[Zacam]

Good answers. I like them. Well, once the inverted boundary boxes issue get's resolved, I'll be more than happy to open and save every model I have with it.

[Kazan]

didn't bob catch the part of that which i missed?

[chief1983]

I don't know, but Bob's sig build from 1/21 has been seemingly working better than your RC2a build.  It seems to be crashing opening a lot of models that his is still handling.

[Kazan]

give me one of those models i've never been able to reproduce these crashes

[chief1983]

http://www.hard-light.net/forums/index.php/topic,51708.0.html

I think the basestar in Raptor's Original BSG pack was one that was crashing it.  It opens and saves fine in Bob's build, but I think it crashed yours, or was taking so long to open I assumed it had crashed.  On another note, the Protector and CylonAF models in that pack saved fine with Bob's build, but have some holes in the collision meshes, not sure if that's inverted BBox related or not.  Specifically, large chunks of the pods and I think the nose of the Protector, and the rear section of the CylonAF.  Probably just model errors though, but Bob's fixes had improved bad model handling so much that I thought he might be interested in seeing these two.

[Kazan]

New RC2b build posted

as far as i can tell the bound box problem should be resolved

oh.. you should recompile from cob to make sure an model suffering from that is completely fixed

[blowfish]

The new RC2b build seems to do weird things to collision detection.  I was working on a model today, and when I compiled it with RC2b, it had gaping holes in its collision detection, even after purging the BSP cache several times and converting to PMF and back again.  I tried it with the RC1 build and it worked fine on the first try.

[Kazan]

bob has been playing with BBox'es .. so talk to him

[Woolie Wool]

~58000 polys

is too high even once we've optimized the hell out of PCS2 - FS2 will kill you in your sleep

I've had no problems with multiple >10000 poly models on screen.

But then again, I have a GeForce 8800 GTS.

[blowfish]

Uhh ... 58000 is greater than 10000.  Do you mean 100000?

[Woolie Wool]

Uhh ... 58000 is greater than 10000.  Do you mean 100000?

Well, get a few of them together, some fighters, and a couple of installations, and boom, 60000 polies.

Edit: One of the ships was 25000, one was 21000, and one was 11000, and one was 9000. That's over 60,000 polies not counting the installations (>1000) and fighters (500-2000). Now just think of all the textures/texels being rendered with four caps, three installations, and four wings of fighters.

[ARSPR]

Just tested with RC2b and Comm2S-01.pof.

 

No way, PCS2 still "corrupts" the original pof which is "Boundary-error" free.

See PCS Mantis 49 for the full info.

[Kazan]

this is maddening.. was the file already suffering from inverted bboxes and you just opened the pof and resaved, or did you convert from cob over?

[edit]
i think i'm a ****ing moron, i think i have it figured out

[e2]
rc2c build posted

[ARSPR]

this is maddening.. was the file already suffering from inverted bboxes and you just opened the pof and resaved, or did you convert from cob over?

Kazan, I'm not a modeler at all so I only do VERY, VERY basic pof editing.    I don't even know what a Bounding Box is ...  

As posted in the mantis report:

• I open Comm2S-01.pof from DaBrain's MediaVPS 3.6.10. beta edition. (Model attached in PCS Mantis). This model doesn't make FS2 to complain about this "Inverted bounding box" issue.
• Then I just save it again. Then the saved model makes FS2 complain about it. And I also see that the Header >> bounding values have changed in PCS. BUT if I want to see these bounding box value changes, I have to reload the model, because just after saving, PCS2 still displays the original values.

(I know I can seem a bit lazy, but as I posted before, please test it yourself because with my little knowledge I really don't know what the hell I'm exactly doing. Maybe even the original Comm2S-01.pof is FUBAR because of some reason  )

EDITED
I've just seen your edits. I'm going to download and test rc2c

Stay tuned...

[Kazan]

(I know I can seem a bit lazy, but as I posted before, please test it yourself because with my little knowledge I really don't know what the hell I'm exactly doing. Maybe even the original Comm2S-01.pof is FUBAR because of some reason  )

i ask people who know how to replicate the bug to test it because this bug has been eluding me like a mother****er

[ARSPR]

Just tested with RC2c.

And at least the main model warning has disappeared. But there's still a lot of warnings about submodels. (Full info in PCS Mantis)

[Kazan]

Just tested with RC2c.

And at least the main model warning has disappeared. But there's still a lot of warnings about submodels. (Full info in PCS Mantis)

try again, this time purge the BSP cache

[ARSPR]

try again, this time purge the BSP cache

IT WORKS !!!!   

Now there are no inverted bounding box warnings. Just two comments (see Mantis report for all the info):

• Bounding box values are still different in original and saved models. Is this supposed to happen?
• The new bounding box values are not shown just after saving. You have to reopen the model in order to see them. Another bug?

 
And now a question. What's the BSP cache? Thanks.

[Kazan]

And now a question. What's the BSP cache? Thanks.

http://www.hard-light.net/wiki/index.php/POF_Constructor_Suite_2#FAQ

[chief1983]

That wiki page actually references an RC2d being the first one to properly check for BSP caching.  Are you holding out on us?