kicad-developers team mailing list archive

["Marco Serantoni" <marco.serantoni@...>]

--- In kicad-devel@xxxxxxxxxxxxxxx, Wayne Stambaugh <stambaughw@...> wrote:
Wayne, 
I've seen that, and i've worked on that, but all the other IDs were still suffering of the problem.

Another way was create functions that support also the wxDC passing but this approach implies deep changes in the code that i want to avoid, as i wantto avoid to inhibit yours great job of cleanup, so i will mantain the mostdefensive way possible.

My patch scope still the one to propose a "snipet" to be included in a greater job, support for MacOSX doesn't have to be an obstacle for the mainstream but a something stimulating a step ahead.

Anyway also if was, i kno that developing code implies being slaves of infinite loops for fixing code, so do your job in full freedom, don't apologize, in the case of i'll signalate an issue here and we'll find a better solution.

== another approach ==

/* Extends the old method */
Clear_PCB(bool,wxDC *){
[does the job]
};

/* Mantains compabitility */
Clear_PCB(bool){
wxClientDC dc( DrawPanel )
Zoom_Automatique(bool,wxDC *)
};

Bye,
Marco
>
> Macro,
> 
> Please take a look at my last commit. I made the load module from board
> command ID a separate event handler which should have fixed the nested
> DC issue in that code path. It appears that I may have fixed something
> other than what you did with your patch below. Although, I am not sure
> that the call stack dump that you posted on the dev list that I looked
> at was complete. A cleaner way to handle this is to create a separate
> event handler for the offending command ID in the switch statement
> rather than make an already overly complex function even more complex.
> An even better fix would be to make each command id its own event
> handler. I don'thave Mac so I have no way of testing my fix it other
> than making sure it still works under windows and linux. I apologize
> for forgetting to add that to the change log when I made the commit.
> 
> Wayne
> << snipped >>