Well, it would require fluid to be able to analyse if some fluid modification would actually mean modification of the header file or not, so it would be quite significant change to the fluid.
But there is "workaround" which works for me: 1) if I have many independent GUI elements, I split it to independent fluid files (each window/widget definition in separate .fl file) and include only used headers 2) isolate GUI and non-gui parts of the program to independent files. Usually I need only one relatively short source to code around particular gui window/widget where the fluid-generated header is included 3) If I need some reference to a GUI object, I usually use it only through a pointer and put general definition like class Foo_Class ; without actual inclusion of Foo_Class.H header in other headers and api definitions 4) And ... (yeah, very boring, I am lazy too...) cleaning after adding substantial part of features (read:bloat) usually helps to reduce headers cross-inclusion and dependency. R. Oguz Akyuz wrote: > Hi Everybody, > > I've been using FLTK for some time now, and I'm generally very satisfied with > it. However, one nuisance that often disturbs me is when I "Write Code" using > fluid, it updates both header and sources files causing the subsequent > compilation take a painstakingly longer time. This is especially a problem if > the project is large and the updated header file causes many files to compile. > > So instead of blindly updating both header and source files, can "Write Code" > be a little more intelligent and update the header file only when necessary > (i.e. when the interface changes, etc.)? > > My apologies if this feature already exists and I'm ignorant about it. > > Thanks, > Oguz _______________________________________________ fltk mailing list [email protected] http://lists.easysw.com/mailman/listinfo/fltk
