On 4/6/25 08:44, Thorsten wrote:
> If it is indeed a bug, you should be able to reproduce it if you enable 
> memory debugging by defining the various macros in pov_mem.cpp (in 
> particular MEM_GUARD) when compiling POV-Ray.
> 
> The other, and more likely, option - if this only occurs during long 
> renders - may very well be memory corruption due to overheating 
> problems. Nowadays this also tends to be a problem for memory due to 
> overall heat being generated.

Thank you for the information. I'd not considered overheating as a 
potential cause.

I spent some time earlier this morning looking at the C style memory 
handling I could find in the source code - as well as what in particular 
was using code from pov_mem.cpp.

Other than functions, which were common to both animations which 
crashed, I wasn't using much beyond the povms code(*). There are still 
too some allocations and frees to set up for possible blobs around the 
rendering of each frame, for example.

My dummy long animation renders have all finished clean. I'll try to get 
a MEM_GUARD build going today and run more of those.

Bill P.

(*) - I found that the messaging code doesn't use the pov_mem versions, 
but rather raw C malloc(), realloc() and free().

Post a reply to this message