Release roadmap for version 1.8

[CABAListic]

Hey folks,

I have just created a new branch (v1-8) for the upcoming major release, 1.8 (Byatis). This means that v1.8 is now feature-frozen; any new feature development must be targeted towards v1.9.
In the next few weeks, our plan is to get as many bug fixes and paper cuts worked into v1.8 as we possibly can. By the end of October / beginning of November we plan to do a beta or RC release of 1.8. The final release is planned for the end of the year.
Somewhere during the RC and the final we also plan to release one last maintenance update to the 1.7 branch.

If you haven't already, I hope that you will give the v1-8 branch a try and let us know about any bugs which need to be fixed before the final release. Thanks!

[Klaim]

Good news and excellent timing for me!

By the way Is the wiki page for 1.8 up to date on the features list?

[CABAListic]

Probably not Unless someone has worked on it recently.

[dermont]

So what is the "latest/development" branch now. is it still default?

Code: Select all

default                     3148:c7e1484a5a50 (inactive)

[spacegaier]

So what is the "latest/development" branch now. is it still default?

Code: Select all

default                     3148:c7e1484a5a50 (inactive)

Yes, it is. The previous state of "default" was now branched into 1-8, so right now those two are the same, but 1-8 is frozen (so only bug fixes and papercuts will be applied) where as "default" is still the dev/unstable branch.

[Wolfmanfx]

Great! I check it out.

[CABAListic]

So what is the "latest/development" branch now. is it still default?

Code: Select all

default                     3148:c7e1484a5a50 (inactive)

I think the attribute "inactive" means that at this moment default is just an "older" version of v1-8. Once someone commits something to the default branch, it'll go away.

[m2codeGEN]

Wat about RenderSystem_GL3Plus?
Thanks

[masterfalcon]

It's definitely not ready for 1.8. After 1.8 is released I'll revisit it and keep working toward getting it into the next major release.

[m2codeGEN]

It's definitely not ready for 1.8. After 1.8 is released I'll revisit it and keep working toward getting it into the next major release.

thus, it is possible to use the functionality of OpenGL 3.1 and GLSL 1.40 in 1.8

[masterfalcon]

No, that's not what I said. Next major release would be 1.9. A couple of the features are already available through extensions though.

[al2950]

Out of interest is the GSOC 2011 illumination techniques going to be merged into 1.8, as it does not seem to be there at the moment? At least the sample is not there.

[Mind Calamity]

Out of interest is the GSOC 2011 illumination techniques going to be merged into 1.8, as it does not seem to be there at the moment? At least the sample is not there.

That's the "Deffered Shading" you're looking for, it has the ambient occlusion, but some features have been dropped (no idea why, if possible, could someone of the OGRE team, or someone knowledgeable please explain), I really miss the RSM (reflective shadow maps) technique, it was one of my most anticipated techniques.

[al2950]

That's the "Deffered Shading" you're looking for, it has the ambient occlusion

That is the old demo, however the new stuff as you said has things like RSM in them or what i am interested in 'Deffered lighting'. Actually its more to do with the changes and additions that were added in the background then the actual light buffer shaders, for example I believe some additions were made to the new instancing system to help with rendering light volumes.

[Mind Calamity]

I did a whole lot of hacking to compile the modern illumination techniques sample a few months ago, I must have overwritten the Deffered shading, though I'm pretty sure that originally the Modern Illumination Demo was supposed to be an extended version of the Deffered Shading demo.

[andrei_radu]

The modern illumination demo isn't merged in because it doesn't work on ATIs(the reason is that I haven't found a profile that supports vertex texture fetch and works with the dx9 renderer; if you know one or a quick work-around please let me know). If you want to play with that, you can get the sources from the repo(they should work if you try a clean build, not merging it into the current trunk). Changes to instancing were really minimal, I just added a really dumb batch type that doesn't assume that it has a position. It would be useful if you get your position in the vertex shader.

[Wolfmanfx]

@cabalistic
You could use this for command line params -->

Code: Select all

INT WINAPI WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, PSTR lpCmdLine, INT nCmdShow) {
	int argc = __argc;
	char** argv = __argv;

[dark_sylinc]

The modern illumination demo isn't merged in because it doesn't work on ATIs(the reason is that I haven't found a profile that supports vertex texture fetch and works with the dx9 renderer; if you know one or a quick work-around please let me know).

What?
Vertex Texture Fetch is in DX9 using Vertex Shader 3.0

You may have encountered the ATI X1000 line (X1800 X1600, etc) which claims to support VS 3.0 but without supporting VTF (which is mandatory).
ATI got away with it because of a loophole in the DX specs (it says VTF is supported, but which texture formats are allowed is up to the vendor. So, they do support vtf, they just don't support any texture format).

As a result developers have to use vs 3.0 and check the RSC_VERTEX_TEXTURE_FETCH flag is present.
Blame ATI for cheating.

The ATI Radeon HD line, all of them support VTF. NVIDIA has always supported it in VS 3.0 (though GeForce series 6000 & 7000 have just a few texture formats available only, I think they were just 4, all of them floating point)

Cheers
Dark Sylinc

[Mind Calamity]

@Dark_Sylinc:

Well if all ATI Radeon HD cards did support VTF, then that would mean that a ATI Mobility Radeon HD 5650 would too, right ?

Unfortunately, it isn't so, I have the said card on my laptop, and the Modern Illumination sample did not work for me (see the broken images here).

The sample runs perfectly well on the NVIDIA GeForce 9500 I have on my PC, even though it's inferior to the laptop one, but I'm not that well informed, as I haven't done much research on the subject.

[dark_sylinc]

Hi, yes ATI Mobility Radeon HD 5650 supports VTF.
I will post in the other thread so this one doesn't get Hi jacked.
Try NewInstancing sample. It has three VTF techniques (VTF+HW Instancing VTF+HW+LUT and VTF alone). They should all be working.

[dermont]

So what is the "latest/development" branch now. is it still default?

Code: Select all

default                     3148:c7e1484a5a50 (inactive)

I think the attribute "inactive" means that at this moment default is just an "older" version of v1-8. Once someone commits something to the default branch, it'll go away.

Yes it was as you said.

How often are the changes from the v1-8 branch merged into default? For example the following fix appears to be applied to v1-8 and not default (therefore the default build fails on Linux). Do we have do a "local" merge until the changes are applied?
http://sourceforge.net/tracker/?func=de ... tid=302997

[CABAListic]

Right now, pretty much all development is focussed on the v1-8 branch to get it release-ready, so default is currently lacking behind. I'll definitely do a merge when we release the 1.8 beta.

Ultimately, I think merging every week is a useful schedule, I'll need to build up a routine to do that regularly

[Mind Calamity]

Well great, and - a question: How much of this list of notes is complete, and how much remains to be done ?

I'm not up to date with the DX11 render system, but I think it doesn't support tessellation yet (I think I read something like that a while ago).

[Mikachu]

Another question : are there plans to integrate Gorilla into Ogre for this release?
If I understand correctly, its original goal was to replace the overlay system, with better perfs and features..

[masterfalcon]

I've been doing some testing of 1.8 and have a couple questions for the community:

Wondering if anyone is having issues with the HDR, Gaussian Blur and TestMRT compositors. Or if the black screen that I'm seeing is OS X or GL only.

Anyone else seeing several problems with the RTSS sample when using GLSL? Again, I'm trying to pin down if this is an OS X or GL render system issue.