|

»

»

aqsis-development

Patch for error ID 2827244, large numbers of RiPoints incorrectly bound

View: New views

7 Messages — Rating Filter: Alert me

	Patch for error ID 2827244, large numbers of RiPoints incorrectly bound by Ricardo Mayor :: Rate this Message: Reply to Author \| View Threaded \| Show Only this Message This patch solves the error id 2827244. I've attached the patch an a rib file with an example. The problem was related with dicing of the ripoints primitive. When the number of points is greater than 256, a CqPoints instance is diceable. The CqPoints.split function spilts the instance into two diferents cqpoints but it doesn't set maximum size of each new instance. This patch fix the problem. `[my_changes.patch]` `379a380,382 > pA->InitialiseMaxWidth(); > pB->InitialiseMaxWidth(); >` ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Aqsis-development mailing list Aqsis-development@... https://lists.sourceforge.net/lists/listinfo/aqsis-development points_256_bug2827244.rib (2K) Download Attachment
	Re: Patch for error ID 2827244, large numbers of RiPoints incorrectly bound by Ricardo Mayor :: Rate this Message: Reply to Author \| View Threaded \| Show Only this Message The patch sended was senseless because it hasn't got any reference to the modified file . I've attached a new one. I hope this is the correct one. Best regards On Sun, Aug 16, 2009 at 9:54 PM, Ricardo Mayor <ricardo.mayor@...> wrote: This patch solves the error id 2827244. I've attached the patch an a rib file with an example. The problem was related with dicing of the ripoints primitive. When the number of points is greater than 256, a CqPoints instance is diceable. The CqPoints.split function spilts the instance into two diferents cqpoints but it doesn't set maximum size of each new instance. This patch fix the problem. `[my_changes.patch]` `--- ../notas_personales/error/points.cpp.old 2009-08-16 21:38:19.144692012 +0100 +++ libs/core/geometry/points.cpp 2009-08-16 21:39:02.084694408 +0100 @@ -377,6 +377,9 @@ pA->SetSurfaceParameters( this ); pB->SetSurfaceParameters( this ); + pA->InitialiseMaxWidth(); + pB->InitialiseMaxWidth(); + KDTree().Subdivide( pA->KDTree(), pB->KDTree() ); aSplits.push_back( pA );` ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Aqsis-development mailing list Aqsis-development@... https://lists.sourceforge.net/lists/listinfo/aqsis-development
	Re: Patch for error ID 2827244, large numbers of RiPoints incorrectly bound by Paul Gregory-4 :: Rate this Message: Reply to Author \| View Threaded \| Show Only this Message Ricardo, Thanks for the patch, we noticed that the original was incomplete, but managed to work out the intent. We were just reviewing the patch before committing. Thanks for sending an update, I'll respond to this mail when it is committed. Paul 2009/8/17 Ricardo Mayor <ricardo.mayor@...> The patch sended was senseless because it hasn't got any reference to the modified file . I've attached a new one. I hope this is the correct one. Best regards On Sun, Aug 16, 2009 at 9:54 PM, Ricardo Mayor <ricardo.mayor@...> wrote: This patch solves the error id 2827244. I've attached the patch an a rib file with an example. The problem was related with dicing of the ripoints primitive. When the number of points is greater than 256, a CqPoints instance is diceable. The CqPoints.split function spilts the instance into two diferents cqpoints but it doesn't set maximum size of each new instance. This patch fix the problem. ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Aqsis-development mailing list Aqsis-development@... https://lists.sourceforge.net/lists/listinfo/aqsis-development ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Aqsis-development mailing list Aqsis-development@... https://lists.sourceforge.net/lists/listinfo/aqsis-development
