[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [MiNT] XAAES slow?



tir, 21,.06.2005 kl. 18.04 -0400, skrev Standa Opichal:
> Ozk,
> 
> Odd Skancke wrote:
> 
> > What the hell is the problem with you all? I'm about to give up this shit!
> 
> With a little patience (I know I don't have one sometimes either) you 
> would not discourage the rest of the comunity left I belive.

 No, I would not do that. Sometimes I get frustrated.

> 
> You didn't even read his thoughts completely as it appears..... :(

 But I did. Because wind_calc() dont work on existing windows, it works
in the "moment" -> current theme. And his comments implied that
implementing runtime-changable themes is a bug. That pissed me off.

> 
> The rule is to keep explaining and think about what was wrong in the 
> previous attempt which wasn't understood. :)

 I dont know if it can be made any clearer looking at the previous
posts. But here is a scenario that hopefully clears things up;

 applications knows the coordinates of a dialog it wants to put into a
window. It calls wind_calc() to get the full-extent of a window with a
selected set of widgets. Then uses wind_calc()'s result to create/open
the window.

 Now the user changes theme, which changes the full/work area
relationships completely.

 application wants to open a second window with the exact same dialog,
and does exactly the same as for the first window. Everything will work
excellent!

 Now, if wind_calc() is used to get deltas for window #1 it will be
getting wrong coordinates, because that theme is no longer 'current'.
For window #2 it will be correct, since this window is using the still
'current' theme.

 If wind_get(WF_WORKXYWH/WF_CURRXYWH) is used to get the EXACT SAME INFO
as wind_calc() can provide, everything will work perfect on both
windows. Why? Because wind_get() works on a specific window.

 How to define wind_calc()'s context-lessness? I think the only way to
define that is to say that wind_calc() works on current values, current
theme. As soon as an instance of the 'moment' is snapped, by creation of
a window, all info about that snapshot should be fetched from that
snapshot.

 Implementing this is a bug because no other AES have this? Or is it a
bug to ignore the contextless nature of wind_calc()? 

> 
> Take care and keep up the good work.

 Thanks.

> 
> Standa