[http://sourceforge.net/p/ufoai/bugs/1957 Item 1957] imported from sourceforge.net tracker on 2013-01-28 19:14:56

For a given map, the csv-files created by ufo2map and the debug_croute cmd comprise slightly different xy-ranges.
In order to compare the csvs for the excavation map with WinMerge, I had to adjust them a bit:

void Grid_DumpClientRoutes_f (void)
{
ipos3_t wpMins, wpMaxs;
VecToPos(map_min, wpMins);
VecToPos(map_max, wpMaxs);
wpMins[0] += 1;
wpMins[1] += 1;
wpMaxs[0] -= 3;
wpMaxs[1] -= 3;
RT_WriteCSVFiles(clMap, "ufoaiclient", wpMins, wpMaxs);
}

Of course, this is not a fix, just a workaround. But it illustrates that ufo and ufo2map have a different 'understanding' of the mins and maxs, which is potentially dangerous.

map_min and map_max are calculated in RT_GetMapSize() with a long series of boxtraces. Right before I started to look for a bug in that function, a question came to my mind:
We have the mins and maxs for every bsp (at the very beginning of the routing table) we add to a map, and we know the offset where we add it. With this data we could easily calculate the new map_mins and _maxs with a few additions.
Should be more exact and much faster.

Or did I miss something ?
===== Comments Ported from Sourceforge =====

====== wilminator (2009-01-01 18:07:24) ======

As the maps are loaded into ufo, the mins and maxs are buffered, increasing the size of the extents. This isn't bad per se- it is only used for optimizing functions from what I've seen, and if the functions scan where there is nothing, it does not hurt anything.
====== wilminator (2009-01-01 18:08:26) ======

If you want me to, I can look into not buffering the extents of the map, but it would be a low priority for now. Let me know.
====== aduke1 (2009-01-01 22:49:40) ======

As your recent fix seems to have blown away most - if not all - of the pathfinding problems with the exeption of stairs, let's hope we don't need the csv files anymore :)

But what do think of the suggested concept to simply calculate the new extents as
...
MapMaxs[i] = (TileMaxs[i] > MapMaxs[i]) ? TileMaxs[i] : MapMaxs[i];
...
instead of doing some 16,000 boxtraces in the void, however cheap they may be cpu-wise ?

====== wilminator (2009-05-18 03:30:57) ======

Closing because 1) it is an old ticket, and 2) the methods used to determine map bounds differ between ufo2map and ufo. I do agree that this should be looked into further, but imho is not critical.