v1.2.2

Changes:

• more tests and improved coverage
• internal and external API JSDocs cleanup
• extensive cleanup of (internal) vertex priority queue
• manual inlining of common GluHalfEdge.dst getter
• switched to official Closure Compiler gulp plugin
• dependency updates

Performance

• 3-11% performance improvement, depending on input geometry
• libtess.min.js 1090 bytes smaller (330 bytes smaller after gzip)

all commits in release

v1.2.1

Changes:

• cleanup of docs and gulpfile
• dependencies update
• fix jscs linting
• 2dc0ba5 - remove last vestiges of starting vertex cache (moderate performance increase and code-size reduction)

all commits in release

v1.2.0

Changes:

• libtess.assert statements removed from libtess.cat.js and libtess.DEBUG set to false to more accurately match an unminified libtess.min.js in terms of performance, JS engine behavior, etc.
• libtess.debug.js added to fill the need for a debug build. asserts still in place and libtess.DEBUGstill set to true
• correctness testing on real building footprint data, courtesy of @bcamper

all commits in release

v1.1.0

Changes:

• libtess will no longer emit triangle fans or strips, only triangles. Adding a GLU_TESS_EDGE_FLAG callback now only affects if that function is called back with edge information. Fully backwards compatible:

  • if you used GLU_TESS_EDGE_FLAG before, tessellation output is identical, whether or not you now use the flag
  • if you did not use GLU_TESS_EDGE_FLAG before, you will no longer receive rendering primitives other than GL_TRIANGLES. However, since you had to handle GL_TRIANGLES output anyways (e.g. for suboptimal strips), little change should be required. Also, it's not 1994 anymore, don't use triangle fans.

  See #9 for more details.

• testing, expectations viewer, etc