MinersC wrote:
> I have a POV file, created in L3P, which rended fine in 3.6.1c but does not
> render correctly with 3.7.0.beta.16. 

Well, for starters, could you please be a bit more precise? Is there an
error message, a crash, something else?

	Thorsten

Post a reply to this message

I can report the same Problem.
I have a Scene what renders in Pov 3.5 and 3.6.x, but causes an access
violation during parsing in Pov 3.7 (tested in Beta 15 and 16, pvengine.exe
and pvengine-sse2.exe).

It is a Blob-Object on which I distribute lots of Mesh-objects with help of
the trace - command.

Here you can find the File:
http://blue.unix-ag.uni-hannover.de/~schwanck/Download/inc/Affe_1.27.zip

OS is Win XP Prof SP2 on Intel Core 2 Duo 6400 / 2 GB RAM

Post a reply to this message

Schwan379 wrote:
> I can report the same Problem.
> I have a Scene what renders in Pov 3.5 and 3.6.x, but causes an access
> violation during parsing in Pov 3.7 (tested in Beta 15 and 16, pvengine.exe
> and pvengine-sse2.exe).
> 
> It is a Blob-Object on which I distribute lots of Mesh-objects with help of
> the trace - command.
> 
> Here you can find the File:
> http://blue.unix-ag.uni-hannover.de/~schwanck/Download/inc/Affe_1.27.zip

Thanks for this, I have been able to replicate the problem here.

-- Chris

Post a reply to this message

While I know why it's crashing (it relates to an invalid element when copying
a CSG object), I don't yet know why it's getting that way since it takes so
long to get to the point where the crash happens (about 10 minutes of parsing
in debug mode).

It would help a lot if you could, if possible, simplify the scene such that
the crash happens sooner. In case it helps, the problem is triggered by the
line 'object{MA2_HaR}' at line 2912 of the file. However the real problem is
earlier than that; the object copy mentioned above just demonstrates it.

regards,

-- Chris

Post a reply to this message

Chris Cason <del### [at] deletethistoopovrayorg> wrote:
> While I know why it's crashing (it relates to an invalid element when copying
> a CSG object), I don't yet know why it's getting that way since it takes so
> long to get to the point where the crash happens (about 10 minutes of parsing
> in debug mode).
>
> It would help a lot if you could, if possible, simplify the scene such that
> the crash happens sooner. In case it helps, the problem is triggered by the
> line 'object{MA2_HaR}' at line 2912 of the file. However the real problem is
> earlier than that; the object copy mentioned above just demonstrates it.
>
> regards,
>
> -- Chris

Yes, I can help you there.
The simplified Scene would not crash anymore :(,
but if I include the Object in a whole Scene it would crash faster (after
0:40 min instead 2:12 min on my machine):

http://blue.unix-ag.uni-hannover.de/~schwanck/Download/Pov37BetaTest/MuB3.7Test.zip

While making this Zip-archive I found another access violation while
cancelling the parsing (pressing Alt+G and using the Stop-button result in
an error) or using the #error-Stream...
Just change Line 27 in "#local AffenFell = off;" and try again, I
constructed an #error statement what causes this error.

yours,

Thomas

Post a reply to this message

Schwan379 wrote:
> While making this Zip-archive I found another access violation while
> cancelling the parsing (pressing Alt+G and using the Stop-button result in
> an error) or using the #error-Stream...
> Just change Line 27 in "#local AffenFell = off;" and try again, I
> constructed an #error statement what causes this error.

I didn't get this error - is there a particular time that the button has to
be pressed at ?

It would help me if you could do whatever you need to do to make this happen
and save a mindump when that option is offered to you, then zip it up and
email it to min### [at] povrayorg.

-- Chris

Post a reply to this message

Chris Cason <del### [at] deletethistoopovrayorg> wrote:
> Schwan379 wrote:
> > While making this Zip-archive I found another access violation while
> > cancelling the parsing (pressing Alt+G and using the Stop-button result in
> > an error) or using the #error-Stream...
> > Just change Line 27 in "#local AffenFell = off;" and try again, I
> > constructed an #error statement what causes this error.
>
> I didn't get this error - is there a particular time that the button has to
> be pressed at ?
>
> It would help me if you could do whatever you need to do to make this happen
> and save a mindump when that option is offered to you, then zip it up and
> email it to min### [at] povrayorg.
>
> -- Chris

The Minidump is on its way to your Mailbox...

How I get the access violation:

Just run "Moosaffen_unter_Baum.pov" until Pov crashes.
On my Computer it crashes if I attempt to use the error-stream, too. That is
just that what I build into the File (for testing ;) ).

Just bevor ist crashes, there is some debug-text like
"Zeile: " a number and a random-looking string of ".", "F", "3" and
whitespace.

I hope, you can find it now.

-- Thomas

Post a reply to this message

Schwan379 wrote:
> I can report the same Problem.
> I have a Scene what renders in Pov 3.5 and 3.6.x, but causes an access
> violation during parsing in Pov 3.7 (tested in Beta 15 and 16, pvengine.exe
> and pvengine-sse2.exe).
> 
> It is a Blob-Object on which I distribute lots of Mesh-objects with help of
> the trace - command.
> 
> Here you can find the File:
> http://blue.unix-ag.uni-hannover.de/~schwanck/Download/inc/Affe_1.27.zip
> 
> OS is Win XP Prof SP2 on Intel Core 2 Duo 6400 / 2 GB RAM

This is fixed now, thanks.

-- Chris

Post a reply to this message

Schwan379 wrote:
> The Minidump is on its way to your Mailbox...

got it, thanks - could you also do one with the non-SSE2 build and send it?

thanks,

-- Chris

Post a reply to this message

Chris Cason <del### [at] deletethistoopovrayorg> wrote:
> Schwan379 wrote:
> > The Minidump is on its way to your Mailbox...
>
> got it, thanks - could you also do one with the non-SSE2 build and send it?
>
> thanks,
>
> -- Chris


Uups...

Yes, of course!

-- Thomas

Post a reply to this message