Hi,

if I "render" (parse) the following scene, I get 5 warnings that in Line 10
(object{Test}) the pigment is missing. Well I can image why this may be
intended (of course one could texture the union as a whole) - but is this
really the way it should be? In older versions it was different, mentioning
the line, where the untextured object was defined, wasn't it?

#declare Test=
union{
sphere{0,1 pigment{rgb 1}}
sphere{0,2}
sphere{0,3}
sphere{0,1}
sphere{0,2}
sphere{0,3}
}
object{Test}

Marc-Hendrik

Post a reply to this message

"Marc-Hendrik Bremer" <Mar### [at] t-onlinede> wrote in message
news:3b9a7cd6@news.povray.org...
> if I "render" (parse) the following scene, I get 5 warnings that in Line
10
> (object{Test}) the pigment is missing. Well I can image why this may be
> intended (of course one could texture the union as a whole) - but is this
> really the way it should be? In older versions it was different,
mentioning
> the line, where the untextured object was defined, wasn't it?

The image will render on both 3.1g and 3.5b1 for me just fine (well, no
lights, of course).  The older versions handled it the same way (as a
warning rather than an error), except it did not return a line number like
3.5 now does.  I don't really see what the problem is.

-Ian

Post a reply to this message

Ian Burgmyer wrote:

> The image will render on both 3.1g and 3.5b1 for me just fine (well, no
> lights, of course).  The older versions handled it the same way (as a
> warning rather than an error), except it did not return a line number like
> 3.5 now does.  I don't really see what the problem is.

I issued a bugreport a long time ago about no line number given for no
pigment warnings in v3.1g. I presume this is the fix that Ron had
promised. He mentioned that it was not an easy thing to do. For example
if your no pigment warning comes from deep within a deeply nested loop
isolating the exact location and giving a line warning is extrememly
difficult to isolate. It may create more overhead than is acceptable
to get this level of error checking.

-- 
Ken Tyler - POV-Ray Technical Assistance Group

Post a reply to this message

Oh, sorry! I thought, that it reported the line number in which the object
was first declared. But as I see now there was no line number in 3.1 at all.
I stumbled over this, because an old scene of mine reported several missing
pigments in line with only on "object"-statement. Those line numbers don't
help that much, but they are of course better then the behaviour in 3.1.

So it's a feature, no bug :-)

Marc-Hendrik

Ian Burgmyer schrieb in Nachricht <3b9a7e98@news.povray.org>...

>The image will render on both 3.1g and 3.5b1 for me just fine (well, no
>lights, of course).  The older versions handled it the same way (as a
>warning rather than an error), except it did not return a line number like
>3.5 now does.  I don't really see what the problem is.
>
>-Ian
>
>

Post a reply to this message

"Ken" <tyl### [at] pacbellnet> wrote in message
news:3B9A8208.726A47FF@pacbell.net...
> I issued a bugreport a long time ago about no line number given for no
> pigment warnings in v3.1g. I presume this is the fix that Ron had
> promised. He mentioned that it was not an easy thing to do. For example
> if your no pigment warning comes from deep within a deeply nested loop
> isolating the exact location and giving a line warning is extrememly
> difficult to isolate. It may create more overhead than is acceptable
> to get this level of error checking.

Oh reeeeeeeealy?

I figured it'd be sort of easy to grab a line number from a script file, but
I dunno.  Can anyone give me a more technical explaination (Ron?)?  Pretty
please?

-Ian

Post a reply to this message

Ian Burgmyer wrote:

> Oh reeeeeeeealy?

Yes, really.
 
> I figured it'd be sort of easy to grab a line number from a script file, but
> I dunno.  Can anyone give me a more technical explaination (Ron?)?  Pretty
> please?

Ron Parker said and I quote "It's a little harder to fix than one might
think. The warning is printed in post_pigment, which is called when the
object finishes parsing,  potentially several lines after the texture is
parsed. Still, we should at least be able to get the line number of the
end of the relevant object declaration.".

-- 
Ken Tyler - POV-Ray Technical Assistance Group

Post a reply to this message

"Marc-Hendrik Bremer" <Mar### [at] t-onlinede> wrote in message
news:3b9a8140@news.povray.org...
> Oh, sorry! I thought, that it reported the line number in which the object
> was first declared. But as I see now there was no line number in 3.1 at
all.
> I stumbled over this, because an old scene of mine reported several
missing
> pigments in line with only on "object"-statement. Those line numbers don't
> help that much, but they are of course better then the behaviour in 3.1.

It is certainly more noticeable in 3.5 than with 3.1g :-)

-Ian

Post a reply to this message

"Ken" <tyl### [at] pacbellnet> wrote in message
news:3B9C26D1.62719E56@pacbell.net...
> > Oh reeeeeeeealy?
>
> Yes, really.

Reeeeeeeally?  *hrm...* Sorry.

> > I figured it'd be sort of easy to grab a line number from a script file,
but
> > I dunno.  Can anyone give me a more technical explaination (Ron?)?
Pretty
> > please?
>
> Ron Parker said and I quote "It's a little harder to fix than one might
> think. The warning is printed in post_pigment, which is called when the
> object finishes parsing,  potentially several lines after the texture is
> parsed. Still, we should at least be able to get the line number of the
> end of the relevant object declaration.".

Oh reeeeeeeeally?  *hrm...* Okay, I'll stop.

-Ian

Post a reply to this message