[GRASS-user] v.delaunay Bug in Mac OS X

Michael_Perdue · June 16, 2009, 10:50pm

Hi,

v.delaunay is generating incorrect results on Mac OS X on my machine (Intel Mac Pro, Leopard, W. Kyngesburye's GRASS 6.4-RC5 binaries). The following commands give the results below;

echo "456490.2375|5905949.8625|1
456690.09166667|5905895.5625|2
456570.93333333|5905821.65416667|3
456579.22916667|5905881.9875|4
456494.00833333|5905859.3625|5
456673.5|5905819.39166667|6
456601.85416667|5905782.4375|7
456604.11666667|5905944.58333333|8" | v.in.ascii -n output=DelaunayPoints cat=3

v.delaunay input=DelaunayPoints output=DelaunayTriangles
  Building topology for vector map <DelaunayTriangles>...
  Registering primitives...
  12 primitives registered
  24 vertices registered
  Building areas...
  100%
  4 areas built
  1 isles built
  Number of nodes: 8
  Number of primitives: 12
  Number of points: 0
  Number of lines: 0
  Number of boundaries: 12
  Number of centroids: 0
  Number of areas: 4
  Number of isles: 1
  Number of incorrect boundaries: 4
  Number of areas without centroid: 4
  WARNING: Unable to calculate area centroid z coordinate
    100%
  Building topology for vector map <DelaunayTriangles>...
  Attaching islands...
    100%
  Attaching centroids...
    100%
  Number of nodes: 11
  Number of primitives: 15
  Number of points: 0
  Number of lines: 0
  Number of boundaries: 12
  Number of centroids: 3
  Number of areas: 4
  Number of isles: 1
  Number of incorrect boundaries: 4
  Number of areas without centroid: 1

g.region vect=DelaunayPoints

d.mon x0
d.vect DelaunayTriangles
d.vect map=DelaunayPoints color=red width=0 icon=basic/circle size=10

(attachments)

v.delaunay problem.tiff (31 KB)

William_Kyngesburye · June 16, 2009, 11:57pm

I don't see any changes in v.delaunay itself from RC4, though maybe there are changes in the vector libraries. Did this work in RC4?

A big difference in RC5 for OSX is that it's now 64bit. Does this work:

arch -i386 v.delaunay in=DelaunayPoints out=DelaunayTriangles

On Jun 16, 2009, at 5:50 PM, Michael Perdue wrote:

Hi,

v.delaunay is generating incorrect results on Mac OS X on my machine (Intel Mac Pro, Leopard, W. Kyngesburye's GRASS 6.4-RC5 binaries). The following commands give the results below;

echo "456490.2375|5905949.8625|1
456690.09166667|5905895.5625|2
456570.93333333|5905821.65416667|3
456579.22916667|5905881.9875|4
456494.00833333|5905859.3625|5
456673.5|5905819.39166667|6
456601.85416667|5905782.4375|7
456604.11666667|5905944.58333333|8" | v.in.ascii -n output=DelaunayPoints cat=3

v.delaunay input=DelaunayPoints output=DelaunayTriangles
  Building topology for vector map <DelaunayTriangles>...
  Registering primitives...
  12 primitives registered
  24 vertices registered
  Building areas...
  100%
  4 areas built
  1 isles built
  Number of nodes: 8
  Number of primitives: 12
  Number of points: 0
  Number of lines: 0
  Number of boundaries: 12
  Number of centroids: 0
  Number of areas: 4
  Number of isles: 1
  Number of incorrect boundaries: 4
  Number of areas without centroid: 4
  WARNING: Unable to calculate area centroid z coordinate
  100%
  Building topology for vector map <DelaunayTriangles>...
  Attaching islands...
  100%
  Attaching centroids...
  100%
  Number of nodes: 11
  Number of primitives: 15
  Number of points: 0
  Number of lines: 0
  Number of boundaries: 12
  Number of centroids: 3
  Number of areas: 4
  Number of isles: 1
  Number of incorrect boundaries: 4
  Number of areas without centroid: 1

g.region vect=DelaunayPoints

d.mon x0
d.vect DelaunayTriangles
d.vect map=DelaunayPoints color=red width=0 icon=basic/circle size=10

<v.delaunay problem.tiff>

I've tested it on the same machine running Linux and had no issue. It also ran without issue on Mac OS X with GRASS 6.3. Can anyone verify that it isn't working properly anymore or could it be a problem with my installation?

Cheers,

Mike_______________________________________________
grass-user mailing list
grass-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-user

-----
William Kyngesburye <kyngchaos*at*kyngchaos*dot*com>
http://www.kyngchaos.com/

"Those people who most want to rule people are, ipso-facto, those least suited to do it."

- A rule of the universe, from the HitchHiker's Guide to the Galaxy

Michael_Perdue · June 17, 2009, 3:15am

This is the first time I noticed the problem, but I haven't used v.delaunay in a while. I tried the command modification you suggested, but with the same outcome. I also went back and recompiled a svn checkout from about 1.5 months ago as a 32 bit app, but v.delaunay behaved identically. I'm not sure whether v.hull and v.voronoi share much of the same code or not, but they both run fine.

