findPolysAroundCircle and findPolysAroundShape were only using distances
to compute the costs. This was inconsistent with other APIs that
performed cost-based searches and accepted filters as parameters.
This changes the Dijkstra searches in findPolysAroundCircle and
findPolysAroundShape to append the results when we visit nodes in the
main loop. The algorithm guarantees that we will never visit them again
after this, so at this point we are guaranteed that we cannot find a
better path.
Previously results would be appended the first time we saw a polygon
which meant later improvements on the paths to those polygons would not
be reflected in the results.
Fix#204
When creating the tile's links, offmesh connections inside the tile were
added twice, leading to the array of links being full and not being able
to store the links computed after (portals between tiles).
This changes dtTileCache::update to add an optional parameter that
indicates whether the tile cache is up to date. This is useful to
determine whether calling the function again would do anything.
These are by no means complete coverage of the functions they test, but are at least a start. For consistency, all floating point comparisons use Approx(), regardless of whether they need to or not.
Removed unnecessary C++11 unique_ptr stuff, now that rcHeightfield has a constructor and destructor.
These could lead to undefined behaviour if e.g. a negative value from ftell
was used to allocate memory.
Also store result of ftell in a long;
The result may previously have been truncated on some platforms
m_sample was being checked for null after being used, which looks like a
null dereference, though actually if m_sample becomes null, the tool will
be deactivated anyway. Regardless, this code looks more sensible.
These were reported by Coverity Scan. They're extremely unlikely but still
possible - if all edges of a poly are FLT_MAX or further away from the
input point, a negative index may be used to access the verts array of a
poly.
The fix is to arbitrarily pick the first edge as the closest in case all
are tied as being very far away.
Would crash if selecting a test with geometry then a test with missing geometry
Would use the previous sample type if selecting a sample that doesn't exist in the list
Two of the test cases had an incorrect sample name
dtNavMeshQuery now fails initialization if too many nodes are requested
in the node pool. This could cause wrong paths and infinite loops to
happen if the node indices started overflowing dtNodeIndex or
dtNode::pidx.
Fix#178
Fix three warnings added in fc5df2c, and increase the text pool size in
the demo. This would frequently be overflowed, even just by building
with the default settings and expanding the log.
This adds a version of queryPolygon that takes an interface which is
passed batches of polygons in the search area. This new API is then used
from the old queryPolygons to collect polygons, and from findNearestRef
to find the nearest poly in an arbitrarily sized area. Previously
findNearestPoly had an arbitrary poly limit of 128 which could cause it
to return potentially wrong results.
Fix#107
* Added virtual qualifier to a few destructors in subclasses.
* Removed a few empty destructors that didn't need to be explicitly defined.
* Fixed a few typos
* Removed unnecessary empty virtual destructors in some classes with parent's who derrived from a class with a virtual destructor
