The code that follows should be rendering a sphere, but it doesn't for 
me in POV-Ray 3.7.1-alpha.8638272. It would be useful if others running 
the current master branch could confirm whether the same is seen - 
especially if seen in windows. Thanks.

Bill P.

//---------------------- Start SDL ----------------------------
// Function bug with POV-Ray 3.7.1-alpha.8638272 Ubuntu 14.04
global_settings { assumed_gamma 1 }
background { color rgb <0.5,0.5,0.5> }
camera { location <0,0,-12> look_at 0 sky y }
light_source { <100,300,-100> 1 }
#declare fn00 = function(x,y,z,r) { -exp(-(x*x+y*y+z*z)/r) }
#declare fn01 = function(x,y,z,r) { -exp(-(x*x+y*y+z*z)/r)+0.005 }
#declare fn02 = function(x,y,z,r) { -exp(-(x*x+y*y+z*z)/r)+0.00005 }
#declare fn03 = function(x,y,z,r) { -exp(-(x*x+y*y+z*z)/r)+0.000005 }
isosurface {
   function { fn00(x,y,z,1) }
   contained_by { box { -3,3 } }
   threshold -0.09
   accuracy 0.001
   max_gradient 2.0
   pigment { rgb <0.7,0.4,0.1> }
   finish { phong 1 }
}

// 3.7-stable and 3.6.1 render the sphere.
// Believe -0.00012340980408667956 is max value at container side.
// fn01=fn03 suggests values coming back too negative.
// Tried compile with optimization/fastmath off and no change.
//------------------- End SDL ----------------------------

Post a reply to this message

Am 16.06.2016 um 16:25 schrieb William F Pokorny:
> The code that follows should be rendering a sphere, but it doesn't for
> me in POV-Ray 3.7.1-alpha.8638272. It would be useful if others running
> the current master branch could confirm whether the same is seen -
> especially if seen in windows. Thanks.

I can confirm this issue for the Windows version as well. Some `#debug`
output of function results indicate that the function VM operates ok, so
we seem to be looking at an issue in the isosurface algorithm. (Also
note how fn01 through fn03 give a spherical shape, but broken shading.)

If anyone has sufficient time at hand to go through previous versions
and track down which version was the first to exhibit this problem, I'd
be much obliged.

Post a reply to this message

Am 16.06.2016 um 17:38 schrieb clipka:

> If anyone has sufficient time at hand to go through previous versions
> and track down which version was the first to exhibit this problem, I'd
> be much obliged.

Ah, never mind -- looks like I broke /something/ in commit a10629b
(2016-04-24, "Internal changes to user-defined function handling").

Post a reply to this message

Le 16/06/2016 17:38, clipka a écrit :
> Am 16.06.2016 um 16:25 schrieb William F Pokorny:
>> The code that follows should be rendering a sphere, but it doesn't for
>> me in POV-Ray 3.7.1-alpha.8638272. It would be useful if others running
>> the current master branch could confirm whether the same is seen -
>> especially if seen in windows. Thanks.
> 
> I can confirm this issue for the Windows version as well. Some `#debug`
> output of function results indicate that the function VM operates ok, so
> we seem to be looking at an issue in the isosurface algorithm. (Also
> note how fn01 through fn03 give a spherical shape, but broken shading.)
> 
> If anyone has sufficient time at hand to go through previous versions
> and track down which version was the first to exhibit this problem, I'd
> be much obliged.
> 

If it can help, the bug was not introduced yet in hgpovray-master, which should limit
the search between 9 April 2015 to now.

Do not expect more from me before the week-end or even next month, the work that pay
is exhausting me so far.

Post a reply to this message

Am 16.06.2016 um 18:10 schrieb clipka:
> Am 16.06.2016 um 17:38 schrieb clipka:
> 
>> If anyone has sufficient time at hand to go through previous versions
>> and track down which version was the first to exhibit this problem, I'd
>> be much obliged.
> 
> Ah, never mind -- looks like I broke /something/ in commit a10629b
> (2016-04-24, "Internal changes to user-defined function handling").

... and indeed I did. While overhauling the user-defined function
handling, I had eliminated a trivial function called `Vector_Function()`
in `isosurface.cpp`, which essentially just called the isosurface
function... /and/ subtracted the threshold from the result right away, a
fact that I had apparently missed.

Post a reply to this message

Am 16.06.2016 um 16:25 schrieb William F Pokorny:
> The code that follows should be rendering a sphere, but it doesn't for
> me in POV-Ray 3.7.1-alpha.8638272. It would be useful if others running
> the current master branch could confirm whether the same is seen -
> especially if seen in windows. Thanks.

Should be fixed with the brand-new master.

Post a reply to this message

On 06/16/2016 12:52 PM, clipka wrote:
> Am 16.06.2016 um 16:25 schrieb William F Pokorny:
>> The code that follows should be rendering a sphere, but it doesn't for
>> me in POV-Ray 3.7.1-alpha.8638272. It would be useful if others running
>> the current master branch could confirm whether the same is seen -
>> especially if seen in windows. Thanks.
>
> Should be fixed with the brand-new master.
>
Yes indeed, it looks good. Thanks much!

Bill P.

Post a reply to this message