some comment-replies to Brahma :)

git-svn-id: svn://svn.code.sf.net/p/actiongame/code/trunk/ac@5259 e016b1c5-ed68-44aa-a2f7-31cac88cffde
This commit is contained in:
flowtron
2010-07-01 20:55:09 +00:00
parent 5d2d625d52
commit 13734e6961
+7 -4
View File
@@ -627,9 +627,9 @@ bool load_world(char *mname) // still supports all map formats that have
if(f) delete f;
c2skeepalive();
int cwx, cwy;
loopi(7) mapdims[i] = 0;
loopi(7) mapdims[i] = 0; // this is not needed Brahma - see code below - it's basically a good thing, but the code clearly will not fail w/o it.
loopk(4) mapdims[k] = k < 2 ? ssize : 0;
loopk(cubicsize) // FIXME // it is really necessary to loop again?
loopk(cubicsize) // FIXME // it is really necessary to loop again? // yes it, basically, is
{
switch(world[k].type)
{
@@ -650,9 +650,12 @@ bool load_world(char *mname) // still supports all map formats that have
}
}
loopk(2) mapdims[k+4] = mapdims[k+2] - mapdims[k];
/* the results of this function do not make sense:
/*
the results of this function do not make sense:
ac_desert: 070:070 - 185:185 : maxW 115 maxH 115 : Area 10216
ac_douze: 075:067 - 187:189 : maxW 112 maxH 122 : Area 9926 */
ac_douze: 075:067 - 187:189 : maxW 112 maxH 122 : Area 9926
what exactly is nonsensical about them? the area is all non-SOLID cubes.
*/
int *mapll;
mapll = new int[ssize]; // for testing
loopj(ssize)