	Re: Patch for error ID 2827244, large numbers of RiPoints incorrectly bound by Chris Foster-5 :: Rate this Message: Reply to Author \| View Threaded \| Show Only this Message On Tue, Aug 18, 2009 at 3:39 AM, Ricardo Mayor<ricardo.mayor@...> wrote: > > The patch sended was senseless because it hasn't got any reference to the > modified file . I've attached a new one. I hope this is the correct one. Thanks! However when I apply this patch, I get a segfault when trying to render Matthias' nebula scene which I'm using for testing. I suspect the InitialiseMaxWidth() calls need to come after the KDTree().Subdivide() call. Another issue: Should really bother re-calculating the maximum width every time we subdivide a point set? Using the max-width of the full cloud for all the sub-clouds would give correct results but would potentially be inefficient if a point set had a very large range of widths. I expect that a given call to RiPoints() would usually contain points of much the same diameter. Thoughts? ~Chris. ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Aqsis-development mailing list Aqsis-development@... https://lists.sourceforge.net/lists/listinfo/aqsis-development
	Re: Patch for error ID 2827244, large numbers of RiPoints incorrectly bound by Ricardo Mayor :: Rate this Message: Reply to Author \| View Threaded \| Show Only this Message I tested the patch using the regression tests without problems. But I think it's better to use a more specific test. Where can I find the nebula scene? On Tue, Aug 18, 2009 at 6:15 AM, Chris Foster <chris42f@...> wrote: On Tue, Aug 18, 2009 at 3:39 AM, Ricardo Mayor<ricardo.mayor@...> wrote: > > The patch sended was senseless because it hasn't got any reference to the > modified file . I've attached a new one. I hope this is the correct one. Thanks! However when I apply this patch, I get a segfault when trying to render Matthias' nebula scene which I'm using for testing. I suspect the InitialiseMaxWidth() calls need to come after the KDTree().Subdivide() call. Another issue: Should really bother re-calculating the maximum width every time we subdivide a point set? Using the max-width of the full cloud for all the sub-clouds would give correct results but would potentially be inefficient if a point set had a very large range of widths. I expect that a given call to RiPoints() would usually contain points of much the same diameter. Thoughts? ~Chris. ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Aqsis-development mailing list Aqsis-development@... https://lists.sourceforge.net/lists/listinfo/aqsis-development ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Aqsis-development mailing list Aqsis-development@... https://lists.sourceforge.net/lists/listinfo/aqsis-development
	Re: Patch for error ID 2827244, large numbers of RiPoints incorrectly bound by Chris Foster-5 :: Rate this Message: Reply to Author \| View Threaded \| Show Only this Message On Tue, Aug 18, 2009 at 8:25 PM, Ricardo Mayor<ricardo.mayor@...> wrote: > I tested the patch using the regression tests without problems. But I think > it's better to use a more specific test. Where can I find the nebula scene? It's rather large I'm afraid (50Mb). For simplicity I've reproduced the segfault in a small test scene by using the "width" parameter to RiPoints rather than "constantwidth" (see the attached RIB). In any case, I know how to fix this... The main issue to decide on is whether we want to recalculate the max width every time we split the point cloud in two, or whether we just want to pass m_MaxWidth to the new child clouds without recalculating. I'm not sure what's best; either would work but relative efficiency would depend on the scene. ~Chris. ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Aqsis-development mailing list Aqsis-development@... https://lists.sourceforge.net/lists/listinfo/aqsis-development large_points.rib (4K) Download Attachment
	Re: Patch for error ID 2827244, large numbers of RiPoints incorrectly bound by Chris Foster-5 :: Rate this Message: Reply to Author \| View Threaded \| Show Only this Message On Tue, Aug 18, 2009 at 10:15 PM, Chris Foster<chris42f@...> wrote: > In any case, I know how to fix this... The main issue to decide on is > whether we want to recalculate the max width every time we split the > point cloud in two, or whether we just want to pass m_MaxWidth to the > new child clouds without recalculating. I'm not sure what's best; > either would work but relative efficiency would depend on the scene. Well, I've settled on duplicating m_MaxWidth for the child surfaces rather than recalculating it on each call to Split(). The changes are pushed into the trunk now. Cheers, ~Chris. ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Aqsis-development mailing list Aqsis-development@... https://lists.sourceforge.net/lists/listinfo/aqsis-development