This should be fixed in upstream version 1.1.3; you can either pull the patches from SVN, upgrade to 1.1.3 entirely (as it contains only bug fixes), or recompile 1.1.2 to use floats instead of doubles.
This should be fixed in upstream version 1.1.3; you can either pull the patches from SVN, upgrade to 1.1.3 entirely (as it contains only bug fixes), or recompile 1.1.2 to use floats instead of doubles.