TGEA 1.8.1 Progress Update
by Alex Scarborough · 01/29/2009 (12:01 pm) · 39 comments
Hello all! It's time for a TGEA 1.8.1 progress update. At a high level, we're progressing quite well towards our goals of a faster and more polished product. With over 40 issues resolved, TGEA 1.8.1 is already as fast or faster than TGEA 1.7.1.
Below is a list of issues which are already resolved and guaranteed to be fixed by 1.8.1
Here is what remains on our todo list at this point
There were a couple of things we weren't able to reproduce in our own testing. If you have any additional information about them, it would be greatly appreciated
As before, keep submitting your bug reports. The vast majority of the issues above were community reported and some were even community fixed, which is just awesome. Every fix we get into 1.8.1 makes it a better product, so go find some more!
Below is a list of issues which are already resolved and guaranteed to be fixed by 1.8.1
- Fullscreen support on OS X
- Fix lockup with OS X and X1600
- Legacy terrain performance improvements
- Fix for incorrect terrain texture paths
- Fix bad sun color on legacy terrain in OS X
- Fix audio bug in Projectile
- Fix lighting preferences
- Fix for ShapeBase::getCameraTransform
- Fix so changing datablock in mission editor actually changes the object datablock
- Remove need for DDS mipmaps
- Render bitmap gui controls with linear texture filtering
- Fix FontRenderBatcher::render assert
- Fix scrolling with touchpad on OS X
- Fix for WMIVideoInfo init
- Fix LightManager initialization
- Fix fullscreen in Vista
- Disable fences by default
- Implement/Fix waterFind function in ShapeBase
- Fix mismatched method signatures in GameBase
- Fix for Splash emitters and explosions
- Color Picker support on OS X
- Mac: Fix sound volume issues post mission start
- Fix issue where cursor is offset after drag operations
- Fix billboard meshes
- Reimplement/fix BillboardZ meshes
- Mac: Key events processed by action map while GuiTextEditCtrl is active
- Mac: Modifier keys incorrect
- Fix getScheduleDuration, getTimeSinceStart
- DSQ import crashes/has incorrect orientation
- World editor defaults overwriting world editor prefs
- Win32: Cursor offset after resizing window
- Fix IFLs
- Pink lines in gui
- MegaTerrain match up problem
- Fix crash if DDS fails to load
- Fix memory leak in legacy terrain
- Clipmap not updated to reflect changing/unsetting of previously set terrain texture slots
- Fix inverted shift-snapping logic in GuiSliderControl
- Fix bitmap operator in terrain terraformer
- Terrain center in newMission moved to 0, 0
- Fix crash caused by dangling pointer held by world editor
- Win32: Attempting to use OpenGL device when using prefs from mac OS
- Terrain terraformer locked to first terrain block turned up by server query
- Terrain edge files getting out of sync
Here is what remains on our todo list at this point
- Fix terrain heightfield loading
- Update OpenAL dll
- Dedicated server should not open window
There were a couple of things we weren't able to reproduce in our own testing. If you have any additional information about them, it would be greatly appreciated
As before, keep submitting your bug reports. The vast majority of the issues above were community reported and some were even community fixed, which is just awesome. Every fix we get into 1.8.1 makes it a better product, so go find some more!
About the author
#22
01/29/2009 (11:49 pm)
@Mikael: Opening and closing the console should resolve that. Anyhow, a full fix would be to go into platformInterface.cpp and change WindowManager->getFocusWindow() to WindowManager->getFirstWindow() in Platform::setWindowLocked. This has been done in 1.8.1.
#23
What I did to patch this behavior is, in particleEmiterNode.cpp, in ParticleEmitterNode::advanceTime, change :
So, before emitting, it just checks to see if the emitter is not NULL.
Nicolas Buquet
www.buquet-net.com/cv/
PS : there is certainly a better way to patch this.
01/30/2009 (12:01 am)
I had too with TGEA 1.8.0 the annoying bug on a particleEmitterNode when leaving a mission : it crashes because the particleEmitterNode wants to emit particle although it is deleted.What I did to patch this behavior is, in particleEmiterNode.cpp, in ParticleEmitterNode::advanceTime, change :
mEmitter->emitParticles(emitPoint, emitPoint,
emitAxis,
emitVelocity, (U32)(dt * mDataBlock->timeMultiple * 1000.0f));which is reponsible for the paricle emiision, to[b]if (mEmitter != NULL )[/b] mEmitter->emitParticles(emitPoint, emitPoint,
emitAxis,
emitVelocity, (U32)(dt * mDataBlock->timeMultiple * 1000.0f));So, before emitting, it just checks to see if the emitter is not NULL.
Nicolas Buquet
www.buquet-net.com/cv/
PS : there is certainly a better way to patch this.
#24
01/30/2009 (12:31 am)
@Nicolas: Fixed in 1.8.1. And yes, while there is a better way to fix this, it would probably require some very non-trivial changes. So, instead, we have a trivial fix (if !NULL) to a non-trivial problem (crashing).
#25
Damn, I'll have to port my game once again now.. no more excuses.
01/30/2009 (3:51 am)
Alex, this is very impressive! I probably can't say anything new, but the pace at which you guys have made these fixes is amazing.Damn, I'll have to port my game once again now.. no more excuses.
#26
Well done with the rest can not wait for the final one.
