Hi,
As I think has been discussed, valgrind has a few complaints about livarot. Most of these are fixed with the following patch:
Index: src/livarot/ShapeRaster.cpp =================================================================== RCS file: /cvsroot/inkscape/inkscape/src/livarot/ShapeRaster.cpp,v retrieving revision 1.7 diff -u -r1.7 ShapeRaster.cpp --- src/livarot/ShapeRaster.cpp 1 Apr 2004 00:16:12 -0000 1.7 +++ src/livarot/ShapeRaster.cpp 19 Apr 2004 00:22:24 -0000 @@ -1754,6 +1754,7 @@ }
swrData[no].calcX=swrData[no].curX+(to-step-swrData[no].curY)*swrData[no].dxdy; swrData[no].guess=-1; + swrData[no].misc = NULL; } void Shape::DestroyEdge(int /*no*/,float /*to*/,float /*step*/) {
Would anyone with valgrind like to verify this?
More importantly, could a livarot expert (fred?) take a look and see if this patch is sensible? The problems seem to stem from swrData[...].misc being uninitialised, but it could be that this value should be set somewhere else. In that case, my patch would probably just cover up a bug rather than fixing it.
Cheers
Carl