octave-maintainers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Release candidate for Qhull 2012.1


From: Rik
Subject: Re: Release candidate for Qhull 2012.1
Date: Sun, 29 Jan 2012 22:45:31 -0800

On 01/29/2012 07:42 PM, Brad Barber wrote:
At 08:50 PM 1/29/2012, Ben Abbott wrote:

>On Jan 29, 2012, at 8:46 PM, Dmitri A. Sergatskov wrote:
>
>> On Sun, Jan 29, 2012 at 7:38 PM, Ben Abbott <address@hidden> wrote:
>> 
>>> On Matlab ...
>>> 
>>> cube = [0 0 0;1 0 0;1 1 0;0 1 0;0 0 1;1 0 1;1 1 1;0 1 1];
>>> [h, v] = convhulln (cube);
>>> 
>>> h =
>>> 
>>>     3     4     2
>>>     2     4     1
>>>     6     2     1
>>>     5     6     1
>>>     4     8     1
>>>     8     5     1
>>>     7     3     2
>>>     6     7     2
>>>     7     4     3
>>>     7     8     4
>>>     7     6     5
>>>     8     7     5
>>> 
>>> 
>>> v =
>>> 
>>>     1
>>> 
>>> [h, v] = convhulln (cube, {'Qt'})
>>> 
>>> h =
>>> 
>>>     3     4     2
>>>     2     4     1
>>>     6     2     1
>>>     5     6     1
>>>     4     8     1
>>>     8     5     1
>>>     7     3     2
>>>     6     7     2
>>>     7     4     3
>>>     7     8     4
>>>     7     6     5
>>>     8     7     5
>>> 
>>> v =
>>> 
>>>     1
>>> 
>>> I get the same with Octave (tip below) on MacOSX 10.7.2 (Lion).
>>> 
>>> $ hg tip
>>> changeset:   14282:b4d7de953066
>>> tag:         tip
>>> user:        Michael Goffioul <address@hidden>
>>> date:        Sun Jan 29 12:53:05 2012 +0000
>>> summary:     Export more required symbols [Win32].
>>> 
>>> Does Fedora really return a 6x4 matrix for the first test?
>>> 
>>> Ben
>> 
>> octave:1> cube = [0 0 0;1 0 0;1 1 0;0 1 0;0 0 1;1 0 1;1 1 1;0 1 1];
>> octave:2> [h, v] = convhulln (cube);
>> octave:3> h
>> h =
>> 
>>   4   1   2   3
>>   5   6   2   1
>>   8   5   1   4
>>   6   7   3   2
>>   7   8   4   3
>>   8   7   6   5
>> 
>> Fedora 16 x86_64, octave 3.6.0, qhull 2003.1
>> 
>> Dmitri.
>
>Using qhull version < 2011 I got the same thing.
>
>Doesn't this result imply that the 6 faces of the cube are being returned and not the 12 triangulated faces ?
>
>Given the default option is "Qt" shouldn't the result be triangles?
>
>Ben
Note the version from Fedora.  It is using qhull 2003.1 which introduced option 'Qt'.  

My guess is that the Fedora build did not set 'Qt' as the default option list.  

All targets for Octave should use the same default option list.
Note the Octave version is 3.6.0 which passes 'Qt' as a default option.  It really appears that that something may have changed in 2012.  I finally built 2012 for Kubuntu 10.04 this evening and I now get the same test failure as Ben.  This is not present if I use version 2009 which is shipped with 10.04 by default.

Here are the results from running the same code linked against the 2009 and 2012 versions.  The only difference I see is "run-id".

convhulln (cube, 'Qt FO')
Options selected for Qhull 2009.1 2009/06/14:
  Qtriangulate  _pre-merge  _zero-centrum  _max-width  1
  Error-roundoff 1.4e-15  _one-merge 9.7e-15  _near-inside 4.9e-14
  Visible-distance 2.8e-15  U-coplanar-distance 2.8e-15  Width-outside 5.5e-15
  _wide-facet 1.7e-14
ans =

   4   1   2   3
   5   6   2   1
   8   5   1   4
   6   7   3   2
   7   8   4   3
   8   7   6   5

convhulln (cube, 'Qt FO')
Options selected for Qhull 2012.1 2012/01/26:
  run-id 1461090698  Qtriangulate  _pre-merge  _zero-centrum  _max-width  1
  Error-roundoff 1.4e-15  _one-merge 9.7e-15  _near-inside 4.9e-14
  Visible-distance 2.8e-15  U-coplanar-distance 2.8e-15  Width-outside 5.5e-15
  _wide-facet 1.7e-14
ans =

   3   4   2
   2   4   1
   6   2   1
   5   6   1
   4   8   1
   8   5   1
   7   3   2
   6   7   2
   7   4   3
   7   8   4
   7   6   5
   8   7   5

--Rik

reply via email to

[Prev in Thread] Current Thread [Next in Thread]