01/30/2009 (4:06 am)
Hi , can we not by default add database support to MSSQL, MYSQL. I have managed to incorporate connection etc in TGEA 1.8 but now I need to copy the classes in again. How do we make such things part of the actual project.Well done with the rest can not wait for the final one.
#27
01/30/2009 (4:09 am)
ETA?
#28
Didn't realize that you have posted there.
Otherwise the fixed things look very nice, will definitely help.
01/30/2009 (6:13 am)
added a note to the OpenAL problem thread.Didn't realize that you have posted there.
Otherwise the fixed things look very nice, will definitely help.
#29
To see the issues, do the following: Load up 1.8.0, go into the stronghold mission, and edit the scene. Turn full reflect off, and save the mission. Exit.
Reload the mission -
1. Everything is now dark until you get high in the mountains (above the top of the waterblock scale?)
2. Go down and look at the water. The water is completely transparent, though the shader is working correctly (you can see the waves)
3. Edit the water block and turn OFF "Render Fog Mesh".
4. The scene is bright again, like its supposed to be.
5. Look at the water - The cube map is now displaying, but it's on it's side.
6. No shading anymore - no waves.
What I thought was going to happen: The reflection would be drawn with just the cube-map, shaded.
edit: I know the easy answer is "Don't turn off full reflect", but if you want to have multiple waterblocks, or you need to recover some performance on low end cards, this is the best thing to do. If the shader and cube-map were working, the loss in quality wouldn't even be that bad.
01/30/2009 (6:45 am)
@Alex: Did you fix the water block problems when Full-reflect is off?To see the issues, do the following: Load up 1.8.0, go into the stronghold mission, and edit the scene. Turn full reflect off, and save the mission. Exit.
Reload the mission -
1. Everything is now dark until you get high in the mountains (above the top of the waterblock scale?)
2. Go down and look at the water. The water is completely transparent, though the shader is working correctly (you can see the waves)
3. Edit the water block and turn OFF "Render Fog Mesh".
4. The scene is bright again, like its supposed to be.
5. Look at the water - The cube map is now displaying, but it's on it's side.
6. No shading anymore - no waves.
What I thought was going to happen: The reflection would be drawn with just the cube-map, shaded.
edit: I know the easy answer is "Don't turn off full reflect", but if you want to have multiple waterblocks, or you need to recover some performance on low end cards, this is the best thing to do. If the shader and cube-map were working, the loss in quality wouldn't even be that bad.
#30
@Jaimi: It's not in the list, so no, it hasn't been fixed. I will look into it. Also, it's silly that you have to reload the mission. I will look into that as well.
01/30/2009 (9:41 am)
@CDK: That is way out of the scope of a bugfix update, and, honestly, out of the scope of a Torque release in general. However, I've taken some steps that should make bringing over your changes (or merging our changes into your project) a bit easier.@Jaimi: It's not in the list, so no, it hasn't been fixed. I will look into it. Also, it's silly that you have to reload the mission. I will look into that as well.
#31
01/30/2009 (4:50 pm)
@Alex - I hope you're able to fix it. This is a big win on low powered cards. :)
#32
01/30/2009 (4:53 pm)
@Jaimi: Not only is it fixed, but when you change fullReflect in the mission editor it takes effect immediately. And with a bit of help from the good guys at Sickhead, changing the scale of a waterblock takes effect immediately and you no longer have to reload the mission after creating a new waterblock.
#33
01/31/2009 (5:47 am)
Opengl support is only for mac?, when i try opengl on windows the system crash.
#34
01/31/2009 (8:10 am)
I would like to have opengl working in windows as well.
#35
That can be discussed here. For now, I will simply reiterate my previous response.
01/31/2009 (11:24 am)
@Javier, CarlThat can be discussed here. For now, I will simply reiterate my previous response.
Quote:If the OpenGL layer and all new functionality could exhibit identical behavior on OS X and Windows, it would be something we would look into. Of course, if this was the case the OpenGL layer would work on Windows right now and this wouldn't be a problem.
The idea of OpenGL 'just working' on multiple OSs is great in theory, but in practice that doesn't seem to be working out so well.
#36
Possibly out of scope, but it's a serious and limiting bug.
02/01/2009 (2:17 pm)
TGEA 1.8.0 Bug - Polysoup, only first submesh receives shadows.Possibly out of scope, but it's a serious and limiting bug.
#37
02/01/2009 (3:02 pm)
@Stefan: Unfortunately it is out of scope for 1.8.1 at this point, but it's been logged and will hopefully be addressed by a later TGEA update (should we do one) and/or Torque 3D.
#38
TGEA 1.8.0 Bug - WaterBlock can't handle proper grid sizes.
02/03/2009 (6:08 am)
From one frustration to another:TGEA 1.8.0 Bug - WaterBlock can't handle proper grid sizes.
#39
02/04/2009 (2:05 am)
I'am not sure it's an TGEA 1.8 bug, because I tried the demos on a mac mini with intel graphics, but the water is rendered as a black plane. 
Torque Owner Jason "fireVein" Culwell
edit: As an update I haven't been able to get bloom to work on my GeForce 6600. It used to, but doesn't anymore; no matter what settings I adjust. Could have something to do with the fact that my card overheats and I have 4 fans taped to it. Not too worried about it, though, in a few days I'll be running an 8600gt that has been sitting on my shelf staring at me. *waits patiently*
Actually I just made that edit to see if the edit feature is working properly, now. :P