#295 OpenGL error after renderbin draw

[Anonymous]

Originally created by: y...@sablonier.ch

*What steps will reproduce the problem?*
1. Loading aircraft where reflect shader is used (737-100, b29, IAR80)

*What is the expected output? What do you see instead?*
OpenGL error after renderbin draw

*What version of the product are you using (when using GIT version, please
mention date or Hudson build number)?*
latest from today

*What operating system and graphics card?*
OSG 2.9.9, OSX 10.6.6, ATI 5750

*Please provide any additional information below or as attachment (Avoid
expiring external links, such as to imageshack/pastebin/...).*
Tried also to revert patch from Lauri, without success.

[Anonymous]

Originally posted by: y...@sablonier.ch

I tried to revert two commits from Fred (patches from Lauri):
[re3ee29a59a69e67a85c64fde58fb6bdc3bac14f3]
[rf9bf35bc55be83615663c7130183cd66f9a366c3]

Unfortunately without success. I still get OpenGL error and invalid operation error after renderbin draw. All other shaders work as expected, I get the errors only with "reflect" and "bumpsec-reflect". "bumpspec" runs without problems.

Model-default is defined technique 10, working bumpspec technique 10, reflect/reflect-bumspec is defined technique 9. Maybe that doesn’t matter, but I have no idea where to start to look.

[Anonymous]

Originally posted by: y...@sablonier.ch

Not a ATI/OSX issue, it is also reported (see devellist) with:
Linux, 2.6.38, nVidia 8600 GT, 270.30 drivers, GIT from today, osg 2.9.10

[Anonymous]

Originally posted by: emili... (code.google.com)@gmail.com

Yup, happens over here when enabling for the first time the transition effects,

Warning: TangentSpaceGenerator: unknown primitive mode 9
Warning: TangentSpaceGenerator: unknown primitive mode 9
Warning: TangentSpaceGenerator: unknown primitive mode 9
Warning: detected OpenGL error 'invalid enumerant' at after RenderBin::draw(..)

On subsequent changes to it,  doesn't dispaly anything in the console anymore.

system specs as posted above :). Replied to the list first.

[Anonymous]

Originally posted by: y...@sablonier.ch

Just to add, I tried also to use OSG 2.9.7, without success. Reflect.vert/.frag still broken (reflect and bumpsec-reflect). While all other shaders pass OpenGL validation for my card without any problems, reflect and bumpspec-reflect doesn’t. Maybe driver related, I don’t know. Any other ideas?

[Anonymous]

Originally posted by: gasgur... (code.google.com)@gmail.com

..screenshots please, so we can compare, are you seeing duller colors when you enable material shaders and brighter colors when you disable them?  Etc.

[Anonymous]

Originally posted by: cumuluni... (code.google.com)@gmail.com

Is this still an issue?
Changing to Stalled for now.

Labels: Rendering
Status: Stalled

[Anonymous]

Originally posted by: grg.gdsr (code.google.com)

It is still an issue with latest version from git Wednesday 14-09-11. I am using Windows 7 and a Nvidia Palit GTS 450 card. Driver is: 8.17.12.8026

[Anonymous]

Originally posted by: grg.gdsr (code.google.com)

Here are the screenshots for shaders off and on.

[Anonymous]

Originally posted by: zakalawe@mac.com

Is the 'TangentSpaceGenerator: unknown primitive mode 9' the cause of the RenderBin errors, or not? I believe the TangentSpaceGenerator problems can be fixed by updating the loader code to convert polygons to quads/tris, based on comments from Tim in the past.

Cc: timoor...@gmail.com

[Anonymous]

Originally posted by: y...@sablonier.ch

I will try again ;-) Wrong comment some minutes ago. Sorry. This issue is still ONLY related to the REFLECT SHADER. No other shader throws this issue.

[Anonymous]

Originally posted by: zakalawe@mac.com

Discussing a variation on this (water shader, OS-X / Ati) on the mailing list at the moment. Issue looks to be lack of 'classic' texture units on Ati+Mac, and effects assuming many more.

[Anonymous]

Originally posted by: gijsrooy

(No comment was entered for this change.)

Summary: OpenGL error after renderbin draw

[Anonymous]

Originally posted by: gijsrooy

Is this now fixed as of https://gitorious.org/fg/fgdata/commit/28ecedf10f0f6dc3ffeedeac031f1ff314571d23 ?

Cc: zakalawe@mac.com
Status: Testing

[Anonymous]

Originally posted by: zakalawe@mac.com

Hopefully, but needs wider testing than just me.

[Anonymous]

Originally posted by: y...@sablonier.ch

I will test it too the next days. Many thanks looking to this issue.

[Anonymous]

Originally posted by: gijsrooy

Issue 608 has been merged into this issue.

[Anonymous]

Originally posted by: bre... (code.google.com)@gmail.com

Should be fixed. Report otherwise.

Status: Fixed

[Anonymous]

Originally posted by: l...@bitwisetech.com

Either 608 is no dulpicate or is not fixed
608: Is urban shader related, start in town, move urban shader quality to max then to min, silver pools appear in place of urban texture and console floods with OpenGL error messages.
Have asked for help on forum, on IRC, reported here over three weeks, have asked for help isolating problem, little response  Thks.

[Anonymous]

Originally posted by: chandram... (code.google.com)@gmail.com

Hi,

I am observing this error still in Windows Vista. I tried to download and install the latest version of FlightGear (2.4.0). I am not able to run the program. Attached is the screenshot depicting the error.

[Anonymous]

Originally posted by: chandram... (code.google.com)@gmail.com

(No comment was entered for this change.)

[Anonymous]

Originally posted by: gijsrooy

Could you try the 2.6.0 release candidate? You can download it from http://www.flightgear.org/news/flightgear-v2-6-release-candidates/

[Anonymous]

Originally posted by: b...@ucw.cz

Still seeing this with latest next (SimGear (098441f) and FlightGear (5ca0192) and master data (591eae6)). I only started seeing this after update sometime in January. I could bisect it, but at that update the setting of shaders changed and I had the "quality/performance" setting at 0 before that, so the shader wasn't enabled before, so I am not sure whether it's worth. I am also not sure whether to bisect SimGear or FlightGear; doing both is more complicated.

This is Debian/GNU Linux on amd64 with Radeon HD 6450 using radeon 1:6.14.3-2+exp1 and egl 8.0-2 driver (unlike the fglrx driver from AMD, this one does display runway lighting correctly and now I tried precipitation again and it now works too (wasn't earlier)).

Fog is now only applied if I turn of the "Generic" shader and even than it's not really correct, because trees seem to be unaffected. Some other shadered objects also become unaffected. Landmass at any non-0 setting, forrest background with transition on, urban at levels 1-3 (it's completely broken at 4 and 5) and water from 2 up.

Before the settings change while I used to have quality always at 0, I did have the transition and crop shaders checked and random vegetation enabled and never noticed problems with fog, so something probably changed beyond the setting. So could me trying to bisect something help? Or trying to get some logs/test something (I know C++ well, but never did anything with OpenGL)?

[Anonymous]

Originally posted by: b...@ucw.cz

Retrying with revision SimGear d7931a2, FlightGear 484d3c6, data 07be94a (dated 7th January), fog is correctly applied to all shaders. I am going to bisect and write it as new bug.

[Anonymous]

Originally posted by: b...@ucw.cz

Ok, I have initially though it would be the same problem, because it has the same symptom (fog not working), but I checked further and reported #699.