Yes, agreed. Makes sense. :-)

Post a reply to this message

Here an update:

During the rendering, my unofficial version crashed. I did a debugging
with the VS, and post here some of the errors (at the end, in German
language).

I changed both locations inside the blob.cpp file ("const DBL
DEPTH_TOLERANCE = 1.0e-6" and "const DBL INSIDE_TOLERANCE = 1.0e-12"). I
will reduce both values (3 and 9). If that doesn't help, I will just
keep the first change (3) as suggested by William.



-----------------------------

Ausnahmefehler bei 0x00000000778DFFC2 (ntdll.dll) in pvengine64.exe:
0xC0000374: A heap has been corrupted (Parameter: 0x0000000077957470)


---


Ausnahme ausgelöst bei 0x000000014032B230 in pvengine64.exe: 0xC0000005:
Zugriffsverletzung beim Lesen an Position 0xFFFFFFFFFFFFFFFF.

Falls ein Handler für diese Ausnahme vorhanden ist, kann das Programm
möglicherweise weiterhin sicher ausgeführt werden.


---


Ausnahmefehler bei 0x000000014032B230 in pvengine64.exe: 0xC0000005:
Zugriffsverletzung beim Lesen an Position 0xFFFFFFFFFFFFFFFF

Falls ein Handler für diese Ausnahme vorhanden ist, kann das Programm
möglicherweise weiterhin sicher ausgeführt werden.


---


Ausnahmefehler bei 0x000000014032B230 in pvengine64.exe: 0xC0000005:
Zugriffsverletzung beim Lesen an Position 0xFFFFFFFFFFFFFFFF

Post a reply to this message

During the new compiling with the reduced values, I am seeing this message:

1>  blob.cpp
1>..\..\source\core\shape\blob.cpp(2586): warning C4267: "=":
Konvertierung von "size_t" nach "int", Datenverlust möglich

= conversion from "size_t" to "int", data loss possible

Could that be the source of the crash? I am really no C++ programmer,
just wondering. Still, it is interesting to try creating my own
PVEngine64.exe... :-)

Post a reply to this message

Am 13.01.2016 um 05:23 schrieb Sven Littkowski:
> This time, the compilation succeeded ... but:
> 
> AVG reports a thread upon finishing:
> 
> Found Win32/DH{cQ?}
> Object Name:
> h:Setup\Applications\3D\POV-Ray\C++\povray-master\windows\vs10\bin64\pvengine64.exe
> 
> I am sure, this is just a false alarm, as I downloaded the ISO of the
> German VS straight from their website, and as my   computer is otherwise
> virus-free. Thus, I will tell AVG to ignore this "threat".

Given that you also experience problems when running the build, I'd
actually be more cautious. Proper procedure would be to submit the
binary to AVG for analysis.

Post a reply to this message

Am 13.01.2016 um 06:08 schrieb Sven Littkowski:

> During the rendering, my unofficial version crashed. I did a debugging
> with the VS, and post here some of the errors (at the end, in German
> language).

For debugging, you really want to pick "Debug" instead of "Release" from
the combo box up there. It should give you more information about the
problem, and may also take you directly to the offending code.

Also, you want to figure out whether the problem is caused by your
changes (which seems highly unlikely to me, though I wouldn't rule it
out for now), by something breaking your build process, or by something
unrelated to your particular build.

To figure out whether it is your changes that cause the problem should
be easy: Just revert back to the original code, rebuild, then run it in
the same way as before.

To figure out whether it is a problem with your build setup or a more
general issue, I'd recommend trying out the latest semi-official
release, routinely available at
https://github.com/c-lipka/povray/releases. As the latest build is only
available as a 32 bit version, you'll want to test those (possibly both
the 32-bit and 32-bit SSE2 version), but be aware that if it works fine,
there might still be a fundamental problem with the 64 bit version; in
that case you may also want to check the latest available semi-official
64 bit build, though again if that one works fine there still may be a
64 bit problem introduced later. Thus if all the semi-official builds
work fine, you may still want to ask for futher support, providing us
with the scene that causes the crash.


Also, you may want to do the usual stuff to narrow down the root cause:
Trim down your scene to the minimal set that still reproduces the crash.


FYI, a few days ago during development I did encounter a nasty crash
myself that I couldn't attribute to any specific thing I was messing
with at that time, so I wouldn't be too surprised if you did stumble
across an actual bug in the current master branch.

Post a reply to this message

AVG says, it is a potential thread. Not an actual virus. My system is
clean (was clean before). And I want to assume, that GitHub is also clean.

I am compiling that 64bit version now without own changes, and will make
some tests. And afterwards, if I still get crashes, I will compile those
other versions of POV-Ray as suggested by you (and test for viruses or
potential threads). :-)

Post a reply to this message

Sven Littkowski <jam### [at] yahoocom> wrote:

> I am compiling that 64bit version now without own changes, and will make
> some tests. And afterwards, if I still get crashes, I will compile those
> other versions of POV-Ray as suggested by you

As for those other versions, my suggestion was not for you to compile them, but
for you to test the pre-built binaries on your system and with the scene you're
using.

Post a reply to this message

Oh, okay. Yes, will do. Updates will follow.

Post a reply to this message