Skip to main content ↑ ↓
Topic: r1943 - Water level of ppm preview on new map window is one level high (Read 10136 times) previous topic - next topic

r1943 - Water level of ppm preview on new map window is one level high

simutrans r1943 GDI

I repoted this before and made a patch.
But this problem is NOT solved yet.
I don't know why you don't like my patches in spite of working well.

Code: [Select]
Index: gui/welt.cc
===================================================================
--- gui/welt.cc	(r1943)
+++ gui/welt.cc	(copy)
@@ -322,7 +322,7 @@
 					uint16 B = (uint8)fgetc(file);
 
 					while(karte_x<=rest_x  &&  karte_x<preview_size) {
-						karte[(karte_y*preview_size)+karte_x] = reliefkarte_t::calc_hoehe_farbe( (((R*2+G*3+B)/4 - 224+16)/16)*Z_TILE_STEP, sets->gib_grundw****er() );
+						karte[(karte_y*preview_size)+karte_x] = reliefkarte_t::calc_hoehe_farbe( (((R*2+G*3+B)/4 - 224+16)>>4)*Z_TILE_STEP, sets->gib_grundw****er() );
 						karte_x ++;
 					}
 					rest_x += skip_x;

Re: r1943 - Water level of ppm preview on new map window is one level high

Reply #1 –
This is a compiler error then. It works fine on my compiler. Furthermore for an unsigned number >> 4 and /16 are identical. Hmm. Ok, can change it, although I do not like hidden divisions in not speed critical sections.

It works even for the main routine ... maybe should be same in both places.

Re: r1943 - Water level of ppm preview on new map window is one level high

Reply #2 –
AFAIK, both vc++ and gcc are same at this point.
-5/2 = -2
-5 >> 1 = -3

And my first report is for v100.0 which you made it.
My code is the old simutrans code before you changed it. I dig it.

Re: r1943 - Water level of ppm preview on new map window is one level high

Reply #3 –
I apologize. Since the numbers are unsigned, I was ****uming that the compiler would make an unsigned shift. And even more, I was so used to /16 equal >>4 that I overlooked your point. (Strangley division by 16 is still used in the main code though).