename of the id member variable in Fl_Pixmap
>>> and Fl_Bitmap as a proof of concept, but I didn't really test it
>>> other than compiling and running test/pixmap_browser (and it works).
>>>
>>> Thus, I propose to rename the id variable, although it is public,
>>> but marked as "for internal use".
>>>
>>
>> I fully agree it would be better to rename this id variable,
>> even if this conflict is not likely to happen often
>> because it occurs only for objective-c++ source files and only
>> with gcc-3.
>
> Yes, please lets rename it.

Okay, I'm just testing.  Commit to follow shortly

> I beleive the actual functionality is "stride".

Actually it's used as a pointer to an Fl_Offscreen or an Fl_Bitmask or
something very similar (implementation dependent).

What I did until now:

  - renamed "id" to "id_"
  - renamed "mask" to "mask_"

to distinguish these member variables from other variables (trailing
underscore) in Fl_Image (class Fl_RGB_Image), Fl_Bitmap, and
Fl_Pixmap (.h and .cxx, resp.).

I also made both members *private* w/o any negative effects until now.

Some more code cleanups by renaming local vars.

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

Reply via email to