This has been driving me crazy all day!!! There's no other SOP with display flag on besides the one I'm working on right now and yet I can see geometry from at least one other geo container outside the current. It happens in either h11 and gl3.2 modes and the only way I've found to temporarily fix this is by closing my houdini session and starting a new one. Is anybody else having the same issue? Is there a better workaround than mine?
I'm using 12.5.428, on centOS 6.2, Nvidia diver version 295.20.

Cheers

If you can determine a set of steps to reproduce this problem, please submit a bug. This situation is very difficult to reliably recreate in order to fix. Details such as whether you use Show/Ghost/Hide Other Objects, Show Current Node or Display Node, and how you were editing the network around the time you first notice it are helpful.

You sure you don't have Edit -> Preferences -> Default Geometry Display set to “Show Current Operator” instead of “Show Display Operator”? That would result in geometry being shown that didn't have its Display Flag set.

@twod:

No, there's no specific set of operations, I'm just doing regular usage, diving in an out of geo containers, turning display flags here and there. Absolutely nothing out of the ordinary, otherwise if I had noticed a pattern I would have logged a bug.

@wolfwood:

It happens even with “Show Display Operator”, and it's displaying geo that's not even in the same container as the one I'm working with…

Cheers

There was a very small change made in 12.5.479 that might help the situation, and though it was a fix for a different issue, it might help.

rafaels
This has been driving me crazy all day!!! There's no other SOP with display flag on besides the one I'm working on right now and yet I can see geometry from at least one other geo container outside the current. It happens in either h11 and gl3.2 modes and the only way I've found to temporarily fix this is by closing my houdini session and starting a new one. Is anybody else having the same issue? Is there a better workaround than mine?
I'm using 12.5.428, on centOS 6.2, Nvidia diver version 295.20.

Cheers

I'm trying to test this, are you using imported geo?

is there any other nodes apart from sops ie dops, particles etc?

What type of nodes are in the scene? blast, poly extrude, shaders etc

thanks

Edit: I'm using 12.5.490, oS X, AMD 7950, doing mainly with fluids work and haven't seen this problem.

I have the same issue with 12.5.411 / CentOS 6.4 / Nvidia Quadro 2000 driver ver 310.40
sometimes geometry just freezes in viewport and it doesent matter which context is current it stays forever. Sometimes it dissapears when switching several contexts, sometimes I have to restart houdini. Unfortunately no way to recreate it. I`ll try latest daily build as soon as I can to see if this bug is still present.
P.s. The type of geo doesent matter. Volumes/Geos/Visualisation SOPs - i have this issue with all of them

same issue here. cant help in reproducing the problem since it happens randomly. Generally i think the 3d viewport needs improvement. Especially when it comes to Modelling. Its just frustrating sometimes

i have a suspect, the for each seems to be often involved, but can not say it for sure

Try switching to an adjacent pane tab so your viewport is no longer visible and switch back. Might save some restarting

sanostol
i have a suspect, the for each seems to be often involved, but can not say it for sure

More often for me too, like using Template display on another node.

Is everyone using 2 monitors when you get the error?

thanks

MartybNz
Is everyone using 2 monitors when you get the error?

thanks

3 for my part… and i have the same problems with 2 monitors.

fsimerey

MartybNz
Is everyone using 2 monitors when you get the error?

thanks

3 for my part… and i have the same problems with 2 monitors.

If you get the chance, can you test with 1 monitor. I've only starting getting some errors when using 2 monitors, no errors when using 1.

Cheers.

I can confirm that this can also happen on 1-monitor configurations.

I remember that I also had for-each sops involved, whith the scene viewer pinned to upper levels (outside the for-each)…

pinning will always hold the display flag. is this the so-called display ‘error’?

Edit: to recreate:
create sphere
Dive inside: pin viewer
go back to scene level and click the display flag. The sphere is ‘stuck’ in the viewer


see attachments

no this is not the display error, this is normal behavior when You pin a viewer, the display error shows geometry that should not be visible
i had the case that even after a new scene command the geometry was still visible in the viewport, no object in the scene.

MartybNz
pinning will always hold the display flag. is this the so-called display ‘error’?

Edit: to recreate:
create sphere
Dive inside: pin viewer
go back to scene level and click the display flag. The sphere is ‘stuck’ in the viewer


see attachments

Cool - there's a good example on Odforce to test if you can. I managed to replicate the error first time but not since then. The subnets used, VoroniFracture, have the ForEach nodes in them as well.

Last post, page 2. mantragora's post

http://forums.odforce.net/index.php?/topic/18202-h125-viewport-displays-geometry-that-its-not-supposed-to/page__st__12 [forums.odforce.net]

Thanks (and thanks to Mantragora). I will give it a spin.

Edit: As it turns out, the video starts after whatever steps were taken to end up in that situation. So unfortunately, while it's a good demonstration of the results of the bug, it's not very good to reproduce the bug. I did attempt to follow the steps he mentioned, but couldn't get it to produce the bug. There is a bug fixed in 12.5.496 (tomorrow's build) that corrects a couple of cases of missing geometry, though.

Cool - I was able to create the error using a SopSolver with a pointjitter inside. Working on making it consistent.