[Re] Vertex pipeline replacement

[Tobias Jakobi]

Short feedback:
- Max Payne 2: works as expected, I don't see any visual changes
- Morrowind: 3D rendering breaks completly, minor (blending?) issues in
the menu as well - however ingame scenes are fully black now

How should I report this? Open a bug? Should I do any regression testing
(since your patchset consists of more than one patch)?

Greets,
Tobias

[James McKenzie]

Tobias Jakobi wrote:
> Short feedback:
> - Max Payne 2: works as expected, I don't see any visual changes
> - Morrowind: 3D rendering breaks completly, minor (blending?) issues in
> the menu as well - however ingame scenes are fully black now
>
> How should I report this? Open a bug? Should I do any regression testing
> (since your patchset consists of more than one patch)?

Do a regression test to locate which patch breaks Morrowmind. 

If you know which patch might have broken this, start there.  If you do
not, a normal regression test is a good idea.

James McKenzie

[James Hawkins]

On Thu, Oct 30, 2008 at 10:07 PM, James McKenzie
<jjmckenzie51(a)earthlink.net> wrote:
> Tobias Jakobi wrote:
> > Short feedback:
> > - Max Payne 2: works as expected, I don't see any visual changes
> > - Morrowind: 3D rendering breaks completly, minor (blending?) issues in
> > the menu as well - however ingame scenes are fully black now
> >
> > How should I report this? Open a bug? Should I do any regression testing
> > (since your patchset consists of more than one patch)?
>
> Do a regression test to locate which patch breaks Morrowmind.
>
> If you know which patch might have broken this, start there.  If you do
> not, a normal regression test is a good idea.

Tobias is referring to Stefan's RFC concerning his vertex pipeline
replacement patch.  Stefan asked for users to test the patch to see if
any games break.

-- 
James Hawkins

[Austin English]

On Thu, Oct 30, 2008 at 11:22 PM, James Hawkins <truiken(a)gmail.com> wrote:
> On Thu, Oct 30, 2008 at 10:07 PM, James McKenzie
> <jjmckenzie51(a)earthlink.net> wrote:
> > Tobias Jakobi wrote:
> > > Short feedback:
> > > - Max Payne 2: works as expected, I don't see any visual changes
> > > - Morrowind: 3D rendering breaks completly, minor (blending?) issues in
> > > the menu as well - however ingame scenes are fully black now
> > >
> > > How should I report this? Open a bug? Should I do any regression testing
> > > (since your patchset consists of more than one patch)?
> >
> > Do a regression test to locate which patch breaks Morrowmind.
> >
> > If you know which patch might have broken this, start there.  If you do
> > not, a normal regression test is a good idea.
>
> Tobias is referring to Stefan's RFC concerning his vertex pipeline
> replacement patch.  Stefan asked for users to test the patch to see if
> any games break.
>
> --
> James Hawkins

Apply a single patch at a time until it breaks.

-- 
-Austin

[Henri Verbeet]

2008/10/31 Austin English <austinenglish(a)gmail.com>:
> Apply a single patch at a time until it breaks.

You'll probably just end up with patch 12, which is huge and enables
the entire thing.

[Stefan Dösinger]

> If you know which patch might have broken this, start there.  If you do
> not, a normal regression test is a good idea.
> It will always be the last patch. This patch does the real thing, the others
> are just making the infrastructure more flexible. So a regression test in
> this case is mostly useless

[Tobias Jakobi]

Stefan Dösinger wrote:
> > If you know which patch might have broken this, start there.  If you do
> > not, a normal regression test is a good idea.
> > It will always be the last patch. This patch does the real thing, the others
> > are just making the infrastructure more flexible. So a regression test in
> > this case is mostly useless

So what should I do then?

What about opening a metabug and adding all apps what break with the new
code? Or do you wanna have all this in separate bug reports?

Greets,
Tobias

[Stefan Dösinger]

> What about opening a metabug and adding all apps what break with the
> new
> code? Or do you wanna have all this in separate bug reports?
> Just file a bug report, assign it to me, and make it clear that this new
> code breaks things.

I have one reported problem fixed already, and I am currently working on
AoM. I'll send an updated set of patches later today.

My sense is that there aren't many bugs, but more like 2-3 problems breaking
a lot of games.

[Stefan Dösinger]

Here is an updated version of the patches. Can you retest the games? (the
ones that worked already as well as the broken ones)

What are the differences?

-> Proper handling of RHW vertices which do not come from a vbo(should fix
Star Wars Jedi Knight 1)
-> Proper handling of the alpha value in the lighting code(+ a test). Fixes
the water in AoM.

Those issues had a quite big effect potentially, so more games may be fixed
now.

[Rico Schüller]

Stefan Dösinger schrieb:
> Here is an updated version of the patches. Can you retest the games? (the
> ones that worked already as well as the broken ones)
>
> What are the differences?
>
> -> Proper handling of RHW vertices which do not come from a vbo(should fix
> Star Wars Jedi Knight 1)
> -> Proper handling of the alpha value in the lighting code(+ a test). Fixes
> the water in AoM.
>
> Those issues had a quite big effect potentially, so more games may be fixed
> now.
>
> ------------------------------------------------------------------------

Oblivion has rendering errors in the menu (white boxes). Also the 
status/action bar isn't drawn correctly (both with fbo). Should I file a 
bug?

WarCraft 3 had some glitches with the previous version but it is fixed 
in version 2 of the Vertex pipeline replacement patch.

Cheers Rico

[Stefan Dösinger]

> Oblivion has rendering errors in the menu (white boxes). Also the
> status/action bar isn't drawn correctly (both with fbo). Should I file
> a
> bug?
> Yes, file bugs please. I think I have Oblivion here somewhere, this should
> be easier to debug than Morrowind.

Note that it will be a while before I send another version of this. I did
some benchmarking, and there isn't much of a performance improvement. In
most cases the replacement pipeline is slower. We still need it some day, so
I'll continue the work, but this makes it less of a priority. I also think
that we can improve the performance quite a bit, but it will need quite a
bit more work.