The libart library | |||
---|---|---|---|

<<< Previous Page | Home | Up | Next Page >>> |

ArtVpath* art_vpath_from_svp (const ArtSVP *svp); ArtSVP* art_svp_from_vpath (ArtVpath *vpath); ArtSVP* art_svp_union (const ArtSVP *svp1, const ArtSVP *svp2); ArtSVP* art_svp_intersect (const ArtSVP *svp1, const ArtSVP *svp2); ArtSVP* art_svp_diff (const ArtSVP *svp1, const ArtSVP *svp2); |

SVPs cannot be created in many different ways: art_svp_from_vpath and art_vpath_from_svp can be used to generate SVPs from a Vpath and to generate VPaths from SVPs.

Once SVPs have been created, it is possible to manipulate them and apply a number of interesting operations on them. Most notably, it is possible to do the union of two SVPs with art_svp_union (Given two SVPs, the union of these two SVPs is an SVP representing the area obtained by the concatenation of the underlying areas of the two original SVPs). art_svp_intersect calculates the intersection of two SVPs: the resulting SVPs contain points which were in the two original SVPs. Finally, art_svp_diff calculates the difference of two SVPs: the resulting SVPs contains all the points which were in only one of the original SVP. These operations are sumamrized below:

ArtVpath* art_vpath_from_svp (const ArtSVP *svp); |

Converts the sorted vector path ` svp` into standard vpath form.

ArtSVP* art_svp_from_vpath (ArtVpath *vpath); |

Converts a vector path into sorted vector path form. The svp form is more efficient for rendering and other vector operations.

Basically, the implementation is to traverse the vector path, generating a new segment for each "run" of points in the vector path with monotonically increasing Y values. All the resulting values are then sorted.

Note: I'm not sure that the sorting rule is correct with respect to numerical stability issues.

:vpath | ArtVPath to convert. |

Returns : | Resulting sorted vector path. |

ArtSVP* art_svp_union (const ArtSVP *svp1, const ArtSVP *svp2); |

Computes the union of the two argument svp's. Given two svp's with winding numbers of 0 and 1 everywhere, the resulting winding number will be 1 where either (or both) of the argument svp's has a winding number 1, 0 otherwise. The result is newly allocated.

Currently, this routine has accuracy problems pending the implementation of the new intersector.

ArtSVP* art_svp_intersect (const ArtSVP *svp1, const ArtSVP *svp2); |

Computes the intersection of the two argument svp's. Given two svp's with winding numbers of 0 and 1 everywhere, the resulting winding number will be 1 where both of the argument svp's has a winding number 1, 0 otherwise. The result is newly allocated.

Currently, this routine has accuracy problems pending the implementation of the new intersector.

ArtSVP* art_svp_diff (const ArtSVP *svp1, const ArtSVP *svp2); |

Computes the symmetric of the two argument svp's. Given two svp's with winding numbers of 0 and 1 everywhere, the resulting winding number will be 1 where either, but not both, of the argument svp's has a winding number 1, 0 otherwise. The result is newly allocated.

Currently, this routine has accuracy problems pending the implementation of the new intersector.