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
This adds the ability for geometry sets to store build settings that can
be automatically applied when they are loaded. This should allow sharing
of .gset files to demonstrate problems with certain settings on certain
files. It also allows people to diagnose problems more easily by being
able to dump their own triangle meshes and settings and load them in the
demo, with all of its visualization options. .gset files can be created
from the current mesh and settings by pressing the 9 key, which will
generate it in the same folder as the input mesh.
Also converts more of the demo to use STL.
The readme was looking lonely after the BitDeli badge was taken away. Here are some nice badges to replace it, which also helpfully show how active a project this has become! They show recent average time to close issues and PRs.
There is a class and a struct called BuildContext in two different files,
one .h and one .cpp. Depending on how the user structures their program,
it's possible that this would cause a compile or link error, as seems to
have happened in #44. I've just renamed the struct to get around this.
Fixes#44
It appears that BitDeli is defunct; their website has had a banner about a change in the GitHub API for a long time, which is a dead link to a blog article, and none of the example stats have been updated for years. Let's just remove it!
When the detail mesh sampling sees an unset height it previously tried
to recover by looking for heights of neighbors. When the heightfield was
not eroded this recovery would fail since the contour simplification
could mean that the nearest valid cell was further away than a neighbor.
We now walk adjacent cells in a spiral to find a height and do this
depending on the max simplification error.
Fix#153
This changes the detail sampling to find height data by using the flood
fill method for polygons that were merged because of a border vertex
removal. This could potentially cause overlapping polygons to receive
the same region IDs which would later make the fast-path sampling in
getHeightData sample the wrong heights.
Also includes optimizations for this path and getHeightData itself.
Previously the seeding was responsible for 25% of the total time spent
in getHeightData in these cases. We now use the direction and prefer
moving directly towards the center of the polygon. This makes the
overhead of seedArrayWithPolyCenter virtually nonexistent (~2%
according to profiling).
Fix#146
dtNavMesh needs exclusive access to the data passed to addTile since it
will make dynamic changes to this at runtime. For that reason the data
should not be reused until the tile has been removed from the nav mesh
again.
Fix#149
* Removed some unused vars
* Fixed some signed/unsigned comparison mismatch warnings
* Cast some int's to size_t when appropriate
* Removed some redundant casts
* silenced warnings about implicit conversions that lose precision
* changed origMousePos to an array of ints to match mousePos
The demo relied on the build context returning accumulated times in
microseconds. On Windows this was using QueryPerformanceCounter directly
which has a higher resolution, thus returning wrong measured times. We
now report the accumulated times in microseconds on all platforms.
* Tests live in a "Tests" folder alongside the other components. Inside the "Tests" folder, tests are split into folders by the component they test. For example, the example unit test of "rcVdot" (which is implmemented in Recast/Recast.h) lives in "Tests/Recast/Tests_Recast.h".
* Uses the Catch testing framework
* One example test of "rcVdot"
* Tests are run on Travis and Appveyor after every build. Failing unit tests will fail the build in both case.
* Added instructions on running unit tests to the readme
* Renamed a bunch of variables in main.cpp to be more descriptive.
* Removed unnecessary SDLMain.h and SDLMain.m OSX objective-c class as well as the plist, strings, xib and icns files, which are not needed.
* included cstdio in imguiRenderGL since SDL2 doesn't do it for us.
* Updated premake5 script to support SDL2 and to set the debug directory, as well as copy the SDL2.dll to the target directory on Windows.
* Updated readme with more descriptive, platform-specific demo project setup instructions
* Updated appveyor build script to build vs2015 as well.
* Updated Travis build script to build SDL2 from source, because they use Ubuntu 12.04 which doesn't have the libsdl2-dev package in its repositories.
