On 1/22/2011 9:36 AM, Thomas de Groot wrote:
> Something which is becoming increasingly important is an undo switch,
> especially for the Groups/Lines section, but may be also for the
> Groups/Subdivision section.
>
> Thomas
>
>
>
New undo feature in the latest beta of PoseRay.

http://mysite.verizon.net/sfg0000/

later,

FlyerX

Post a reply to this message

Am 04.05.2011 09:44, schrieb FlyerX:
> New undo feature in the latest beta of PoseRay.
>
> http://mysite.verizon.net/sfg0000/
>

FWIW:
PoseRay is still crashing whith any contemporary ATI graphics card and 
model geometry above approx. 15 MB!
After getting frustrated about using Blender as converter (I really 
loved PoseRay for that purpose!) I did some research and used the 
VisualStudio debugging tools and finally found (at least a temporary) 
workaround:
The access violation happens within the atioglxx.dll module and it seems 
there is a function call with an uninitialized pointer. NVidia and *old* 
ATI drivers are obviously more forgiving about this issue but 
contemporary ATI drivers are not. As this is not a driver bug it is not 
to be expected that any future update of the ATI OpenGL graphics drivers 
will fix the problem.
But if it happens that you own a ATI card (and its installation CD) from 
2004 (anything newer will not work) you can do the following:

Search the installation CD for the file atioglxx.dl_   (<- note the 
underscore).

Copy this file into the PoseRay installation directory.

Run a console window and navigate to the PoseRay installation directory.
(e.g. "cd C:\CGI\PoseRay"). If you are using Vista or 7 and PoseRay is 
installed within the "Program Files" folder you will need also 
administrator rights.

Enter "expand atioglxx.dl_ atioglxx.dll".

PoseRay should work again. This will in no way mess up your graphics 
driver/OpenGL installation as *only* PoseRay is effected.

Note that any catalyst drivers from the ATI update side (going back to 
2008) are still too *new* and will not work.


Well, just in case you have no old driver cd available, you may want to 
send me a mail...

-Ive

Post a reply to this message

On 5/4/2011 6:18 AM, Ive wrote:
> Am 04.05.2011 09:44, schrieb FlyerX:
>> New undo feature in the latest beta of PoseRay.
>>
>> http://mysite.verizon.net/sfg0000/
>>
>
> FWIW:
> PoseRay is still crashing whith any contemporary ATI graphics card and
> model geometry above approx. 15 MB!
> After getting frustrated about using Blender as converter (I really
> loved PoseRay for that purpose!) I did some research and used the
> VisualStudio debugging tools and finally found (at least a temporary)
> workaround:
> The access violation happens within the atioglxx.dll module and it seems
> there is a function call with an uninitialized pointer. NVidia and *old*
> ATI drivers are obviously more forgiving about this issue but
> contemporary ATI drivers are not. As this is not a driver bug it is not
> to be expected that any future update of the ATI OpenGL graphics drivers
> will fix the problem.
> But if it happens that you own a ATI card (and its installation CD) from
> 2004 (anything newer will not work) you can do the following:
>
> Search the installation CD for the file atioglxx.dl_ (<- note the
> underscore).
>
> Copy this file into the PoseRay installation directory.
>
> Run a console window and navigate to the PoseRay installation directory.
> (e.g. "cd C:\CGI\PoseRay"). If you are using Vista or 7 and PoseRay is
> installed within the "Program Files" folder you will need also
> administrator rights.
>
> Enter "expand atioglxx.dl_ atioglxx.dll".
>
> PoseRay should work again. This will in no way mess up your graphics
> driver/OpenGL installation as *only* PoseRay is effected.
>
> Note that any catalyst drivers from the ATI update side (going back to
> 2008) are still too *new* and will not work.
>
>
> Well, just in case you have no old driver cd available, you may want to
> send me a mail...
>
> -Ive
>
>
>

Thanks for the workaround. To this day I have not been able to find the 
origin of the bug. I used an OpenGL debugger and all the calls were 
correct. Most likely it is a bug (or maybe an outdated function) in 
PoseRay but without the ATI drivers running there is no way to check 
exactly what is happening. If there was a way of emulating an ATI card 
as a layer over the Nvidia drivers, for programming/debugging purposes, 
that would be great.


FlyerX

Post a reply to this message

Am 04.05.2011 20:20, schrieb FlyerX:

> Thanks for the workaround. To this day I have not been able to find the
> origin of the bug. I used an OpenGL debugger and all the calls were
> correct. Most likely it is a bug (or maybe an outdated function) in
> PoseRay but without the ATI drivers running there is no way to check
> exactly what is happening. If there was a way of emulating an ATI card
> as a layer over the Nvidia drivers, for programming/debugging purposes,
> that would be great.
>

Yeah, I understand your problem very well. While stepping through 
hundrets of lines of assembler code I did suspect the calls
glTexImage2D and/or glGetTexImage to be the critical parts but it might 
very well be the case that some initialization before did already fail. 
Sorry, this is all I can tell as I'm just an OpenGL n00b ;)

While I did play around with the debugger (and accidentally did corrupt 
the ATI drivers) I did also find out that in this case the 
Windows-OpenGL-layer falls back to a GDI software emulation mode that 
works (at least with an i7) not as bad as it sounds. Maybe a button 
within PoseRay to force software emulation (if this is possible) would 
be a cleaner workaround than moving dll's from the stone-age around?

-Ive

Post a reply to this message

On 5/4/2011 3:19 PM, Ive wrote:
> Am 04.05.2011 20:20, schrieb FlyerX:
>
>> Thanks for the workaround. To this day I have not been able to find the
>> origin of the bug. I used an OpenGL debugger and all the calls were
>> correct. Most likely it is a bug (or maybe an outdated function) in
>> PoseRay but without the ATI drivers running there is no way to check
>> exactly what is happening. If there was a way of emulating an ATI card
>> as a layer over the Nvidia drivers, for programming/debugging purposes,
>> that would be great.
>>
>
> Yeah, I understand your problem very well. While stepping through
> hundrets of lines of assembler code I did suspect the calls
> glTexImage2D and/or glGetTexImage to be the critical parts but it might
> very well be the case that some initialization before did already fail.
> Sorry, this is all I can tell as I'm just an OpenGL n00b ;)
>
> While I did play around with the debugger (and accidentally did corrupt
> the ATI drivers) I did also find out that in this case the
> Windows-OpenGL-layer falls back to a GDI software emulation mode that
> works (at least with an i7) not as bad as it sounds. Maybe a button
> within PoseRay to force software emulation (if this is possible) would
> be a cleaner workaround than moving dll's from the stone-age around?
>
> -Ive
>
>
>
>

New beta available.

http://mysite.verizon.net/sfg0000/

The software mode idea was a good one and I added a mode to PoseRay to 
run in software mode. It is very slow but seems to work. It defaults to 
OpenGL 1.1.


FlyerX

Post a reply to this message