【转帖】odgsbasevectorizeviewpolylineout coordinates
 

odgsbasevectorizeview::polylineout coordinates
odgsbasevectorizeview::polylineout coordinates
hi,
in what coordinates are coded the input points in the odgsbasevectorizeview:olylineout() function (actually overridable in a odgsbasevectorizeview class)? why they are dependent by the current view settings? is there a way to force it to input world coordinates only?
currently i crack this by calling
m_pmodeltoeyeproc->setworldtoeyetransform(odgematrix3d::kidentity) ;
in each myview::draw(const odgidrawable* pdrawable) call.
thanks in advance for any help.
regards
chudomir
hi chudomir,
you can control output coordinates via odgibasevectoriser::seteyetooutputtransform(). however in your case you just need to get identity world-to-eye-transform. it is set to identity in odgsbasevectorizeview's constructor and is controlled by odgsbasevectorizeview's::setview();
setview(odgepoint3d::korigin - odgevector3d::kzaxis, odgepoint3d::korigin, odgevector3d::kyaxis, fieldwidth, fieldheight, odgsview::kparallel) will set it to identity. fieldwidth, fieldheight affect projection that is not used in common case (on odgsbasevectorizeview level), odgsbasevectorizeview::getnumpixelsinunitsquare() that is used for deviation calculation.
m_pmodeltoeyeproc->setworldtoeyetransform(odgematrix3d::kidentity) ; is probably the shortest way and i don't see any problem you to use it.
last edited by dmitry a. novikov; 18th may 2004 at 06:27 amfff">.
>>currently i crack this by calling
>>m_pmodeltoeyeproc->setworldtoeyetransform(odgematrix3d::kidentity) ;
hi cudomir,
seems this is the better way:
// add eye-to-world transform to get geometry in wcs
seteyetooutputtransform(getworldtoeyetransform().i nverse());
// restore world-to-eye transform
seteyetooutputtransform(odgematrix3d::kidentity);
thank you for your reply,
but where should i call it and restore it?
i call m_pmodeltoeyeproc->setworldtoeyetransform(odgematrix3d::kidentity) ; in the draw() member function, and all seems to work?
regards
chudomir
it probably depends... i do the next to build solid cache on each view:
code:
void exgsopenglvectviewworldcache::update()
{
// temporary caching implementation:
odgematrix3d xw2e = getworldtoeyetransform();
::glmatrixmode( gl_modelview );
::glloadmatrixd((const gldouble*)&xw2e.transpose());
if(!m_list)
{
m_list = ::glgenlists(1);
invalidate();
}
if(!m_blistvalid)
{
::glnewlist(m_list, gl_compile_and_execute);
seteyetooutputtransform(xw2e.inverse());

exgsopenglvectorizeview::update();

// restore world-to-eye transform
seteyetooutputtransform(odgematrix3d::kidentity);

::glendlist();

// set to valid only if fully regenerated
m_blistvalid = !regenabort();
}
else
{
::glcalllist(m_list);
}
::glmatrixmode( gl_modelview );
::glloadidentity();
}
i see. well, in our system we use openinventor - it have a camera object and we have to fill the geometry in world coordinates. the camera takes care for the output.
thanks for your help anyway
regards
chudomir