On 16-Jun-09, at 5:57 PM, William Kyngesburye wrote:

I don't see any changes in v.delaunay itself from RC4, though maybe there are changes in the vector libraries. Did this work in RC4?

A big difference in RC5 for OSX is that it's now 64bit. Does this work:

arch -i386 v.delaunay in=DelaunayPoints out=DelaunayTriangles

On Jun 16, 2009, at 5:50 PM, Michael Perdue wrote:

Hi,

v.delaunay is generating incorrect results on Mac OS X on my machine (Intel Mac Pro, Leopard, W. Kyngesburye's GRASS 6.4-RC5 binaries). The following commands give the results below;

echo "456490.2375|5905949.8625|1
456690.09166667|5905895.5625|2
456570.93333333|5905821.65416667|3
456579.22916667|5905881.9875|4
456494.00833333|5905859.3625|5
456673.5|5905819.39166667|6
456601.85416667|5905782.4375|7
456604.11666667|5905944.58333333|8" | v.in.ascii -n output=DelaunayPoints cat=3

v.delaunay input=DelaunayPoints output=DelaunayTriangles
  Building topology for vector map <DelaunayTriangles>...
  Registering primitives...
  12 primitives registered
  24 vertices registered
  Building areas...
  100%
  4 areas built
  1 isles built
  Number of nodes: 8
  Number of primitives: 12
  Number of points: 0
  Number of lines: 0
  Number of boundaries: 12
  Number of centroids: 0
  Number of areas: 4
  Number of isles: 1
  Number of incorrect boundaries: 4
  Number of areas without centroid: 4
  WARNING: Unable to calculate area centroid z coordinate
  100%
  Building topology for vector map <DelaunayTriangles>...
  Attaching islands...
  100%
  Attaching centroids...
  100%
  Number of nodes: 11
  Number of primitives: 15
  Number of points: 0
  Number of lines: 0
  Number of boundaries: 12
  Number of centroids: 3
  Number of areas: 4
  Number of isles: 1
  Number of incorrect boundaries: 4
  Number of areas without centroid: 1

g.region vect=DelaunayPoints

d.mon x0
d.vect DelaunayTriangles
d.vect map=DelaunayPoints color=red width=0 icon=basic/circle size=10

<v.delaunay problem.tiff>

I've tested it on the same machine running Linux and had no issue. It also ran without issue on Mac OS X with GRASS 6.3. Can anyone verify that it isn't working properly anymore or could it be a problem with my installation?

Cheers,

Mike_______________________________________________
grass-user mailing list
grass-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-user

-----
William Kyngesburye <kyngchaos*at*kyngchaos*dot*com>
http://www.kyngchaos.com/

"Those people who most want to rule people are, ipso-facto, those least suited to do it."

- A rule of the universe, from the HitchHiker's Guide to the Galaxy

Paul_Kelly · June 26, 2009, 2:21pm

Hello all,
Last week I asked the author of the new v.delaunay if he could have a look at this, but I haven't heard back from him yet. FWIW I can't reproduce the problem with today's 6.4 release branch on a 32-bit Linux system (Slackware 12.2).

On Tue, 16 Jun 2009, William Kyngesburye wrote:

I don't see any changes in v.delaunay itself from RC4, though maybe there are changes in the vector libraries. Did this work in RC4?

A big difference in RC5 for OSX is that it's now 64bit.

Does this affect all builds on OS X? If so, it seems like a rather drastic change when the 6.4 release branch is in critical bugfix only mode preparing for release - I fear a change to 64-bit binaries could trigger obscure bugs that might not be found for a long time. Not something to do just before the release IMHO.

Paul

William_Kyngesburye · June 26, 2009, 2:40pm

On Jun 26, 2009, at 9:21 AM, Paul Kelly wrote:

Hello all,
Last week I asked the author of the new v.delaunay if he could have a look at this, but I haven't heard back from him yet. FWIW I can't reproduce the problem with today's 6.4 release branch on a 32-bit Linux system (Slackware 12.2).

On Tue, 16 Jun 2009, William Kyngesburye wrote:

I don't see any changes in v.delaunay itself from RC4, though maybe there are changes in the vector libraries. Did this work in RC4?

A big difference in RC5 for OSX is that it's now 64bit.

Does this affect all builds on OS X? If so, it seems like a rather drastic change when the 6.4 release branch is in critical bugfix only mode preparing for release - I fear a change to 64-bit binaries could trigger obscure bugs that might not be found for a long time. Not something to do just before the release IMHO.

Michael did try it on 32bit OSX with the same results. And also said it happened on a previous 6.4RC that was 32bit only.

As far as the 'drastic' change, this was a bugfix - fixed for trunk and dev6

http://trac.osgeo.org/grass/ticket/431

and backported to release:

http://trac.osgeo.org/grass/changeset/36838

I didn't see it attached to a ticket, though I did notice the backport in:

http://trac.osgeo.org/grass/ticket/597

Paul
_______________________________________________
grass-user mailing list
grass-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-user

-----
William Kyngesburye <kyngchaos*at*kyngchaos*dot*com>
http://www.kyngchaos.com/

"This is a question about the past, is it? ... How can I tell that the past isn't a fiction designed to account for the discrepancy between my immediate physical sensations and my state of mind?"

- The Ruler of the Universe